<%BANNER%>

Estimating Consumers' Willingness-to-Pay for Country-of-Origin Labels in Fresh Apples and Tomatoes: A Double-Hurdle Prob...

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110318_AAAALO INGEST_TIME 2011-03-18T20:07:06Z PACKAGE UFE0011841_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 5002 DFID F20110318_AABZKH ORIGIN DEPOSITOR PATH mabiso_a_Page_080thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
d2942d59ff5db403e7c6e376156f8085
SHA-1
46fe19fe712cc38f1f952749a914389ff670d357
6144 F20110318_AABZJT mabiso_a_Page_062thm.jpg
1552835c4b9c91e7f5240e5c6ce2fbf3
66d548fac855d7c2acf3317e8f4d46cf516cc98d
1053954 F20110318_AABYHF mabiso_a_Page_095.tif
4e3ad80e236c26b4dfefc78ffbeffbf8
f67793d6574295f2a30de72bbac56b3dba2a3a9a
25271604 F20110318_AABYGR mabiso_a_Page_078.tif
41132138458280cbaf59de0ed6d7e93f
0ddee3c8079e9c15cafa6ddf80e1eaf8c4c46c80
4470 F20110318_AABZKI mabiso_a_Page_081thm.jpg
6efeb409118b1fc6489704d3210dfc68
00f51863496b2d94abdaa2cd0cb454bd207ff824
6068 F20110318_AABZJU mabiso_a_Page_063thm.jpg
8c88747ee3df8045f3f8d91e3ec84b50
4d227b6b14326b7ec89292c5d253ca930a706930
F20110318_AABYHG mabiso_a_Page_096.tif
710fa32b3fedaf21e7ba69607dfe8e24
e0d438bd9d083b564c6ddbafeb96a87ca343aa31
F20110318_AABYGS mabiso_a_Page_079.tif
16486b1c223cb2eae246e9dff25e4abd
a85db629d4fefb1dbd127414c61fa141231cf7c8
5309 F20110318_AABZKJ mabiso_a_Page_082thm.jpg
aaa896e070bd62e1327415a80d8c9e4b
6dbb92d3a89fc3e1feddab7824c09a8d0aa31b96
6250 F20110318_AABZJV mabiso_a_Page_064thm.jpg
8d4cec0efde3cc1688b2348ef50b4859
2f4b3b4997ee87005f76d095b26b5cd08849815c
F20110318_AABYHH mabiso_a_Page_097.tif
7bc05bc2dc6e700847e4182c7926cf8b
b28e9cd59abb7bec88b8ab8c230d39d1a2831f89
F20110318_AABYGT mabiso_a_Page_082.tif
1445cf166e19b30a5a390ed562f39d4f
2048f1ce67e658d801933fc0d0b4ac78c06dd44a
6237 F20110318_AABZKK mabiso_a_Page_083thm.jpg
9a661453b3b245a93e7548e2075b04a3
a00de64208b8bc5c77d7a8c5b00d3252ffbdd809
6055 F20110318_AABZJW mabiso_a_Page_065thm.jpg
36f3379ceffc37482157ae9beb0d74fe
f4373c1bd72790da7ce47de4c74d61fea853670a
F20110318_AABYHI mabiso_a_Page_098.tif
f7e5f01aabd3cf223b228998db654478
b4ac7642724ce4060b5bcd7d705a345d524c22a8
F20110318_AABYGU mabiso_a_Page_083.tif
91856ec0ec27e061bdc7a2a00faf2319
586656df6dccaed51efea44b3ec134944aae9103
2765 F20110318_AABZLA mabiso_a_Page_099thm.jpg
03a16b141113e757e14fa25f9471a292
75b048af918b52b0e6df2188348d7eb8fcafad2b
5679 F20110318_AABZKL mabiso_a_Page_084thm.jpg
8d5ae44e0d0eb0923311fa48161a16d0
59fa9db865a4edab5d867abc679c8d641b9c87e4
4920 F20110318_AABZJX mabiso_a_Page_067thm.jpg
e473003af2a750e3461138cd7a305251
fedef817da9464f66653b12548bcb87b89017c4a
F20110318_AABYHJ mabiso_a_Page_099.tif
990a286d8e44cb1ebb1b95c3b30bde18
202dccb11f3ae0d62a1a0d151baea2f91c6ef7a6
F20110318_AABYGV mabiso_a_Page_084.tif
49e03fc474b02e21f8e1edb1088f2c9f
0e4b259c9b1cdfd4240edc8bdd1abca4c41d3892
5415 F20110318_AABZLB mabiso_a_Page_100thm.jpg
3126fa8641d452eecf9047337b976722
b65e28f2b0a628492e02e923a800c73b1e8b8f70
6048 F20110318_AABZKM mabiso_a_Page_085thm.jpg
f9060f4ee8881fe7971ceee96a98c75a
bc3633895c3fda3947f89fcf4a417b2af64c271a
1949 F20110318_AABZJY mabiso_a_Page_068thm.jpg
8f2057643bd6904452cb5545a5ac37d1
2b99152c58daedcffa160e85c4cfe470c797335a
F20110318_AABYHK mabiso_a_Page_100.tif
73a36668d32ca4b31ccac406bf9ce4d0
9fe4d5ce94dd88fc7e42bd53f0d17b4921b9b8c7
F20110318_AABYGW mabiso_a_Page_085.tif
d98986ab8ad644197b15578e59e86992
c6fb2c83a1c45f79f9f14d20f929444623dcf967
6248 F20110318_AABZKN mabiso_a_Page_086thm.jpg
febfebe7836b2b3eef876f5e23d56155
39f3b434712977dc8b8802238ab26bc19d716e85
5707 F20110318_AABZJZ mabiso_a_Page_070thm.jpg
64eb28b974cf8960553a866b10c437f1
aa54f0d315879b6987ca10fb24728876582c7da6
F20110318_AABYHL mabiso_a_Page_101.tif
d2f03edd8430c90a8634cd8c571ba14f
0d2ae515286404f9127dffa656bc0d11650a780d
F20110318_AABYGX mabiso_a_Page_086.tif
55fc610c73dead3e8303cde0197ef049
86dd57f94d87a8fe0f27cc4d1a89442d02003d7d
6687 F20110318_AABZLC mabiso_a_Page_101thm.jpg
3075cc18763f7346e88faf32b14e8fe9
47e043501bda9401625d2e5069b6f5cdb8020298
6172 F20110318_AABZKO mabiso_a_Page_087thm.jpg
856a15d3dcb55b1631bee402bf5003d7
1c6341397ff58ae2afcfddfccd228fad64595b60
F20110318_AABYIA mabiso_a_Page_118.tif
a1d4ad89693c69283eb0c16f775e2a69
72ad1318b5aacb86ad27f5a95251b8a9b3e5225d
F20110318_AABYHM mabiso_a_Page_102.tif
5c87b3297f84d49a5f8a14142358f176
f2becbd21824c7e9d0147bc6563bf6d3528b4afd
F20110318_AABYGY mabiso_a_Page_087.tif
a90020c2667e3cb91872c5880a912028
a12dd856fb63e0e7f8be6b925d75c854db18854a
4520 F20110318_AABZLD mabiso_a_Page_102thm.jpg
acad3c3dd09ec5aca179d9500a9a72ef
2ac80c2bd5244bdcabbe9e0e576057c0f42645c1
6445 F20110318_AABZKP mabiso_a_Page_088thm.jpg
62e11df3c35a5823105df688e7eee5a5
eca2f53f27e5a6f57ede658481af2336b8b9194f
F20110318_AABYIB mabiso_a_Page_119.tif
20cfa60069d153e72ee0b48f803332fd
c4139cd58427d84fcfaa9e58d72282be7da299aa
F20110318_AABYHN mabiso_a_Page_103.tif
054dccafdb4576ad88ac045d704d9202
799adf2f1140a8b9f21dacbff0e3902706540c9c
F20110318_AABYGZ mabiso_a_Page_088.tif
0c1f449c28ae4ce380c39e76240fff2d
1154c01421d3e51d90750993dbf7a2517b9b89fd
4829 F20110318_AABZLE mabiso_a_Page_103thm.jpg
65576db3c7329a75cdde8203bff44927
ed3a719394170a38b400c1b6db0803374bd8fa52
6401 F20110318_AABZKQ mabiso_a_Page_089thm.jpg
1216b236a2ef8f6b65b3348588818e69
cf35e3745651778e34c102277d0a9b21f7e862b4
F20110318_AABYIC mabiso_a_Page_120.tif
e82ddeb2d489c4cfda50edc0b9ee59ed
e641e5daa181f78ef83db78a4403593ee80f7b3a
F20110318_AABYHO mabiso_a_Page_104.tif
b1126be406b75e884d79750eea00e637
c4312e373268e8ba61af78cad80c1bbc454af2e3
6641 F20110318_AABZLF mabiso_a_Page_105thm.jpg
07755f425bb9f880cc0b9e02de870602
50733fbc4b0fce152654a684f8a5ad3a7c2e4009
3277 F20110318_AABZKR mabiso_a_Page_090thm.jpg
62e3a8606090ca6f4523ee0007e66706
1d0afbfa3f465a692391ae1f26fe6e8668611dae
8423998 F20110318_AABYID mabiso_a_Page_121.tif
8abddb40388f3a2bf2ecdb0f3a1ab345
3abf900da9553b505d8130504d0dbfe08af5577f
F20110318_AABYHP mabiso_a_Page_105.tif
5c719bcc1727381518cec532539b9778
bfbc1d748f5ea77767287627f91b08a60e047bff
6855 F20110318_AABZLG mabiso_a_Page_106thm.jpg
1e298eec54f181a8516b8ed093ee46b3
28800a1865a803117d3980f015707eafbe91259c
5179 F20110318_AABZKS mabiso_a_Page_091thm.jpg
fdcadf4aeced3164f3f5847f94d13c90
dd893e33b62a3d08f065f8e33a625671c66ab909
F20110318_AABYHQ mabiso_a_Page_106.tif
333fe135c2600c0e4d5e18db235cebf5
84567535b1b61ca1a2b0d9201d8aed84fdac99fb
F20110318_AABYIE mabiso_a_Page_122.tif
d101043b781003391e8bff283f6f73a4
f4f7a98b5e4a96bec056f00d32686742a420f0ae
6412 F20110318_AABZLH mabiso_a_Page_107thm.jpg
0fd07f6679cd6b4187fd03654a7ac8e8
6cf268aa994f4e0e4888e49974a1263545d2be62
6367 F20110318_AABZKT mabiso_a_Page_092thm.jpg
54744916b1ade124aeb5c6b1ef3ab6e7
013f8d805d9004e40c6fa34f5f90fcf730d1a05e
F20110318_AABYHR mabiso_a_Page_107.tif
c22b066e5a9265b375ce0b186f73337d
c068fffca9de04c0cad075c149be5e7c6f42e62b
F20110318_AABYIF mabiso_a_Page_123.tif
dec355041510f09ff79be67de21c7be5
2db10c7dbd26e28700aa21e4e5beeb8e7c8cb86a
7147 F20110318_AABZLI mabiso_a_Page_109thm.jpg
c8161b873b15131cad222ba105120598
d532225656632487392e5201b52d76561f50c4e1
6649 F20110318_AABZKU mabiso_a_Page_093thm.jpg
7e1d0531c1066ae05e8f12e45f996953
eec640c1286897bd02844a30de4a14fe8f0d461d
F20110318_AABYHS mabiso_a_Page_108.tif
4df2502819ded86d5ca8a200acf11232
2ccad3a944f688e51c55eb4631a52948ea637f80
F20110318_AABYIG mabiso_a_Page_124.tif
c99824af1de85c13a17070e8d86b9440
32fc319a9fbac3091af1ad93cc2645dee946e66c
5835 F20110318_AABZLJ mabiso_a_Page_110thm.jpg
68d98185a6684d97741ce0f682b06c3c
29648abcf20cb407aa6460115c3880b1521471bd
6283 F20110318_AABZKV mabiso_a_Page_094thm.jpg
dc703e327d78eb5205b3a49f37e84789
33b0dcaa633db7effb1aafc2e9fbeaef1e868d1f
F20110318_AABYHT mabiso_a_Page_109.tif
8da07f7b81d84ed8dbc7c7d1f8447046
e49463fd1ac1dc2a765d1f130d68deefc6f61d14
F20110318_AABYIH mabiso_a_Page_125.tif
a7f0e4e4a52c33bba61fa2b55ebd514c
b48595b55c6f3c258154fb6208976043cb535798
462 F20110318_AABZLK mabiso_a_Page_111thm.jpg
f53bb3efd018f5f82737ff3198741196
4c40e0316261c32ce0701a89b13a58d98ecd33e3
6455 F20110318_AABZKW mabiso_a_Page_095thm.jpg
620a9a1ca1aed1ec23a06fe9672eb0a4
d45c2f519db27c9a701628c3babbdd998f00ce55
F20110318_AABYHU mabiso_a_Page_111.tif
8dbd11d7e1919306e7976a46c334b02f
ca7625d36b4d4fa7839a1089781169971f2b2c9b
F20110318_AABYII mabiso_a_Page_127.tif
0c2ba5eaaa885a20bebadc1a009e4624
168ebe7ac28c2b57994fa1a3a8be2d7838aeefbe
4798 F20110318_AABZMA mabiso_a_Page_127thm.jpg
752ed2042f5c082d060dbe39ff02793b
c3664c9d5005012be27ac6ae8dbeb616e9d8531e
4501 F20110318_AABZLL mabiso_a_Page_112thm.jpg
e92ac3dee6dca336cb31ca3ab5bc5cac
d14204e68f0f50aa9566d5b2ed04e6849893e3b4
6102 F20110318_AABZKX mabiso_a_Page_096thm.jpg
16b35981c46406468ba0316f9ff0014b
7d3e57553df808d4428f2792de2a7c85c5da41b9
F20110318_AABYHV mabiso_a_Page_112.tif
ef4606e8abec5b05cb006ff75e2526d8
7ec10f9e0b2ce969bbca11d235571df370b2c47c
F20110318_AABYIJ mabiso_a_Page_128.tif
69749b1bd0eff43b4ed959916f9ec321
ca84812b1efa736fe3b7ddfd46a56487d346d606
5188 F20110318_AABZMB mabiso_a_Page_128thm.jpg
4a6804fb9010658ba870c7dd076c489e
c8b97934864780660eb2f0d093c3b5e656c1e276
4287 F20110318_AABZLM mabiso_a_Page_113thm.jpg
8bf6b6a4f717dd471d0fbc9bb821a7ea
7fce82f971732d951e73376cc6e81b047e6a867b
6420 F20110318_AABZKY mabiso_a_Page_097thm.jpg
75712ca95a288e7df072f49f22f568c8
deae52b3373d3d9f8ed9061db38362d95abb7756
F20110318_AABYHW mabiso_a_Page_113.tif
29ff0314a6ab589b7666201038329438
b7ff804029dd4eb93b2eb187202066a186a7c8a5
F20110318_AABYIK mabiso_a_Page_129.tif
7e5efea9ea427b3b973fae1981ac02bf
e2915f93d78b2986bc8531f709c24d88649f479a
4587 F20110318_AABZMC mabiso_a_Page_129thm.jpg
51215a3e9f889c650c7cb70f1c87bbe5
1267539ad82cee443771fe1d8571102fc8a44214
6506 F20110318_AABZLN mabiso_a_Page_114thm.jpg
5985c8323bac5eca1dd48af54f04a4bc
cc4606f6df90335cc1dc7093d23895251b4a23e5
5894 F20110318_AABZKZ mabiso_a_Page_098thm.jpg
754429c77021a0cc46d8b86a306590e6
4923e0888e7b5c8c5376d178f7748c775b3c47e9
F20110318_AABYHX mabiso_a_Page_114.tif
9111582edf5e110178d039b6fe28a696
7b1e8e801ee99adcd5d4ca77bef5a9fe159c0e0f
3279 F20110318_AABYJA mabiso_a_Page_006.txt
b830b6143bf32262e17d10c2e1a9e662
26977ed2e2580b4a251d701a22e5a5d0107fc092
F20110318_AABYIL mabiso_a_Page_130.tif
785385f6dbec90970637651378982092
313b7ba0bdcaca818654d6139cbd9525e8fe720a
6382 F20110318_AABZLO mabiso_a_Page_115thm.jpg
534121dadf32926ce65eec3a69ceb544
477c886057bfaf96a5901cf090ebd6c77af540d5
F20110318_AABYHY mabiso_a_Page_115.tif
9978fb46cdedaf583cb98c380dba3cfc
a4d17739de52e178f65fd291612f929ee2cf346a
F20110318_AABYIM mabiso_a_Page_131.tif
f077daa850332ad1c31e39afd1d131be
e218b0ee93d9d9f8a23b72ef718979fffaac0b4f
4985 F20110318_AABZMD mabiso_a_Page_130thm.jpg
50985629930e206dac49ea6645f3cdaa
463814394af430901a58c8787b75daf102376fd0
2872 F20110318_AABZLP mabiso_a_Page_116thm.jpg
936dca38e7230d741566ec6303cd45c8
ee8ed4c1afd5b472e60ea6f30de1788e87094e84
F20110318_AABYHZ mabiso_a_Page_117.tif
479e5f77410e30088623e2afb5843d73
12a375312c07028822a2779d02c9df4f450cd053
255 F20110318_AABYJB mabiso_a_Page_007.txt
72570b63691a4e257d79563fa3817213
3231a3a82bcceb42b2d49ecb281316a9e47e6a12
F20110318_AABYIN mabiso_a_Page_132.tif
16f6d7b5c61b6df0c866390ac8a2a756
7cb5b5dc64cd8d6febe9532d43f5426e6d72679b
4296 F20110318_AABZME mabiso_a_Page_131thm.jpg
5c75fe9bb8b94c519b63fffdbb1dc2fd
41e832a9217257827795acf22689064c047f2a27
6286 F20110318_AABZLQ mabiso_a_Page_117thm.jpg
47487dbcb43eddf195abc1669986beea
ef2992a6c16d88790fad166376d8d5927ac4e780
2169 F20110318_AABYJC mabiso_a_Page_008.txt
f65927b17fc9a07806e7a4a5201432d1
a0c8fd31539b50f1fe6fb2d59d8708dccb9b8efe
F20110318_AABYIO mabiso_a_Page_133.tif
c2df098573c4fe7b6f951694452af3e5
9790fd07bf4e378b180a06d3910a2809b5263fd7
2441 F20110318_AABZMF mabiso_a_Page_132thm.jpg
80be9275cc1652ffa961a6d5d9756ac9
dac0de7fee7ca3327788149037356fca76ae0ffb
4444 F20110318_AABZLR mabiso_a_Page_118thm.jpg
cd7fc3cca722eb5bfeb94bd502755e79
9fdb88f7eb328ec46801f99b3f940e87004f1036
1489 F20110318_AABYJD mabiso_a_Page_010.txt
d996762290d98bee3a22db575b2b0855
b926a7f741ac2d64d444d5857570a08031e849a3
F20110318_AABYIP mabiso_a_Page_134.tif
5a420982fdf3d3aced9b864a53121eb9
352fb6bd07b339295cccf76628d344231c04f746
3301 F20110318_AABZMG mabiso_a_Page_133thm.jpg
66f3c827243278ac20361d08d4701fb1
42be1b67e5fc07fe432496345e622ce4f0b239ed
3889 F20110318_AABZLS mabiso_a_Page_119thm.jpg
fa846af40adfc2f51edf0fbb7dd13640
827cd1e8db8c79b2dbd5f94c77da9dbde9cf853e
1808 F20110318_AABYJE mabiso_a_Page_011.txt
c64cc129f9dca0d2ca34453decbd574c
548af5aad1166fb46fa3fd34204b63cc4fbdb83a
F20110318_AABYIQ mabiso_a_Page_136.tif
67879eba9deb19733d349715442779b5
45fb4546fe51c73058397c7a26f27fce0bf55a84
3888 F20110318_AABZMH mabiso_a_Page_134thm.jpg
fca9977473db30a9531b0cc59e46e819
65f7577aa58f1a719a8b962a1f29a9f6331b39e2
5863 F20110318_AABZLT mabiso_a_Page_120thm.jpg
e4841f1c3a22c64b02fde7d5fe2b2b03
f099d3e15308d3b02892f94f8e1dd3e5d9c165e0
1552 F20110318_AABYJF mabiso_a_Page_012.txt
641af556f4e4517ae5c60f8608812f36
fbe86d4609849ecdb6225e5dec74cc0b222f7584
F20110318_AABYIR mabiso_a_Page_137.tif
97f1602d8106290f33fd24428f89cc05
4535724443bd56912462e406b324a86236d5b7e1
2174 F20110318_AABZMI mabiso_a_Page_135thm.jpg
eb15c0decae670205249195f29434f1c
7e6c4169a45a9fa64fb5ab9d08a38f7d4621bf0d
3326 F20110318_AABZLU mabiso_a_Page_121thm.jpg
294525221cdb2143b66dd15c036534df
03ea509ce85ef68b71742b03870daccb9a2c118d
1739 F20110318_AABYJG mabiso_a_Page_013.txt
1e150b3cf99658e53f221fcaa8d97f8a
255ea096554b62adc15f73ec3788cf7e8fbe38c3
F20110318_AABYIS mabiso_a_Page_138.tif
da9c514a831f57995f790b89547e3807
82f73b516cd2f79c09c544fa08591fc06ef37023
6821 F20110318_AABZMJ mabiso_a_Page_138thm.jpg
1622db4d557718bb9c7417ea2d26f85e
43dbe7db4ab5719971b229a5d9f7607506e873ab
3438 F20110318_AABZLV mabiso_a_Page_122thm.jpg
56e479fb1f98edbad6aa665397df2997
c316e8f5f26fdf2839721182e81414d22c055273
1970 F20110318_AABYJH mabiso_a_Page_014.txt
a091e033001ab07abfffb948d47a7b1e
246563a90fcce007f6bef088a93df64719510fbb
F20110318_AABYIT mabiso_a_Page_139.tif
be75710866d9c0e708de915849757c16
e4c70cd3569077aa1be2fc7a2cf614d40307ee1d
6904 F20110318_AABZMK mabiso_a_Page_139thm.jpg
f67b02fef1e40b866faacb6781f0625d
9a3567608dc664c551802f594a2ed93e86862e99
4652 F20110318_AABZLW mabiso_a_Page_123thm.jpg
d25b01183029238dc4ecb5687761fe95
ad86c34ede57c2b948e8af1580eaf3c1b00d6b60
1944 F20110318_AABYJI mabiso_a_Page_015.txt
b4e4b894b663871380a4df2d3164a229
8a57d66a09dc9f9be5846868c4700311a35f3b39
F20110318_AABYIU mabiso_a_Page_141.tif
9c2702febd49b978f90c9cc0c1812394
26c0331c28e0d8147bd5f261538a5d9d3865e57c
5895 F20110318_AABZML mabiso_a_Page_140thm.jpg
f38aa13873ee41e1011305b6aff5c4ed
d9e955be37acb8f3184fed003fdab3534fd5f020
3666 F20110318_AABZLX mabiso_a_Page_124thm.jpg
853a24f45ac5df4b06a2b56e19da4839
b358476d4d17217b73544fc93364613ef13cf8d2
2128 F20110318_AABYJJ mabiso_a_Page_017.txt
4cb572ecb78dfa512334171293f6d185
2bd66b13dacee283cb5376ed8169d5ea07287126
533 F20110318_AABYIV mabiso_a_Page_001.txt
279743017e7eb59789f9f7f12f0eb2d6
4c32a4777341d86e35163030a3c799b168cee82e
2081 F20110318_AABZMM mabiso_a_Page_141thm.jpg
fe5df5be587b10816e7be53599d05333
8e097578684b580b73212193213dfc2cd9f23a02
601 F20110318_AABZLY mabiso_a_Page_125thm.jpg
66d5990df62d91337a85a4ad9e283d45
b52184fd06a677ce4e3e9a5284b6f3ff8d7efb9a
1832 F20110318_AABYJK mabiso_a_Page_019.txt
253210caae5e0630acdea5e815d6aa33
37b481092f921dac36f72d50fde8f874c31dfdad
107 F20110318_AABYIW mabiso_a_Page_002.txt
9b36cc2db361684a25434a6405d84293
a908b506fc65242fd59b318141596ec6073aaab6
586955 F20110318_AABZMN mabiso_a.pdf
43d75fdf4c621f57c3a094c01d86f5bb
cfec06e83f667da464f2fb377643ac8e2537b2ee
5933 F20110318_AABZLZ mabiso_a_Page_126thm.jpg
cf22a43c10eebf2d65a9d6219a7a76f8
5d8ac142c84d661820005f99b2994c14173b90e2
2052 F20110318_AABYKA mabiso_a_Page_036.txt
c21f36bd9fa79497f52529e3666075bd
67e6d55632e38faee32f8737a0234139b1174da7
2029 F20110318_AABYJL mabiso_a_Page_020.txt
16a7a78ac1ae5ef7630fa527442fecfc
8ac9e9495a24bd19c8837a183c7fa3e5ee54f356
82 F20110318_AABYIX mabiso_a_Page_003.txt
c0f82858de8a262db7ff805ef75df187
6ae7301f65cb9a1997bb0c25fd0a47e57186d869
165712 F20110318_AABZMO UFE0011841_00001.mets FULL
20c2dc898721407a1eb08aa7b51f1c47
e24e780a7879b8a76ca3256faf1c9a18d11e1443
2030 F20110318_AABYKB mabiso_a_Page_038.txt
e1658d81e32ea9ec13f714a58e9664d9
ce4644e8d00593cc5b24e7053102bb271a788bf1
1962 F20110318_AABYJM mabiso_a_Page_021.txt
229fdc4ca6a014fc80fa7a876fff225b
c8c39f87c5d4e1814669c7ed79abb80f9b3d91fc
1634 F20110318_AABYIY mabiso_a_Page_004.txt
3503972a5a943d27d13729e8fd032928
3636828d0955598456789f063fc08344b2d86f8a
2001 F20110318_AABYJN mabiso_a_Page_022.txt
bc2e019e56c438bc3a63d0e1107b279a
4af25f0049a1e5f15c5c8e4dddd89e2ead6180ca
2922 F20110318_AABYIZ mabiso_a_Page_005.txt
63453876040e5e7d7db5a892a5aa54c4
ea2438551ecb7aa0a4a78436cddb8f47c8fe29d8
864 F20110318_AABYKC mabiso_a_Page_039.txt
95c91afc739eed59b607a568ff8dde03
5836d13cd531597f6827a1a600e7b1ad36058f68
1921 F20110318_AABYJO mabiso_a_Page_023.txt
f954462144bfe4a8edf495eb13a56b91
6af9b8f4a5874af85fddd450378a0a744cb0c095
1783 F20110318_AABYKD mabiso_a_Page_040.txt
cffea424c4b4f0f66be86cd74e2a6ba0
8da72d56331d58e3e8903634ea1e539496fcb614
2060 F20110318_AABYJP mabiso_a_Page_025.txt
6f809936272ec65088aac09395602fdd
6089e45321d3b7fb93be8187f3cb925659224870
1822 F20110318_AABYKE mabiso_a_Page_041.txt
30dc2dbcbe69dca334eed0905f8dae12
928beb05c6d1bd1d2c15480cb13160d1c8802ae7
1870 F20110318_AABYJQ mabiso_a_Page_026.txt
8cddd7df82eeb076338c4d7683555d65
bb5f926cf31afa3238ebff702526bcc62caf516d
2004 F20110318_AABYKF mabiso_a_Page_042.txt
68f459c10fcd98b90184bd8028c1749e
9729c20cf86036d84335ee324b709b6ef6ecc47c
1956 F20110318_AABYJR mabiso_a_Page_027.txt
e3e0e56455a84da917cf114432e8e06f
a3908555605a3c906fb616af17884fb464c6a7c6
1852 F20110318_AABYKG mabiso_a_Page_043.txt
77dd4459e6c18ae036fbc5af98152789
6c5379da84b3155169472e1c7aae969d46e7b126
1927 F20110318_AABYJS mabiso_a_Page_028.txt
76badbacb187cfd64436aa52ac91aeab
55d1699bf335e98c446b7c340c79bc6549f92e61
1654 F20110318_AABYKH mabiso_a_Page_044.txt
6ccc0e12752a9168b047056bd27aa5e2
4cc418bf6c8cef09537c8d4d15f28f32d5805005
1587 F20110318_AABYJT mabiso_a_Page_029.txt
54c342408598609a05e9f9ca2442e7a3
35ab42337030513cd331191169735791599b47ad
1849 F20110318_AABYKI mabiso_a_Page_045.txt
c4662c28d9bea29ad248f6a868c9078e
8e15743c0cb4eef24406067086fe4f4982730eb8
1255 F20110318_AABYJU mabiso_a_Page_030.txt
e416918f1aec92c27252be4d70150c99
3f4c8626f219a20e83ae0816d8b5a795fd93c451
1504 F20110318_AABYKJ mabiso_a_Page_046.txt
cb68b6ef39db3e933b0c69602a631022
037558c60f12da3bf3237d99bbe6c0250a8eb6fc
1304 F20110318_AABYJV mabiso_a_Page_031.txt
a23c3a08e1d7575c1c008ef1ac786ccb
adb2931c6987c8bb8b13cad2dfe30081f634ebaf
1840 F20110318_AABYKK mabiso_a_Page_047.txt
142365b3cfa63420ad59b718f5e1a417
5a07fbb1bb07d1a49a95e8570b3322cc7131d56c
1532 F20110318_AABYJW mabiso_a_Page_032.txt
c9e65e368cddd00bee310f06f2a62fbf
3b4a5ecf4e944152463e607e18e6322bf1496545
2057 F20110318_AABYKL mabiso_a_Page_048.txt
94407a78dab652ab5dab83df840825f2
bcf031fd45eb7ba13d592e029b4b5966a2db8b39
1320 F20110318_AABYJX mabiso_a_Page_033.txt
db62a0ffb1a20b2b2cf6f76a08ec08fd
e6f1d8e5164033eacad2e194ddf9a46e1b304070
1836 F20110318_AABYLA mabiso_a_Page_064.txt
77a5cb5bdeeb91b8811624215e6771fe
f74dbf1b2aaf7a2f56735e49de5784833e0227bb
2006 F20110318_AABYKM mabiso_a_Page_049.txt
0f611e1e31c2b3b27cd903359e32d80f
fb1dd4988e24924f28f771c2fff805ff982ad215
1772 F20110318_AABYJY mabiso_a_Page_034.txt
b64135bc3e1422604ab5704e7a919984
0728a6da3f3226a911ad53295f0493f01e17a37d
1892 F20110318_AABYLB mabiso_a_Page_066.txt
89806c6db9c1b273136f446480d3702d
6b762948d4c4f8355c5cf40ac1a32cdb581cec4f
2063 F20110318_AABYKN mabiso_a_Page_050.txt
7cdb6fb6a9cc0c66af7dcb096a681e7a
3c38933c018e5386e24a107150f92c31b688dcf8
1966 F20110318_AABYJZ mabiso_a_Page_035.txt
f6c338c2a7a34e270a5a73c5e2dc6848
4fa48e2ec5764aa0eb9d34ad83abec831dadc988
1456 F20110318_AABYLC mabiso_a_Page_067.txt
538849cd2c92742b950b465617abcf19
c3c3aa547d2655d83db487dba1780898a236b593
156 F20110318_AABYKO mabiso_a_Page_051.txt
2cb92ebea7a4714f63e602775d631aa4
5678765ed8f7fbe01d1d51f5e538811c14d48685
1838 F20110318_AABYKP mabiso_a_Page_052.txt
16ca3e94fdcfadf55cb955f0e71f79f8
794cf0468a4ff02a4b49ad32b6a16a724ff32c86
2357 F20110318_AABYLD mabiso_a_Page_068.txt
f504f92f7503fa58c05f8ac84582ec21
ac47922df7824042307d64af25f69e5af30b103d
1640 F20110318_AABYKQ mabiso_a_Page_053.txt
e0c267055857efe010132b28ba1feb58
ca93f153087ca09f34d895232e383e9dd4f48c80
2247 F20110318_AABYLE mabiso_a_Page_069.txt
db1d617d35b8df8d8b972695e4945820
aa09c464aa444cab182f4c217fc0b8eafb7be274
1656 F20110318_AABYKR mabiso_a_Page_054.txt
2bfcc0232808d82e6e043423e5104959
d77697d7dac2c33e0bd029d69cf187726f1ebb58
2821 F20110318_AABYLF mabiso_a_Page_070.txt
27cae5f5ca391e7412bd646ef39f7dcf
b96bb32353bc95686c666ede8c287113bc92553e
1895 F20110318_AABYKS mabiso_a_Page_055.txt
075bcb067cf9d635d80d21f9f441543c
ae9be1a1e13f54aefae297eb642bc8511fe0dab9
2944 F20110318_AABYLG mabiso_a_Page_071.txt
cfee0b1d7ee36cf7e71d5015d90a0ae4
7aafc94c7b37c93e17de794bde77d82907e3222e
1475 F20110318_AABYKT mabiso_a_Page_056.txt
a238d36ba5b55f7007a4bbbf6f48a173
77c0849cbb8486a66b96b8d000b09613236cf94b
2850 F20110318_AABYLH mabiso_a_Page_072.txt
b03a362e5e850f636c2b23738b002a22
385505bb4b1aaf0389f0cccc401907fcd71922fb
1750 F20110318_AABYKU mabiso_a_Page_057.txt
6efe255390609fa9ef5b9cf87f2634ff
6c1c5d83f8d58f6c14f81f0a6cb955d0019be6e2
627 F20110318_AABYLI mabiso_a_Page_074.txt
de76fa1173fe167d216c4b987c557712
8a7f6a95e22ad6e7120d7f7a069956b9dc98e5ef
1695 F20110318_AABYKV mabiso_a_Page_058.txt
d21280f2c661fc3ac0086b64f7c9e8db
632085ef471bcd47fccc6e2cb97831ad7f686925
1697 F20110318_AABYLJ mabiso_a_Page_075.txt
387ee482ced4018e4d0c9a1adbb266d8
2d2a40fd525c40a6f7f69bb3cdbb916a88e42707
339 F20110318_AABYKW mabiso_a_Page_059.txt
7001806d73e99beca8460ad3769e02ad
3ab4ccb807537d00e1ca01fab8f7819d4abb15ca
F20110318_AABYLK mabiso_a_Page_076.txt
9908c07adc8b66306ca5a3d0b3fb7435
66b1b1e7ac5c22d2a42d4bd96f938e10d66e2a36
F20110318_AABYKX mabiso_a_Page_060.txt
af844cedc1dc0735aefe62f36fcad108
77217adfe04047e22651acf6d745cc7a83da190a
1898 F20110318_AABYMA mabiso_a_Page_096.txt
e3538529afc5bb3cba7087711ef25ca7
0e5f896cfc4b5fbacaf9bf0ee104e2f61de67fed
1659 F20110318_AABYLL mabiso_a_Page_077.txt
a5c75c340030cb322eb682da46914467
3c3bb5b568d1f940b600af07602864657a78a4b4
1890 F20110318_AABYKY mabiso_a_Page_062.txt
4bab8c7100351191a0dc2f823da55bd5
0e0ad8daee70a5795beaecede174874bb8e1052f
F20110318_AABYMB mabiso_a_Page_097.txt
d3f4adefb4cfe2e36aee1370117bb5b9
6cd887e82f70981810cba35a9cc2e4e977bc9d27
2024 F20110318_AABYLM mabiso_a_Page_078.txt
137cdf721821e5c88ebe64ffeb4bfb11
784b44323fcb5e868ae2fbfbd05a9ca53bf938b1
1872 F20110318_AABYKZ mabiso_a_Page_063.txt
039f3675f7246ea4043c4d1f76fba8c3
15b70870ce98fdd52aefaf056905d0749ec5fb3d
2396 F20110318_AABYMC mabiso_a_Page_098.txt
03ee52ef5b3183be001a93e346d0e55b
7bb084219c76485d7deb67561692a307e17af6cb
2026 F20110318_AABYLN mabiso_a_Page_079.txt
c4309cbfb8ea73924447ca7dcb7cc383
b8645d8d67a956ca6afe938964fda4b13fe746e5
749 F20110318_AABYMD mabiso_a_Page_099.txt
7bb72a27c8dbd66ecd7896e38779bf9d
7b32214c7497cc406f7f3f40aca40e671cfa035f
1682 F20110318_AABYLO mabiso_a_Page_080.txt
d42cd9e499904855874e595152171d6e
91e15947bc7340d9c312d117a11fa8261eb83e65
1706 F20110318_AABYLP mabiso_a_Page_082.txt
eef3531409f6fa5d3ea5692008baf743
a45316c96dbadbbeb5771b2e76c18368b6ad0d21
2886 F20110318_AABYME mabiso_a_Page_101.txt
dc4c2143298388dab6d66f33c1777fa8
f3126824cea6f9beeb94484a9852dd44700ef316
2916 F20110318_AABYLQ mabiso_a_Page_084.txt
52673671808a97a1b4e77486b4a713cb
9c1a91b199532f2f3ca386fb10ac4ce1b9119c13
2011 F20110318_AABYMF mabiso_a_Page_102.txt
5f6af41d5927e719737adb12ef8ec909
4214a4be40b7322af032e84efd1b2277581f37a4
3106 F20110318_AABYLR mabiso_a_Page_085.txt
45c8798aea564b673103eb4b01eec796
24e34fa98d0d38b8849c03ff5f7864b7f0f8dc35
1790 F20110318_AABYMG mabiso_a_Page_103.txt
abe09999bd63dfa42841c9049585b896
4dce0f8c8bd785c32c3d062e1710ce6ffa8001bd
1967 F20110318_AABYLS mabiso_a_Page_086.txt
b7b974756807009efa6a200f5fab9d17
efc2a68ec650e5598ed78f1beb9961f92035ffee
3116 F20110318_AABYMH mabiso_a_Page_105.txt
2873e6730e219fe40daaceac65b14517
a8107ad9bae4fbcfd63b6b3be0155637972851c0
1998 F20110318_AABYLT mabiso_a_Page_088.txt
2a6eeadd990b585f57bf116f43708a6a
b2fb0241677c794892c5da0222c86a74fdb1bc69
3156 F20110318_AABYMI mabiso_a_Page_106.txt
0dd75bdbaed3f98c19d990d3a1464950
d2b8a1e0ab5d89b843a458d2100519e753b3ffa6
2864 F20110318_AABYLU mabiso_a_Page_089.txt
614a6d95d117e9375a720879e15ac95c
c3d3d1b3fc1cffd77d8a12c6a58fc09e12354b25
2516 F20110318_AABYMJ mabiso_a_Page_107.txt
261c98c4829ff7c484a3355782c80c83
c5c9b38cb8db756a4c8536e32951b01907a2ea3c
993 F20110318_AABYLV mabiso_a_Page_090.txt
72bf874a1a0e2125f97c504523629a40
396d0d349297131b671e913389f174a1ae532a61
3066 F20110318_AABYMK mabiso_a_Page_108.txt
8a0612977e3188bcd42537c8980b8bf8
3c63a832486657be73bd09d565a41cd3c0f189f8
1662 F20110318_AABYLW mabiso_a_Page_091.txt
eb9983ebfa488cffb270282ae00df1d6
32ec9408805462777569d438af24f8057e8b8a52
3330 F20110318_AABYNA mabiso_a_Page_126.txt
fdd10f0025902eb68bfe14dd3ca55f37
cabf49dbf89930a016c32f2860ec983eaa5cd937
3304 F20110318_AABYML mabiso_a_Page_109.txt
117f3c0eafdaa2329421cb0450f9b5c7
c83c99ebc30f28d5c3d64aece60a44dda2227bd2
1974 F20110318_AABYLX mabiso_a_Page_092.txt
3944e3e58e1965a276c668b1e50a1361
08ff2e2e421af53d3f03858960e20983125bb765
2895 F20110318_AABYNB mabiso_a_Page_127.txt
b3a03e2716dd0a3e60ce9882ad1a5d2f
2267e50bd83045176817d83560d8311853ccd318
2485 F20110318_AABYMM mabiso_a_Page_110.txt
305856aa502126b87f0a08061c010039
eea2158afd2512b3a21d0052aa456d36ecd1efb7
F20110318_AABYLY mabiso_a_Page_094.txt
0d5272abf62232e33cb87f2924bfa7f5
fe1999f4919c3c94d3acaaba506da08e42d3b101
68 F20110318_AABYMN mabiso_a_Page_111.txt
1625ee9990f4758d3d47619fdf5ae152
a41fbc9e869b57d0bb51589cd0004a53f7689026
2003 F20110318_AABYLZ mabiso_a_Page_095.txt
51ca9b6e380d96ad467087459bfaab87
06240fdf1c544fda9238fd261f1857a56edadd75
2142 F20110318_AABYNC mabiso_a_Page_128.txt
560ebca432d8f07f49c9953327afb4e4
96bebe0bc4e323a17616397e94e12d1757c64fe4
1817 F20110318_AABYMO mabiso_a_Page_112.txt
f511fd79abfea7100424c7475d3f24fd
4a2f4ce569e0ce2c290135308e38fa71d9162551
1459 F20110318_AABYND mabiso_a_Page_129.txt
dcbd79e1b7d3310a8eec42307dd00816
b85133392007b8b82e96edf813d7f2c5b8443f63
2663 F20110318_AABYMP mabiso_a_Page_113.txt
84da683a66d39f28f952ff3238b30853
68a5c8e3b82ea39a315d43afb202af6cd593dc75
2098 F20110318_AABYNE mabiso_a_Page_130.txt
3cca7b46b4b9e24c17f74e015fde1d65
9b02a49f9e2b4f99ecb7ac663bb003ed0965c590
3168 F20110318_AABYMQ mabiso_a_Page_114.txt
feb8fee3c8edad01df992f3ae3b2f83d
126c2a1910585d28e17a10f53dc340c71c3dc736
2529 F20110318_AABYMR mabiso_a_Page_115.txt
3d3a90ce05d782947c57d5a86114f31b
9aa51c344c31c834f817bbc0006a7fca3e1c86ca
1242 F20110318_AABYNF mabiso_a_Page_131.txt
04164ea0f12b54ebfaf7d179b68e4dac
247dfa2fd21b377bcd0e79f96937feff2ff1497c
2276 F20110318_AABYMS mabiso_a_Page_117.txt
468c9c9938a3529c3446be0a950adaa1
46c7907b3462ae399dbc6b1270ec85a539bb438d
1019 F20110318_AABYNG mabiso_a_Page_132.txt
dc8925a711bd9599076558251e389bd6
940236a094dbb56f119e89c0a2f525e0b507530a
1380 F20110318_AABYMT mabiso_a_Page_119.txt
b3e4e043bfbf9d14e3f7e9644aaa37d4
4adfd76d487095e3d1e91df3a833e82dd20ebda1
1037 F20110318_AABYNH mabiso_a_Page_133.txt
17b5df9833cfdb587599a40093f9c6af
1b881f07fbd01a057a3aa9271bcd06e66e5f8436
2530 F20110318_AABYMU mabiso_a_Page_120.txt
3be1ac50b1b218a04ad0a8a83664a91e
f1c0480943b0457e0b6881c603c5861b91f0c483
990 F20110318_AABYNI mabiso_a_Page_134.txt
40cb9cf97acce42d788cf0e87d045c1b
1f7d60be3c17a25ac0770d75dc11db5c11a26112
962 F20110318_AABYMV mabiso_a_Page_121.txt
0185eb7f7b75f37569324458e1766adf
ecf84bfed02bd75a6a001580af20e7e88c78eb7a
539 F20110318_AABYNJ mabiso_a_Page_135.txt
d2025f73bf1c057fbaf13972c51c8698
303e18a1ac6ae64dff51ca6a02fdbfc1fff45fb7
878 F20110318_AABYMW mabiso_a_Page_122.txt
d0ecf88a41e2eaa026e178c3ab5e27ee
2372c7ec4bc7307dbe6525f44d6549d203994678
2148 F20110318_AABYNK mabiso_a_Page_136.txt
d38c2f79a4a9f1a731faa7686ddaeea9
3af8e5a55a2d0371e5fca77fd999cddf4b249b28
1795 F20110318_AABYMX mabiso_a_Page_123.txt
db7d48ffb0861bea9af29ca727c5ce4f
72966d45dd8a28811f267eff8469eb04c4ad3ca4
41085 F20110318_AABYOA mabiso_a_Page_013.pro
070507d5c5b7ccf1a2bb210b0875e04d
bf3cfadbe21d73a96ca028c8cdea9ac1ae9f8a38
2547 F20110318_AABYNL mabiso_a_Page_137.txt
fa58cdfd35439b7f380e3e6bf3a5e79a
266359ed9e967b46841c12b13e77669d7ff568cb
1689 F20110318_AABYMY mabiso_a_Page_124.txt
1a7c7dff5556da52cd3c78988586a93a
ecf0ed5865a4f742143d9386e3a56bbdd4168c72
49727 F20110318_AABYOB mabiso_a_Page_014.pro
a921286b2934f52972b59cd0e1bdd9c5
b4c09385f7cdeb7524de674e3725cfbf4be38913
2558 F20110318_AABYNM mabiso_a_Page_138.txt
4b7cab715f9db1e595382ab47ecfb644
d17014ae297ae209c1a71789fb7a8486d98086e2
126 F20110318_AABYMZ mabiso_a_Page_125.txt
5cb69169dbc595b253e45050870404f4
a56c50900207042c18e4bfc8b451013c5b9eece8
48504 F20110318_AABYOC mabiso_a_Page_015.pro
efbff7351c4ad18a0d252b44149d2771
56fe2de24906e38c916f40ae52702279b91e2bb1
2018 F20110318_AABYNN mabiso_a_Page_140.txt
5f6a5046d7b561ae5d03ad8e3976d8e4
0ffbf7ab5e573e26879edd0b7c81630b1c8f73e7
47050 F20110318_AABYOD mabiso_a_Page_016.pro
ee90693c0898a15920fd5de44eef3354
08d5de337e78e3c2c9285fe36aa712c8b3444f29
503 F20110318_AABYNO mabiso_a_Page_141.txt
81a02a8514814b0e475e24d0ca4c3e33
97a9ed0ab3b475e1fdf0e85e03bd6ea8108c51f4
50359 F20110318_AABYOE mabiso_a_Page_017.pro
447eaea7334674bafafc65e414e76c22
5b43d395a3924793a90566f45bcfb410e8813add
10211 F20110318_AABYNP mabiso_a_Page_001.pro
3c51066fafca3fe6466808178d377328
22b498f062fffd44ede15bfdc8ec6111b0459d14
44020 F20110318_AABYOF mabiso_a_Page_019.pro
1948a3697420030b1aa52d9bf75d56c8
c00b3b674a8d1947d47e4a669cbfb822acf45ff0
1097 F20110318_AABYNQ mabiso_a_Page_002.pro
14cc09fdb500a217857ba7dc1df4580d
c4cea3fd96d9dc957c0461ce4347b963c32625e9
625 F20110318_AABYNR mabiso_a_Page_003.pro
896ff725fb2eea76b41a0c67c3d4194b
cad2c3696250210a51c5ddc5ba459ba45456784b
50891 F20110318_AABYOG mabiso_a_Page_020.pro
50b8aac287751e7ef5e8dd30446bbb43
ef5098659728a936938d673196e04e969f29b134
40046 F20110318_AABYNS mabiso_a_Page_004.pro
4a217f8c4bc431ef3611db1f2904fda3
d69b570b2337da7057f492f1ca96bb268b52a7ef
49821 F20110318_AABYOH mabiso_a_Page_021.pro
1a571be76dc06150b855a2afb8078ece
6d454a0b85717c414d25b4eaf05037fcbac44751
83516 F20110318_AABYNT mabiso_a_Page_006.pro
b14815db323552bf6bfb0a470b831e0f
3600b4480a93b21bc331626de6f10c3b83da3ca4
50631 F20110318_AABYOI mabiso_a_Page_022.pro
7a20902eac57da87038848156b470280
3519e3c3c9ed693a5e8c7ed6e6905560d431d6d3
6376 F20110318_AABYNU mabiso_a_Page_007.pro
ce18087a743a31617755840e97635502
621626beb709967a651f841921cb1a7c4725a2a7
48754 F20110318_AABYOJ mabiso_a_Page_023.pro
65d23b546f507073af1530569cafcb12
b1a5e3028b6816b7669346c418315e9fba74f77f
54396 F20110318_AABYNV mabiso_a_Page_008.pro
27f1f27bc2aedb7b1508c2d0169e599e
0b59cf5cd3b9a8dee4a902da77fb971398804da4
48856 F20110318_AABYOK mabiso_a_Page_024.pro
d3c3d46a307578b1ed846a53cb7dfe76
b737ce7944946d6274d50250deb481c0232afb9b
50773 F20110318_AABYNW mabiso_a_Page_009.pro
35ba116b50a8fb9b8bb9e6ebffb1b4f6
9d560f17c9554efe2fb5a8e0a9ea59cbbd6b8dd5
43365 F20110318_AABYPA mabiso_a_Page_040.pro
3403d2b11783f8e8ca6c290c06218ff8
c49ff02c5eef7330aefae2763d36db85c6691066
51489 F20110318_AABYOL mabiso_a_Page_025.pro
bdeec83f0d323f05f44ff7f959b67e8e
468339fb3e35af634f55b42a33e79ee02182db50
36323 F20110318_AABYNX mabiso_a_Page_010.pro
3650a654fb9ef63cb7e2ded88fb309ea
b3c005a22973f1de12859f876403319e5568aa7c
50808 F20110318_AABYPB mabiso_a_Page_042.pro
70116f9a9e5820d701494dfcc86503a3
4ffda65346c721d02394cdfe96cfe3f2ba0dca61
47284 F20110318_AABYOM mabiso_a_Page_026.pro
0f297434878bf89a27c3111af51647c1
b7154ca64522ebcb2fc71b3952e55c4f4f470824
41474 F20110318_AABYNY mabiso_a_Page_011.pro
c437470a500180d5ebda5d413156329a
fa5fa9ce3700c4106c6644a6d60c543597008d7d
45433 F20110318_AABYPC mabiso_a_Page_043.pro
89d74fc3e35891d0547699810de38e5b
6909894ccbb67b21dc24fb8b58eafc06396f15a6
49515 F20110318_AABYON mabiso_a_Page_027.pro
c43279b35983ccca492314f0d6449ca4
32e15e73cdb1829f8edc4e6285d4408284c666db
38997 F20110318_AABYNZ mabiso_a_Page_012.pro
c49e2ab919b86880a21226bbb1741ff5
5556f4e64e8c5a0cf1240696795aa55806e89297
39837 F20110318_AABYPD mabiso_a_Page_044.pro
e8a1191b7c379d0ed6be9cd5f8916783
8cab50ee85adc833ab3319d0641bb924acadcbe5
48781 F20110318_AABYOO mabiso_a_Page_028.pro
3be6c02b4dbdce88161b4838b38f2fbf
fa7b3188409f9b599e9e9179270325b1c8ff313b
46014 F20110318_AABYPE mabiso_a_Page_045.pro
1719f3e7e86eb29404f00ac57384ece7
b36a830ee19fc5df6a466c55e778b5cc29e19853
39881 F20110318_AABYOP mabiso_a_Page_029.pro
5e537cc8d4b2e22fcde7214284f8f188
372dc37229211856b966e2854be4128002735801
36010 F20110318_AABYPF mabiso_a_Page_046.pro
ef69905e6a47c96761bd9d8c0976a1f1
8b39d216a825e8dfe931b7754f2022d014cf7850
28686 F20110318_AABYOQ mabiso_a_Page_030.pro
a05cb46c1797d0a9f2c0a05f9d3ebdf8
f4293ae10045e4e94a85c7fa28f905807c0eadbb
45188 F20110318_AABYPG mabiso_a_Page_047.pro
943365b938557189f69d6690f83fe0e2
b9aa8219b65961437c3f2b9e17e9bf0a0f7fd876
23994 F20110318_AABYOR mabiso_a_Page_031.pro
6a83958acc73fa0218a81afc7e75104e
98ddb611e7a3d322327fc3a00abc51cd05277986
34808 F20110318_AABYOS mabiso_a_Page_032.pro
b8a4d7447ea57b1e8e94fd4e7b0e9510
2e59ff7dde594ee8b6ed4d44b53c757074a62915
52281 F20110318_AABYPH mabiso_a_Page_048.pro
8e0b3e795dafebac62f45815c38aca94
b6fd04740959f80e9a02f440dacc77fdd73d979d
25350 F20110318_AABYOT mabiso_a_Page_033.pro
05c3245ee7c5a6adeda5316f9eaa0054
e35493a95fdfc5f3e77ee514671d0a20a3ac02c0
50818 F20110318_AABYPI mabiso_a_Page_049.pro
e6744dabc627ad219e742a99b66e4c06
823478a7edefff15c7eae73562e8bf2ee13d769c
40562 F20110318_AABYOU mabiso_a_Page_034.pro
463f408f1d1942375fb106a3faa0b1b1
edf2c97af8183bc6045515f86a12407ba1b7c5f3
52275 F20110318_AABYPJ mabiso_a_Page_050.pro
a1e091e2673b5a0e8f95d4b543619a36
35cf691630119b170925fcca09ce5af935d8dd48
49248 F20110318_AABYOV mabiso_a_Page_035.pro
7293093a21e65401cc6f2104e89a9b93
839f9cb31e8e1ab349f7686917d3b6a423a7ce87
2806 F20110318_AABYPK mabiso_a_Page_051.pro
90137bed6df4a58960edddbb8528f686
7747a5f9ae990cfa2ecb0abf714171f36adfa990
52315 F20110318_AABYOW mabiso_a_Page_036.pro
fd006ff2615d661744721847d01ad7e3
4c90b51442544b22491a8504ae20cc42536c1e66
45029 F20110318_AABYPL mabiso_a_Page_052.pro
c0fcbf9d1b79ca59891ec470c90ae4c9
8c542f37b9e65b13a4a915aff8ff46e12f4e7e6d
51537 F20110318_AABYOX mabiso_a_Page_037.pro
9c982a0186710527d0485deed2bfd36f
f53a108edffe8777f53c6b58f9d7aed02145b9af
44147 F20110318_AABYQA mabiso_a_Page_068.pro
95087a48589522011e013892a4a1eb6f
4a5642cbbdcf6655b1f21557dda39c941f3f9f7d
38045 F20110318_AABYPM mabiso_a_Page_053.pro
07f68d8a7813201050d76d0a43129f48
91fc0cedd515306c22cdd91e75011531bb6eb3cf
50913 F20110318_AABYOY mabiso_a_Page_038.pro
6df3a59c0faaf478c155677bd7634e8f
724e82ebb2093ee1a1a6f85dc8f851ee31f5b0d8
42536 F20110318_AABYQB mabiso_a_Page_069.pro
3551c0334d8044b132af68bff9cf2028
5a8e750b9a72bdae585883e65c19444f7f59dcc7
36475 F20110318_AABYPN mabiso_a_Page_054.pro
5a65f430027b006d8ebb998542053c67
746664563cde8c2968588a1a897ef5f6ab18ffad
21743 F20110318_AABYOZ mabiso_a_Page_039.pro
8772ee4ab6d38091d9291ca9e242a741
18c28d3d679f16f155004923bcd8d1b277d090c3
62575 F20110318_AABYQC mabiso_a_Page_070.pro
ff018a620c7b37cbb08195c049ea2d5b
269ec58d51ac503415a4c1e42d9e77442f5f09c9
43998 F20110318_AABYPO mabiso_a_Page_055.pro
dc3a635b9983381208f7ae7d8139a467
8be91ac36ba1eb53e78a1aab061ebe59cd01b56a
66900 F20110318_AABYQD mabiso_a_Page_071.pro
a53d746505cba057d995badd794ee8b1
0ed0dd5944b731abc3ddf647b44fb943074051c7
33239 F20110318_AABYPP mabiso_a_Page_056.pro
19cd723d9f977ff70fc7ed6c58e16eac
c4e428b3e00660fa1e306f04872c69076de1607f
63559 F20110318_AABYQE mabiso_a_Page_072.pro
3ebdd60e5bef5d3b4edb9915a3a5d5f1
92b604f034c6799bf0c1534872e9da92ecb739ab
37621 F20110318_AABYPQ mabiso_a_Page_057.pro
7d41eff1b73e15125a3d3461aba18592
2ee70c7f2876e20452522ead9ad3e6fe47f18d49
64937 F20110318_AABYQF mabiso_a_Page_073.pro
8bb1ff4ac8c11fdaf3a6272065f24c2c
1b4464ed1f0ccbfca7790cc2978748ba3b525e9e
41683 F20110318_AABYPR mabiso_a_Page_058.pro
83b34f02d0e5fb26927d44304175ae10
2e1162e70747ba85e2f2ed17122b72b3b0a21a4c
14607 F20110318_AABYQG mabiso_a_Page_074.pro
59b633b021d86a1a0d34639905c9ae01
2b0cd976cda1ebad4a7e4907758ab6e0f602923c
7464 F20110318_AABYPS mabiso_a_Page_059.pro
da07fb016711d4ed638c26bbb6f47d8f
c671e6a087aa9ec8ef4ea0db5d1d2b791976d3e9
41143 F20110318_AABYQH mabiso_a_Page_075.pro
7cd66aa07e2465bc36ab1ea3a2680b4a
1a8df030281d21c7347f32a0fc146ed210abef8d
31014 F20110318_AABYPT mabiso_a_Page_060.pro
7940d68a9fc6d3c3f84f4afd1753a0e8
b74e57a2fbd67834110c2196e4a55a6b4be818ba
45760 F20110318_AABYPU mabiso_a_Page_061.pro
895061b3753a8ba3d616e1a11306e07d
b9c1c436f1fec8dc08ea439d5597dd831f414bc1
45679 F20110318_AABYQI mabiso_a_Page_076.pro
0c61575ef82e6f9264d996e105af40e4
5a1ba0fc1b567b14125afe1fd6d36d4406dc5b66
47869 F20110318_AABYPV mabiso_a_Page_062.pro
3a6f076f71dc21af22b7f6644700b0be
c641fb17dac480f0fbc3254be2a9e10e8a9c54ac
36734 F20110318_AABYQJ mabiso_a_Page_077.pro
674ca4dab8edb359906e62b1ab9e9674
9406f1fb2437fd9454a3de5fc69ce839cb3732b2
47227 F20110318_AABYPW mabiso_a_Page_063.pro
a9604c47d8c793e221ce6334786d6187
83ebc8fbdd5a59034275f580b22dba951c987a27
45660 F20110318_AABYQK mabiso_a_Page_078.pro
7c2978a0e2145fa189c501e735bdaa3b
c29f9f3d59e22660b39424c7839d5d4d2bf404c7
46564 F20110318_AABYPX mabiso_a_Page_065.pro
80dbf11fef35d9679d0bb8cead4a5846
973dac6a69b76d86fdf282271a9f63bca323cc44
49339 F20110318_AABYRA mabiso_a_Page_094.pro
ee62c585878246c6d6b5955ce6a44ae1
2680253699dab041916a8055cb1df8a025584693
43725 F20110318_AABYQL mabiso_a_Page_079.pro
ed15150f5d0fec40036f129db787648b
2ac67056c9eb65ac6c38cb55bcefe5dc33007da2
47033 F20110318_AABYPY mabiso_a_Page_066.pro
0befe659033b04ac26bdccd08ec8adf7
289030cc1a2fec08a9bd7c8d16556b4d66e60f66
50923 F20110318_AABYRB mabiso_a_Page_095.pro
e709e78d71ad30a19cb27223c3ce9432
2de2fc1c91cd95d60ed5a928d76f222f701a58bf
37459 F20110318_AABYQM mabiso_a_Page_080.pro
0df9d9e68b1b30bcffdda22ea5764329
42fe1e9c8fe17be23296701645e2aeb45b8d5143
36358 F20110318_AABYPZ mabiso_a_Page_067.pro
3f98464f959cc1523ff0e9f4de6ad58c
b135c05da198d944b55c722c0265f312952d5e1b
47934 F20110318_AABYRC mabiso_a_Page_096.pro
bbc5fc0a4922baebafd634a970d11331
e50a0904311c804a2cb3e1317244c36c46c09e81
34668 F20110318_AABYQN mabiso_a_Page_081.pro
e12e6bb5ac5d3c1a7f0e2613064a59e1
a25e3c6e2718a4f7d34ac1a7fd28a9a72007b71b
51611 F20110318_AABYRD mabiso_a_Page_097.pro
8cabe02a36a507c612b0fa51437068d9
19d9bcb2d36ffdeba46d80488575300515f989e6
40836 F20110318_AABYQO mabiso_a_Page_082.pro
2a7c320413a96767323aba481b6644d6
a24979634449a69e2b56b2bdebb9ee19531beb35
49392 F20110318_AABYRE mabiso_a_Page_098.pro
21736c49b6eff3f99e9b4405f12a9494
9c7973bb20e298c315d1b77c4d4737ea5cf40b88
87749 F20110318_AABYQP mabiso_a_Page_083.pro
18b827b1b5409031ca50cdf5f9c2117d
f23b9862fe6a0477e883c994abc6e173c9dd29c2
18685 F20110318_AABYRF mabiso_a_Page_099.pro
15c0321451d014b237399834813ea370
2ecc344f83f82aa791c957538196288f39ba850a
63483 F20110318_AABYQQ mabiso_a_Page_084.pro
461fb0380da96400096809e8d6fb1a92
9dc6d2e352266c64fef76b1e94e367c60a8b392a
55613 F20110318_AABYRG mabiso_a_Page_100.pro
8d0b46ad398ffa547ca61721138083c7
42606ed0b1de87f29dd7ec8d97db1be2d16b3c76
69374 F20110318_AABYQR mabiso_a_Page_085.pro
93e28c99ca6010c40fc34d9c437762e0
33effc308e3d127d349e5ac60882810beb96b681
71118 F20110318_AABYRH mabiso_a_Page_101.pro
b005bc1f833d2ebf1415d97c1b50e10b
c0b5c603a5f5ea223813143dcfe276fc417ecf1b
49996 F20110318_AABYQS mabiso_a_Page_086.pro
bd30b3c23a328396e7486b0db1507141
45c9587473b91df381e239a9a59aae4fc5c3e313
47340 F20110318_AABYRI mabiso_a_Page_102.pro
2e18ab1135c29d54c0598eb6e9d4fbe6
6f2f0dbc1c0760af9ee796565fca4999c0cc04c8
67758 F20110318_AABYQT mabiso_a_Page_087.pro
7d4db5e7b756e317d425b92290dc6f1b
90d6b00662a27fb1b52a333e270241a5d43195fe
50664 F20110318_AABYQU mabiso_a_Page_088.pro
624c0e4ad0f12bfa13407fd4228df06e
68c228fb72935adaf7cfc0f1d6b8639954d8c1d0
44629 F20110318_AABYRJ mabiso_a_Page_103.pro
b86fe9eb0125b94c207c27b02a8727b7
933a63e6c2983ed610ca6bb71c0ded30c36e018a
60649 F20110318_AABYQV mabiso_a_Page_089.pro
372816ef34cfc9ccb605915f63c714c1
751e8f8db9b6b16d3622aaf22d95cc6d5fcbe9df
71601 F20110318_AABYRK mabiso_a_Page_104.pro
e4b079641b35aae028bac86d9fc66ec0
d0ef47a6b9ab0d89676a32ac8acc41925827a502
24878 F20110318_AABYQW mabiso_a_Page_090.pro
4c5d5f63e440f0d08ff7b5debd4deae4
b7434a96bd0f759d7996a2a15f5a17f05ca1d23b
78705 F20110318_AABYRL mabiso_a_Page_105.pro
112275f3ef53c8bed6f8cf2263de5faa
976c50ab385d8aa6b3b8e77fb0901a6a2da94397
39371 F20110318_AABYQX mabiso_a_Page_091.pro
a06c469936837b75d3fdd4b3722c0b5a
4ceef580805054a01e9552941ef308ab68086ca6
39761 F20110318_AABYSA mabiso_a_Page_124.pro
4932379c720b2af139628ea34e564da5
47e90103ab18104577f052127b237fe9302ccb1e
78482 F20110318_AABYRM mabiso_a_Page_106.pro
42535883bae53242dbe179e84fc44b80
a0ea7039ab98e16cff532664f366215396dd6a1b
49815 F20110318_AABYQY mabiso_a_Page_092.pro
03b3f141b943aee86b9b6ec1a7a8bee6
4a9cf3e45bc6ec2059aa9455f060cd4c3f938dc6
1992 F20110318_AABYSB mabiso_a_Page_125.pro
5cc722c8e1eb484b645c23b9c9fc4df2
5f4dcb896f8aab00985f56971194553d792327db
64237 F20110318_AABYRN mabiso_a_Page_107.pro
d4969524475bdc04c465c256aa6e557a
cbb67e8a13e7c30d5308a6e8109046cd599a816d
51546 F20110318_AABYQZ mabiso_a_Page_093.pro
7db1f74a9a251f300b0233eea25750be
a494f655b8c3c7b1c1e73996d047f09b9b7cc309
79646 F20110318_AABYSC mabiso_a_Page_126.pro
33888613f92dd0a6139e53b965d103af
53d722ba291729415e700862e4b9c8ebb43a64d4
74543 F20110318_AABYRO mabiso_a_Page_108.pro
d60467628e22d65fb7d3ba0ec3931226
be031831e78c7f25e8d3376ffda27057b3a4ccfe
48024 F20110318_AABYSD mabiso_a_Page_127.pro
ab3b40bebc7fa0a4568e7526135d4852
7247747b4ebecd9704a4d50978dcbce5b91271b4
85095 F20110318_AABYRP mabiso_a_Page_109.pro
b01d16435c765956ab855b31ad8c8fe3
bb7b57e764bf50563f4766a528b2e7158d98b648
40076 F20110318_AABYSE mabiso_a_Page_128.pro
020fb0bbf15e72bddd46136e6ee1d1eb
88383457a2bce48106a2f27d46e7727c5615422f
40514 F20110318_AABYRQ mabiso_a_Page_112.pro
eb953ce0a853c89330413c4b4667c61e
069ed6eb5ab00db850e95521d12b32df588f6941
26124 F20110318_AABYSF mabiso_a_Page_129.pro
758f5294bfb3d8ec3325219a04f5c04e
615e04f14c7c5814b0f8dab1ce31dbcc729cda80
65331 F20110318_AABYRR mabiso_a_Page_114.pro
6961c624edc8de98a811a6e3b71d1686
34751de1cbc06c47201761243c6fc7252ddae9c7
38760 F20110318_AABYSG mabiso_a_Page_130.pro
1a660d515e18c46ccda0aaf95b5e066d
401556a25829959e55e87a6389d5d63086f3b469
56126 F20110318_AABYRS mabiso_a_Page_115.pro
aac15a03541bd1f94200f2c720db1b72
4f65ded153acad971be4eaa4b344adf8fdd5d5ce
24601 F20110318_AABYSH mabiso_a_Page_131.pro
256ef58ea4c88f31a679378dd4c30813
538cf6a6675197cfec7e92e09d521c947f85ebaf
18809 F20110318_AABYRT mabiso_a_Page_116.pro
2a235630ffc5b1efc2081def92af72d2
e4c143cf4af982d8f877437ad2d88215f04400ed
19338 F20110318_AABYSI mabiso_a_Page_132.pro
d278bea89199478ddbcac151e91be457
c05cb00d811e511eb152a840635c9ff9de824833
30528 F20110318_AABYRU mabiso_a_Page_118.pro
6a2d2fbd65c6c21cb4d956806519d299
290607ab58acd9fae13af380005567a1dcd0afb6
20146 F20110318_AABYSJ mabiso_a_Page_133.pro
43b816a6e5c4015dd598740b81020165
e83bc0ac1dabae7628ea23cc36af995b64bda134
28462 F20110318_AABYRV mabiso_a_Page_119.pro
3ec95e4cb0dfd9a38a71ede3465a09e0
b4f1aaf4e9a77d38ccb9b77cd3b7c5a1b4b77850
58571 F20110318_AABYRW mabiso_a_Page_120.pro
9fefab542c7d75db9607e7f74592c716
8491a39ca68aa8fee32ee60ef7d073eb6d91b42f
9100 F20110318_AABYSK mabiso_a_Page_135.pro
1ba73d33aee5cbd6c34ee9c879729f82
617fccbd1fefdd7d299501ff30e65df76acdcb1a
21376 F20110318_AABYRX mabiso_a_Page_121.pro
3fc1016dfac5daa55ee075f1fd928e19
e690a789df82c10cb761d624f777b56d10264ab5
93431 F20110318_AABYTA mabiso_a_Page_006.jpg
0d9e852eaca77bb04a5bba18fb194341
7abbabf23809ba14c183db2e39f4d819f2d7c1e7
52671 F20110318_AABYSL mabiso_a_Page_136.pro
b052dc639c569e55593c395e9c7036aa
9c70bd31f823936cfc13439ef8d68651013ac9d1
21173 F20110318_AABYRY mabiso_a_Page_122.pro
c5a6ebf6c91932c23d31782d756dad42
e5f6317e6d734e63d2887b96c399be2cfd6c7018
23236 F20110318_AABYTB mabiso_a_Page_006.QC.jpg
5316e09b56e9c375c29815d292dc4ec5
93e077b4566d97e7a4956ea1a6a443c6d4ced2a2
62960 F20110318_AABYSM mabiso_a_Page_137.pro
71c8a6ae9cb915774684b6914c9a7d90
223d3e63c9691661562bbd8f9f08d4c3201dc7bf
42072 F20110318_AABYRZ mabiso_a_Page_123.pro
a0f516747d392668b654a35a9b63f3d9
6b5fe95e83f2478419762bbd66799f1d4765818c
10047 F20110318_AABYTC mabiso_a_Page_007.jpg
bf03fcb5500361d35078f195e05ba331
0279db3d94f625daa1367d666a230c77755adf98
63224 F20110318_AABYSN mabiso_a_Page_138.pro
6f0c0a29174ee725a487a00ec88ffd09
0a43cc97271cf5e71dd27d570815d82cafe876a3
2960 F20110318_AABYTD mabiso_a_Page_007.QC.jpg
538e99359195cee045793a0193d04ddb
1c6b1d3e7aa4aa765510224a8774f207d89670fa
62419 F20110318_AABYSO mabiso_a_Page_139.pro
c6defc1a7dc49b0cb11dce5ca259ba61
5fdb4ea0399c1d9effb6f3ee3870f481e9085193
82853 F20110318_AABYTE mabiso_a_Page_008.jpg
2eebcc0c85fbca833547dff94265cf96
531fc910a28d0c17c9437daed47a40f87729c951
49144 F20110318_AABYSP mabiso_a_Page_140.pro
e4e90c2be69a1c70291aa0dbed8ec2ac
59ab5247fc9178e8e197d747e084de0a3bfa3a67
22724 F20110318_AABYTF mabiso_a_Page_008.QC.jpg
3a9fda8343b3ab57c86b2df074c9364f
8a211f1067ffe98e12988b495d99fe2cd2add323
11464 F20110318_AABYSQ mabiso_a_Page_141.pro
e0ae77442dea41871227e25168c933a3
a49567df86581a4c737b37a3525f2904f23e1eaf
79088 F20110318_AABYTG mabiso_a_Page_009.jpg
841a9004cd989ba55735deed2d54df41
4aab8bbe3c54eb12bec5ff13b47d17a5f67707a4
26285 F20110318_AABYSR mabiso_a_Page_001.jpg
15c89e5ccd28230a534ad61ef233a93d
6a3989004486ac299f71373cb9c4dc9ae956c2d4
21787 F20110318_AABYTH mabiso_a_Page_009.QC.jpg
f15b8a5fc3ca6b196e37c384941c8db3
9e3c8dce579efc991e0ed96a2545b38957268a8a
7846 F20110318_AABYSS mabiso_a_Page_001.QC.jpg
51c9406ae5384a76470a6030c495a1d3
e8cb812517fe1e8c67b4dc0d5453f5380ec3e911
51818 F20110318_AABYTI mabiso_a_Page_010.jpg
e6a4c047a57b13e4f6d02eaf0230a332
d3f826e66a03de57e4a97e89764506dd2830246a
4068 F20110318_AABYST mabiso_a_Page_002.jpg
d677426f0c60dab368d69a88f611a8f3
55113dcd65189fea54c929373afc87e4960b0525
14675 F20110318_AABYTJ mabiso_a_Page_010.QC.jpg
52c098f04fffea014cf94a464e13fca5
a2742bb362474ef26db544929ea08833b12af236
1321 F20110318_AABYSU mabiso_a_Page_002.QC.jpg
613592e45c95d1a1235d50b44390c051
77cbdd16b219f833bdb8acfc49e5a9c6d17973a8
74556 F20110318_AABYTK mabiso_a_Page_011.jpg
84a71e905f4d126502650698696f99dc
5e5c4121f08f3371a7bff41f3780a0f6a920e382
3031 F20110318_AABYSV mabiso_a_Page_003.jpg
cee9e0067b625fa3283264e565c7b3fd
f277bbda5a8043b3402ba6d6a3cef239c4d6f636
883 F20110318_AABYSW mabiso_a_Page_003.QC.jpg
e4df857dc0ad122083c9d5734a712385
a3a13d7a5044f3edd97ac29b5dd67f1c53682f87
88376 F20110318_AABYUA mabiso_a_Page_020.jpg
aa82b6bdb1290fb2bb0066031e716c79
570c689375da2fc61e7f817904072a76d5550793
21965 F20110318_AABYTL mabiso_a_Page_011.QC.jpg
bb6ec2d7c75da6779a27770439fa2109
b852801ed221e69942ccf381a7bb300d897bacf6
69085 F20110318_AABYSX mabiso_a_Page_004.jpg
cad16e22af09144041b30a372956f5d1
ca36f77c50c1bd2627bf4e60e6b1aec69282bafb
27709 F20110318_AABYUB mabiso_a_Page_020.QC.jpg
4c868f5832bf5cf8a31ebc885a9fb948
6ef9bc6a6ef513b0baebff8784041c06a785bb63
66974 F20110318_AABYTM mabiso_a_Page_012.jpg
b8d6874d3ad5bfc226b169464425ccc6
1a0219d73a3380ffdc0ffbbd4e88a84e7cbbe357
21439 F20110318_AABYSY mabiso_a_Page_004.QC.jpg
c334920211b6f5cc616e9e3ad87691d6
25908901c518c1a82f0ee8babd49c65f9ed9adde
85568 F20110318_AABYUC mabiso_a_Page_021.jpg
a99bcb8f1363127110f04a0f893ad05e
2e8551f85681f6d735473e395729a3834d1a81cc
21098 F20110318_AABYTN mabiso_a_Page_012.QC.jpg
adbfd935eecdc314de39786a0d1b25b1
f11e9cd4f6e3008ffffd20b082ba66e3fe7501e2
15562 F20110318_AABYSZ mabiso_a_Page_005.QC.jpg
e1df3f10ac68afe11d9c1483aebb1a48
33ae3107fe0bebe6ba1db4d0d55663ae2bde3d68
88010 F20110318_AABYUD mabiso_a_Page_022.jpg
b5f11794e634604dce719b08bd48f6e3
4a3e59052269e39c0ebef63aa105e2feb80f5aa9
73578 F20110318_AABYTO mabiso_a_Page_013.jpg
1b5819048e02e2d6eb3abe8613b1d323
2df59d927c8b13ad06703d71d19f8a5d6e49d2ed
84608 F20110318_AABYUE mabiso_a_Page_023.jpg
f3991110babc1dc24705f5c8156cbea1
eb168f57c6b9cf79563b52b1b7d133b26de83f3f
85133 F20110318_AABYTP mabiso_a_Page_014.jpg
81940ebc46df5557f824e0e10cd63456
1b79f2a73f654874f312b0f1d326c9e172173bf8
26847 F20110318_AABYUF mabiso_a_Page_023.QC.jpg
94e2d08a18b5245a956b94b0001ed68c
5ad1afe6b6219a5ab9f08586b88ac14d42fe7987
26310 F20110318_AABYTQ mabiso_a_Page_014.QC.jpg
db152b69ec20820473f65f1baa843d43
390c65626ea6d329b2c82847bd1345bc903f77cd
83797 F20110318_AABYUG mabiso_a_Page_024.jpg
5641aa3265747db68721b9a8a184b92d
745ce1b23b2a57082d9f75ba96010835d732ae60
26119 F20110318_AABYTR mabiso_a_Page_015.QC.jpg
9bf949eb61c8dfeb0d8cff9bd2b42ec0
c4d520e42149523bed15da379f06b4b61f1e9775
122455 F20110318_AABZAA mabiso_a_Page_105.jpg
4d0ebb628447893e75e3494ffe288073
d176d6fc51f3be398c46937bad40b2ff5ef41e6e
25675 F20110318_AABYUH mabiso_a_Page_024.QC.jpg
3bdaae546ca199d31df4f8c90fc78c2c
3c5a187f383557385d5d48a7bf466ffa40e37401
81260 F20110318_AABYTS mabiso_a_Page_016.jpg
5f9b71ad9eade2bf91c1c69450466ba1
9416104b07af02f073ab3370db5c363a31a101d6
29920 F20110318_AABZAB mabiso_a_Page_105.QC.jpg
ebfa0e0554baff6f362794f10816f991
a2fee84be793b8ba284dc06fa469ac9605bdb38c
90445 F20110318_AABYUI mabiso_a_Page_025.jpg
24b0f9f8831eac6dd40447abacea3782
6c25aeba5f4457b5b3a93dcd43002546438057ed
24448 F20110318_AABYTT mabiso_a_Page_016.QC.jpg
b028d5389852287eab861467429a00ae
3beac892645b15ca588b73ae22d032e662e75cf0
122325 F20110318_AABZAC mabiso_a_Page_106.jpg
84dfd8c8afbe8a86f4f16e5beeef978b
87136d7d7c9740020b2aa31ae71243dc597d2646
28354 F20110318_AABYUJ mabiso_a_Page_025.QC.jpg
e69717c3ec09b464856b5ed3cf070b3f
2e94bcb5bf3f03c81d9f2ef32b228277232ce95b
83172 F20110318_AABYTU mabiso_a_Page_017.jpg
7cac6e859c6d4c226d5e6689bc804b3b
82bb0b31bb513c0a461b7932972090c70c454d7b
30883 F20110318_AABZAD mabiso_a_Page_106.QC.jpg
8cf1cb8547cb972f7c7721384fc824e7
015402e9dfb0d9f60561641db1e30f694ff0b31c
81486 F20110318_AABYUK mabiso_a_Page_026.jpg
65fe6b7aa78eafa63f567604151dc630
07f3c7753b4baaef68f25557643beff991ad8704
23983 F20110318_AABYTV mabiso_a_Page_017.QC.jpg
5205f3e513ffd9cad0faf87af1934a99
4ac778e20c1ad29028a69fab6ec0dd5f1b7eed5c
104763 F20110318_AABZAE mabiso_a_Page_107.jpg
ee2edcb4db5c48370d2f7a0bb3b9fda3
5fb5e886a0124ba6000f567f9b4675954f17cdd5
25333 F20110318_AABYUL mabiso_a_Page_026.QC.jpg
04ac210e76a6f05446dcca96b3f3195d
c56c00ed48d265c82a0a53105f7db3989f1db076
48603 F20110318_AABYTW mabiso_a_Page_018.jpg
e0815cbfb9404b029002adc5cd2c65b6
7c65f9c4961696250815a19c9c0f8ffd5d5461b6
26743 F20110318_AABZAF mabiso_a_Page_107.QC.jpg
0a5d0b5f507d003b1b87afe3555b6f78
f59f0cf10029b327727f6cacacc3d9ddc59f81e5
15075 F20110318_AABYTX mabiso_a_Page_018.QC.jpg
913dbbb5a48e58022e03704c31e1f326
fe511bf3f56b3a3838d354568c8f78a594224059
29264 F20110318_AABZAG mabiso_a_Page_108.QC.jpg
62c7c14d3324deb01d41171f236b6881
34a3d9249687975eac32caa47b820c3748743ef5
70400 F20110318_AABYVA mabiso_a_Page_034.jpg
10fb5e9ba792084de1a3cbefff9f5464
7ca14c0381a2b00475a39bb2cd4b2a22a7efb9c2
84635 F20110318_AABYUM mabiso_a_Page_027.jpg
db318045de58c1adc761e197329006ee
ce6ca367652934a444d2565f5a44f98c49ed1252
78845 F20110318_AABYTY mabiso_a_Page_019.jpg
f8bcd3bf3e2b6ad38b362cf68c6922fe
80e7f60259252227f831985cda2d0ac43cbd31af
129652 F20110318_AABZAH mabiso_a_Page_109.jpg
4f55d078ce429c773a6a7ff48ede8ac9
38bae22a308ef8faa89cc8344be3f71d235f2da9
22277 F20110318_AABYVB mabiso_a_Page_034.QC.jpg
9f681e0f2c8993e3a540509657e3f0fd
b929909fea63075c566fbbd11b5c93d50a431000
26279 F20110318_AABYUN mabiso_a_Page_027.QC.jpg
515988f60d5bb33c571ae4a49a28342a
d8fb6ecfa6b4887f7d13e74380c5354a1c9159d2
24057 F20110318_AABYTZ mabiso_a_Page_019.QC.jpg
cceb67c1871ea480c70a1a15c4e67efd
429ba0eb688d69f793f4f0dc736b928959be3e1c
32057 F20110318_AABZAI mabiso_a_Page_109.QC.jpg
927f20c956733ad36794d70be5df8d8b
a69677d9cd422dd7605da6a95af8a7e84eaf161f
26425 F20110318_AABYVC mabiso_a_Page_035.QC.jpg
6b1e2ca1c75b022f67b416a2fbbab491
ec51895f5ddd10e151dc3a126eb78769e886e687
82791 F20110318_AABYUO mabiso_a_Page_028.jpg
0403aad8bbeb83c4c6acd8d884ee6270
f6c03ec1adc63354536da0a2ebd1a03b80187cfb
96089 F20110318_AABZAJ mabiso_a_Page_110.jpg
90faa3809d712087cc3a3153a2e841ac
368d7ad98406e77a2acd0e226ad743b28be84cee
87824 F20110318_AABYVD mabiso_a_Page_036.jpg
4eeda49dbf36420804418d4d6266b562
71e6c64ff9dc5e284c75c43e7c0b51c8b22ababb
26022 F20110318_AABYUP mabiso_a_Page_028.QC.jpg
812abcd72931636aa04374f13e703d95
4c53d856170700f3175cc70a32f40591db42fb7a
24837 F20110318_AABZAK mabiso_a_Page_110.QC.jpg
fd2b892be8d70c3206991d40269f352f
a639a8c3787dc69c99ca279d23aff7a9be728340
27191 F20110318_AABYVE mabiso_a_Page_036.QC.jpg
4eda352908fc79ea0d39f0f6dff752fd
be4830992c20d8d36b446d12b5193a7bc6174e7c
69381 F20110318_AABYUQ mabiso_a_Page_029.jpg
3b78259979ca7335e4e4f4743048757d
0db4a025f64c71d3c265d94820a12d94f938e188
4515 F20110318_AABZAL mabiso_a_Page_111.jpg
a46086365c9d7ede4b1c4a78b29d617c
3bfac84640e72db8aab3834e2cea3d9bf13f2c1e
86544 F20110318_AABYVF mabiso_a_Page_037.jpg
200e15a08487701f1e69c806d8c97aea
6110550878b2ddb469d9a03eb4d553dc6129bffa
21738 F20110318_AABYUR mabiso_a_Page_029.QC.jpg
abbd2d33eaad7107f5b92fc54e279db4
87ba114220692e8dbba9fb0cc3e6598b3a4c92f4
13666 F20110318_AABZBA mabiso_a_Page_119.QC.jpg
23115a8a72ab91cc8fe68f17a7411a33
c7a127b47b886da9576ef4a5e31adea574730bb9
58229 F20110318_AABZAM mabiso_a_Page_112.jpg
0042eae7b8d6b129c0741b807f17d358
5150c47b4b283ec97f0ced01ae6d9f467f690af2
85443 F20110318_AABYVG mabiso_a_Page_038.jpg
7ed56ba74dc038b8c7687413f0ce5b89
403f283ed88c40cea05cc1e38f7f6a3eb5fd02b5
52265 F20110318_AABYUS mabiso_a_Page_030.jpg
96206f9af3d65a8bb39dae7a9c813f87
977a9884651f8e61f252953109cf7fa531a17606
86664 F20110318_AABZBB mabiso_a_Page_120.jpg
058d2c3dca20745d7c0b728e12bd2cc4
eb4efb5539ee0deaaaca30fe46a5fcf60eb0a4bf
16483 F20110318_AABZAN mabiso_a_Page_112.QC.jpg
73bc29f1570c3dbb1dda1dcc1544a12a
23a82945cb511b1eea5eac7f7b64c03e2bf7b428
27557 F20110318_AABYVH mabiso_a_Page_038.QC.jpg
c7cc9cdd5419c321e9e10dde7afa4c86
2071be21d8f041941823d1a0d8ecc938a5bc876c
16872 F20110318_AABYUT mabiso_a_Page_030.QC.jpg
851673e2648cf562978e6c9ffe8981b9
014c86db851ea89163257d3a5d55b1b2883df5cc
24482 F20110318_AABZBC mabiso_a_Page_120.QC.jpg
11a10bd11a67188236963a6e592856de
d9c99bfea17e1d61b269d0c4a355d7918fdc1b74
16290 F20110318_AABZAO mabiso_a_Page_113.QC.jpg
7bf1528e9da5cbf03506e1840dab6ce8
4321d2a21e5d555f61cee8c404840c39697fcec2
38358 F20110318_AABYVI mabiso_a_Page_039.jpg
8657c14c8957577249ffa50c04b9436a
ab8836cf58eeb21610aeb5197bfd445a47f3c540
52900 F20110318_AABYUU mabiso_a_Page_031.jpg
ab19f72cedeab888314dd642c6f9e3c1
8ed6b486ccbd9de91efa91bb9d4913f282b989ba
39974 F20110318_AABZBD mabiso_a_Page_121.jpg
26a1457f9ad599204b5101dd8cf23b2b
1ed359367d65b9b2f4deea0ade310999aa6922bd
75395 F20110318_AABZAP mabiso_a_Page_114.jpg
a3de43e5c7dad6173259ec94be05486a
b542a17f5003e00bf1d8f339995fb045067e1d8d
11978 F20110318_AABYVJ mabiso_a_Page_039.QC.jpg
89bfb7259e5c2c0fd079da548b7ae4f4
9eb68152109e60651f568aea0d68bee1d2ea71bc
18244 F20110318_AABYUV mabiso_a_Page_031.QC.jpg
f6256535bfb3fb0f194c76514c07114d
53dc1d0e858698acd99f2de0195c8c18755d912b
12338 F20110318_AABZBE mabiso_a_Page_121.QC.jpg
940d70ede033463a632496c807490a44
37a39a9d5d8cbbf614eec58da5925ba139767331
22862 F20110318_AABZAQ mabiso_a_Page_114.QC.jpg
1753fc1fc0e336682a0bc7a5b7f4bd00
c4ded0c2c5ade092287b2e5cb9514d2d2cbcef9b
76443 F20110318_AABYVK mabiso_a_Page_040.jpg
6191cc4c1ea719cbf376647289c6a7cf
a3a36efbc35fd18e4820565586b88cf6b36a378b
69926 F20110318_AABYUW mabiso_a_Page_032.jpg
86e1c7d6c12efdca17adeb086c266384
de139e6b3acc4364e4828bf6569d9ff66243dac9
40207 F20110318_AABZBF mabiso_a_Page_122.jpg
0f0cf7d099e57e6e6602cdd84aaecaf1
d2bb8e28949aa9f2b3b4424425ae9711660c612a
82022 F20110318_AABZAR mabiso_a_Page_115.jpg
fb0facdc9d0f02b066cdacfdc0533616
0138f9089f00fc944b70f2890c447c18aa2c0c54
23686 F20110318_AABYVL mabiso_a_Page_040.QC.jpg
e0e72d1005a0eae5103755d555eddd6f
788f9b661a8e8eddc29200bd806f43c853013012
12429 F20110318_AABZBG mabiso_a_Page_122.QC.jpg
275635b4b30b9af7514a383b8a64325c
2990497f48824208a4adbfbbf43115eb52a2e772
85893 F20110318_AABYWA mabiso_a_Page_050.jpg
043db1ef3ef861f09f36e3b21c88b156
4475a7c6d661877b6efa708164462543f9dd6231
23678 F20110318_AABZAS mabiso_a_Page_115.QC.jpg
99b3661eaad679cf482ff1a6c6e619a8
09776a33b988199a983ab93225d995584a2395ee
79917 F20110318_AABYVM mabiso_a_Page_041.jpg
f6f849c69b94b7777a53b21188ea0f60
8006f84b57eaada05098ca4685a8c4f10c3eb073
22285 F20110318_AABYUX mabiso_a_Page_032.QC.jpg
918b316954fdfaece00e41991888a83b
e7c75034349c9ddca725e4ec99ba851e96c117b5
69946 F20110318_AABZBH mabiso_a_Page_123.jpg
8b455663444fa76ee7f78d5ee785fa27
a4b64ba2f79d6d533c8680af0c9d9afa97b73371
26433 F20110318_AABYWB mabiso_a_Page_050.QC.jpg
6a7909083d615647a05c3a54225ac0da
4d49e9acb03b1139a373992672824e6572a4b18f
31542 F20110318_AABZAT mabiso_a_Page_116.jpg
dba3b64562eff446fd7afc836411c78f
5f40fe189890c4f83ec9b1e52bfe7ed2a021923d
39434 F20110318_AABYUY mabiso_a_Page_033.jpg
c20a8cb92a1999d5fa318226424d696a
99beb4c2ffb1f6393c2e2f5fd0b55b2abe4b1b29
19586 F20110318_AABZBI mabiso_a_Page_123.QC.jpg
aadf844e737c9fbef01e6e66234f554c
0f19e8cae4f4385c3be90dbacd410640cc34e2e5
6719 F20110318_AABYWC mabiso_a_Page_051.jpg
624298cecf784c21c479f9fb6d495c70
c24db15d7020278e13a13e2c1a6c5da8f0fc4724
9195 F20110318_AABZAU mabiso_a_Page_116.QC.jpg
4b2df863e9eaf009a4f897251003291e
5893b479e2340759cc0ff49f5a0c8e5ffc4f8dbc
24489 F20110318_AABYVN mabiso_a_Page_041.QC.jpg
236274d0b3bf13aa64bc22259075062d
d9b80e00e1064e4186fae570a2b46a4ab592ea31
14275 F20110318_AABYUZ mabiso_a_Page_033.QC.jpg
c0e150d7403c62974d9986029ed0c21c
99ed6352a2c9e936b7ec9eb465529c970080ea13
61349 F20110318_AABZBJ mabiso_a_Page_124.jpg
2dfe63e0468e26bc479de2362d3e82d4
445211393fcc8fb838d242091225cc6d63dfd040
2434 F20110318_AABYWD mabiso_a_Page_051.QC.jpg
1e8d765050d9b7886c00ea3907145c65
91791d5c7e892b88c24ed1050f5de0a8f5db42df
73874 F20110318_AABZAV mabiso_a_Page_117.jpg
3520a5c445be3b13718bef7e80a13c59
3fb855cafe5b5de6acc8ea1272ea7899f6a539d3
83987 F20110318_AABYVO mabiso_a_Page_042.jpg
68f8b788769d735d20ab3f3e696e5aeb
0958533a72274364622d50d29f15878f0592223a
17069 F20110318_AABZBK mabiso_a_Page_124.QC.jpg
3e8a745d3918cb516e94e87c20d004dc
468e67fd9423087c8cbafe5c397c881299d8e4b5
77493 F20110318_AABYWE mabiso_a_Page_052.jpg
2a981b5867ed14369b376c7a0287996d
0cfb0531e6d42de2bac8796843879c0a3be28c84
22114 F20110318_AABZAW mabiso_a_Page_117.QC.jpg
11f89cb41bbb2d5afd8fdebc872aa7c2
af1add3e4592c648461b4391a4619342db2867c4
26453 F20110318_AABYVP mabiso_a_Page_042.QC.jpg
04f3cf5f86d5aec0afc4b075c523abf2
3594e4f479cf011897ae16b29f5a165281f6e8b5
5545 F20110318_AABZBL mabiso_a_Page_125.jpg
3c6df9bded85e7d45b56001d39fcd4f3
38ebd0a62e570a659613056229f107ece6d3954f
23839 F20110318_AABYWF mabiso_a_Page_052.QC.jpg
57eadc4bb6babe7097ae908733596ee9
2fd3dbf168a624c6e65d8dcadff4f4f816100938
73076 F20110318_AABYVQ mabiso_a_Page_043.jpg
8669dd59ecefe06efbf5743b96ad5721
5d428350dca672afacf227f38de643fb08b3a281
1629 F20110318_AABZBM mabiso_a_Page_125.QC.jpg
fc886e5c0e5fd2afc7870ea8597fffdb
185db92d23353c87b5cc2d37ca352bca57ee61df
71787 F20110318_AABYWG mabiso_a_Page_053.jpg
5e3ea891e7bf236a2537765f6dd0a61c
e59667c194ddb55e83d90d1ee2699566403141f4
46491 F20110318_AABZAX mabiso_a_Page_118.jpg
f3e2e50a7fa6dad0137db97f23443a7a
0d781da1d471562d26619dc9f35d690aabca1dff
23875 F20110318_AABYVR mabiso_a_Page_043.QC.jpg
72f415b5d0b0a368815533dbbdd28e96
34370fefdf3e5a944b6b1cacd1a5d42d3553e652
32263 F20110318_AABZCA mabiso_a_Page_133.jpg
97830ad91695afdae5dfddad9632d995
9d165585e9880366fea18d3e87190be33413c425
108264 F20110318_AABZBN mabiso_a_Page_126.jpg
e17fa83af80fa96350782577e374f086
acc41516383ee8395ac39369e059c15e0619d2f2
22927 F20110318_AABYWH mabiso_a_Page_053.QC.jpg
7c4821115d5fddaf29db2c0a89447bc3
8a92885a22f5f24454a863cb8adc81e6a87c8eb7
15055 F20110318_AABZAY mabiso_a_Page_118.QC.jpg
080fa1e0b8f2d340d97edd760f235fb3
0637cc49fb49f9f7c9d1eeeb884bf5f97c26b4ab
22581 F20110318_AABYVS mabiso_a_Page_044.QC.jpg
86e1e5fe7a711bae4f4280e8ac0fe106
5736913ed24628d1307c97f7bdedd682f7af4f9d
11151 F20110318_AABZCB mabiso_a_Page_133.QC.jpg
91a53033fab0b0925551a251d6f345df
f2b2176846f121edd010e3374a1e7051fb777110
27432 F20110318_AABZBO mabiso_a_Page_126.QC.jpg
6102e3a174bb6210047df44fe3a39004
9c6c839e111b2dadf6ab5c6ee418d4414cf1dc1d
62446 F20110318_AABYWI mabiso_a_Page_054.jpg
9ba4459abd07df20c59f694d6845934c
db297dbb5b2689c28fec9029d2ad9a8b4c8231d0
45452 F20110318_AABZAZ mabiso_a_Page_119.jpg
af3222ba143b2dd2de753834b5b997f1
cb42bc33379254dacfd132ddd485eb2138a44683
77245 F20110318_AABYVT mabiso_a_Page_045.jpg
bd47a837f1da61cb0c57a84e29acb9a5
82a4a1f38cf3416f2d34102d54af2fd2eb79cc32
42586 F20110318_AABZCC mabiso_a_Page_134.jpg
273b1bd13b1845a5c602ab75e318621e
ad8e1bb250f4fd3627fe8cc0c3f34863ed14f345
56028 F20110318_AABZBP mabiso_a_Page_127.jpg
e4fc3cc5803167e07f56f51f03b37d0c
c6434b82db977ec72a78d30aa78f781a64c1636f
20148 F20110318_AABYWJ mabiso_a_Page_054.QC.jpg
10b5e8bddda5b3149b140529323ec730
2a66eda40ee77b9c30ccd7cd4f209e75a1327079
23999 F20110318_AABYVU mabiso_a_Page_045.QC.jpg
ce8387df9b0dd744847383ea011cc96f
3ae65aa0732f890a4713bbbb37ee5a348d5f6f51
13532 F20110318_AABZCD mabiso_a_Page_134.QC.jpg
0b0e8926c164541b0c751963ca995ae4
35443f77a06da915de7a4e2a98f14ce666d7d3d9
17369 F20110318_AABZBQ mabiso_a_Page_127.QC.jpg
883c445f50d1a9068b777e38b633ccaf
bfec6756556dba16a99827e4f2bac818757523f4
74657 F20110318_AABYWK mabiso_a_Page_055.jpg
42630c6fdc71634cdb78e610a21b44a8
f4747747203e82c87f9c1f9be470246fdc0ddb66
16639 F20110318_AABYVV mabiso_a_Page_046.QC.jpg
9440d02021c799ac4149fccdd0b16429
6ec36c3ce7ac91d61a0d27c78864f322ed440a41
22730 F20110318_AABZCE mabiso_a_Page_135.jpg
cbf44266748c6536529bbcf1c432ab7b
98046176b020ad54463080030ce300f16e888e47
68308 F20110318_AABZBR mabiso_a_Page_128.jpg
ff2357e4d878eecec50db2dd43901a58
20cff4bfacd064dc142ed7243eebfc0058ed3474
23856 F20110318_AABYWL mabiso_a_Page_055.QC.jpg
2c7de312017ca71add3e56d55a905d74
c42db4c332d37f07bfa870afbdded6ea6fc0876f
74240 F20110318_AABYVW mabiso_a_Page_047.jpg
6a79b96089c863738fa4fdbaa707d366
3af57d02ed74e9a8a57e226cdfd991f0c8e809f4
7478 F20110318_AABZCF mabiso_a_Page_135.QC.jpg
abc9f02d725354f9fd7ad7337b4aea3b
0368ae193e3b545bd34229daf495024e491f898c
20751 F20110318_AABZBS mabiso_a_Page_128.QC.jpg
edfdc1a0ce68b0f57a3e18ac2c429396
03da077f68c6f688a07f80bf032843a66cc06180
55620 F20110318_AABYWM mabiso_a_Page_056.jpg
ea1ed32b52acb74e9544625eb19d8831
4ae8b5b134fad602e6f262b9eb9fcdac4425f1bf
86916 F20110318_AABYVX mabiso_a_Page_048.jpg
41d0ee3a5b4da8ba450751025afadfcf
2c742bee436deab24341109645554a0b53fa2b51
87633 F20110318_AABZCG mabiso_a_Page_136.jpg
ee929c5e4725240b1ab1143c60964536
c09ed07454c1261012a0861b954af84f1c8f0f64
24492 F20110318_AABYXA mabiso_a_Page_063.QC.jpg
459f95885ecc1f043bcb5cdb6decd9ea
9b8652297e75847d5a2df98f400aaadd84865ed8
52591 F20110318_AABZBT mabiso_a_Page_129.jpg
0345b0c180358e9150d1b8b92bf0671e
6acd379040296f012db6365a21b07e9315adc32a
17516 F20110318_AABYWN mabiso_a_Page_056.QC.jpg
4544faea3e14deb58dcb14aa3d4f18a9
df7c05a8e5dd1292a18d653591d0bb9ab8c9ae6c
26999 F20110318_AABYVY mabiso_a_Page_048.QC.jpg
52798deec0c30a6215019f353d6caf8f
809befc613a1217d100bd1b352ea76c3625dfe88
F20110318_AABZCH mabiso_a_Page_136.QC.jpg
98f306758c6fd32da4e47fcbef53ecef
b88ce6ae7f888a9f442ef36900b8c20689f51be2
81685 F20110318_AABYXB mabiso_a_Page_064.jpg
a999619087a6bf314be35dff3c4494bf
2745a624ae28eb09cae7ecd305624172325ada40
16854 F20110318_AABZBU mabiso_a_Page_129.QC.jpg
8551263351a5d950cb33154acd8676b1
0b9fd61cbb0d2fbe0ebf4764f581d9cd80e0a424
26150 F20110318_AABYVZ mabiso_a_Page_049.QC.jpg
0149fef23d875bc596b1fe101500d432
a605665afd7515412f8ed7ff99c8034512327c35
111410 F20110318_AABZCI mabiso_a_Page_137.jpg
f4976235601217124102f41f528b9d47
af64c6ae3363a89175e74f46856f0a73691d6cb5
25188 F20110318_AABYXC mabiso_a_Page_064.QC.jpg
3e041bd2096ebb4503b65f686247d0e2
5e59551debe68754aa3eac0895bd0e119ceecc46
65495 F20110318_AABYWO mabiso_a_Page_057.jpg
05ef4d055527af4aaf96cc738e316130
04bdbae2a3d9c74ff9532025624cc71ab73710ba
29282 F20110318_AABZCJ mabiso_a_Page_137.QC.jpg
a20e69cf50a505b2a26f8e911436cffc
d4ff97ab3fb06415e52293d73f8acde47e21e2e7
77797 F20110318_AABYXD mabiso_a_Page_065.jpg
39e77edc01030ceb4df973b5f3e57d15
8fa3ce34c4a6d918be6454ab9e45bda6dd027f62
69310 F20110318_AABZBV mabiso_a_Page_130.jpg
bbdf38e2640905c1c172bf54a59889d9
133e79259173898dd9be3e12534c5ff76bdc50ae
21119 F20110318_AABYWP mabiso_a_Page_057.QC.jpg
d7e08a82008230b92998b9997736d905
f90ef2f703f4153e2255bfcd6427eccb7acfa554
107394 F20110318_AABZCK mabiso_a_Page_138.jpg
a1cfe38e116eb4b43339f67c018bd4ab
dc9db466ed7932ed0435471f2bcb389702214054
24480 F20110318_AABYXE mabiso_a_Page_065.QC.jpg
6eaf7206249d6853c2434b6f32f072ce
268c4c32e85990fbd12ad2186b3e180a36ce2f1e
20676 F20110318_AABZBW mabiso_a_Page_130.QC.jpg
8385db346ea30213bf1a49713f0444c8
5815563f81a35e344cb0109ef170577e3ad435fb
75214 F20110318_AABYWQ mabiso_a_Page_058.jpg
a578df036222d7c1f1a2c8a0c526a6cd
ee286cc28c83a96683b5d1e545b4a2507241ab12
29546 F20110318_AABZCL mabiso_a_Page_138.QC.jpg
ae44c374fdd12deb4176906a128a3b41
8e8b5009b2df94f639d3087bb97b4b7c48b49d92
81205 F20110318_AABYXF mabiso_a_Page_066.jpg
714476d70fcb7e8089c75940f98b8dc8
a5209c7abb0070bd3533785af42cac2d94bdd55d
49831 F20110318_AABZBX mabiso_a_Page_131.jpg
2463b63b52ef755e62ca3cc3aece1608
37113f93a2af8349e62c4da19e40dc67241d41bc
23225 F20110318_AABYWR mabiso_a_Page_058.QC.jpg
7e7bf413eeb1adcb047815bd4b926496
c5530d18d5bcfe6503946dc8ed7ccbcc7602db29
1051974 F20110318_AABZDA mabiso_a_Page_009.jp2
49d68028438fb745302b5807d5f0b871
38071b01d9ee891b82b570b0508b3af98f7ea499
106613 F20110318_AABZCM mabiso_a_Page_139.jpg
533dd379d275bef4bffbfb3b42eb6e6c
aedb7665154b09c5b1dc6950fa7ef5e986432391
25760 F20110318_AABYXG mabiso_a_Page_066.QC.jpg
ebe6e9b196fdf3be389ea485e57490dc
16c3670dbd49926d74ec60390c9323b690ed72e4
15163 F20110318_AABYWS mabiso_a_Page_059.jpg
b0393261ae0059c328ec0979b57db955
84e699cad0c2150bbba8b66fdd866066b31f8b30
1051961 F20110318_AABZDB mabiso_a_Page_010.jp2
a19de16a2753cd2d7c88a0354570246b
daf3248a70ad62ccdb20994d535d5fe32ab0f8d7
28744 F20110318_AABZCN mabiso_a_Page_139.QC.jpg
11b4fbd2e17143a96ab0c039d49d58b9
4fc22fcb478cccbf68a50d2efb878a9acda662fa
28452 F20110318_AABYXH mabiso_a_Page_068.jpg
729cce8c80ee128d3060d15e5aa732df
a75262f39628ed7f37ccb6887af1e4f1e4ad9149
15513 F20110318_AABZBY mabiso_a_Page_131.QC.jpg
719d7c2056a007c7ba32dbc65edaf399
e3411e43887eefaf93f6a810d3ec7bbecd867562
5017 F20110318_AABYWT mabiso_a_Page_059.QC.jpg
5c4fda6b77a802d204bd06bc31a42f50
4121c4a029397fe8df5e051664f38f59a9c7640d
92770 F20110318_AABZDC mabiso_a_Page_011.jp2
dba941d0307658e2d58bb1e543739300
7192938e5e0b24664161f7b9b74b818aedb29396
89265 F20110318_AABZCO mabiso_a_Page_140.jpg
955917ea9dbcc44e08124400e4ec91ee
2cab952cc7843dae80fe646224973db7280016b9
8180 F20110318_AABYXI mabiso_a_Page_068.QC.jpg
d0d12848d4828e72efcf29dd53a9450b
39a2872a971f95b750a8c217952390723434f322
36199 F20110318_AABZBZ mabiso_a_Page_132.jpg
032a5a80d2a0472a420e583c7d65c2ea
4a13de135886538107f47b875c053e4ae90d2966
51045 F20110318_AABYWU mabiso_a_Page_060.jpg
def282a4420f203a07c78b7c61e98be2
68b2232aae2b699fb28aa6a75f781e34d22ffc94
1030323 F20110318_AABZDD mabiso_a_Page_013.jp2
01d2d6cf4e32bb3d53106b63d1f59244
47ac4df5aeea5d11e7d13052d65efef35428fc46
24481 F20110318_AABZCP mabiso_a_Page_140.QC.jpg
19eb0ab2ac5eb0713d09a89726bcd6af
6421043e38afaf9c5b370f1e942c3fa3bbc269ea
8391 F20110318_AABYXJ mabiso_a_Page_069.QC.jpg
2959d10f61a704a5cde2ab9f580465a2
5409b89fec160eb5ba19ad7277e4f3934df123ae
16460 F20110318_AABYWV mabiso_a_Page_060.QC.jpg
a8882d47e15073a8e9d60f0caea80633
6e84a257e7b35ce8ebc7a7292c8c3f2d2457b756
1051944 F20110318_AABZDE mabiso_a_Page_014.jp2
000fe1dbf83196a40c956b9292f1504f
7d4673a5e6963f94e9f1ceb54c76e8f4eca56525
23720 F20110318_AABZCQ mabiso_a_Page_141.jpg
f22889d5723f3075f5a709a439892d8a
09c496abc905f4df7b1e246303c4aa11d3d0c3e0
77437 F20110318_AABYXK mabiso_a_Page_070.jpg
08dfb5ecbfb792206096ba2a8e146c32
01364a3c2ac40361c2764ba28376123091130a5b
78014 F20110318_AABYWW mabiso_a_Page_061.jpg
e9832f9087c63c1d8df3d1f1153ef61f
d4e7b30817c6bd22ea1144786bc37c95cad8456a
1051972 F20110318_AABZDF mabiso_a_Page_015.jp2
6db5cc0f7a2f96000dd0f819f645a03b
fa6e14095af58916e19b4384ebe50f472e1cc440
7574 F20110318_AABZCR mabiso_a_Page_141.QC.jpg
d43b83ada7349d292bf24fcf0bb5e697
f557f5f25e509bd76ee821ab83a41068879e8e1f
22823 F20110318_AABYXL mabiso_a_Page_070.QC.jpg
35cb66f83a716d122a59bfa45fbe3888
15cd65fa999390d0bcd9641d5486b1455c00effc
24389 F20110318_AABYWX mabiso_a_Page_061.QC.jpg
1040f0080c20816be6e2e4d1019ff300
b8af52d01e95a434aee6f9aeac73239541a954c2
101861 F20110318_AABZDG mabiso_a_Page_016.jp2
3b250294226f1e7c770d2b16ead84332
81d07cdb11c82f0960a8f0e3d9c43793a807b8cf
22883 F20110318_AABYYA mabiso_a_Page_078.QC.jpg
407ce95d83144f08bbbe8cb70f91c3e5
ad633bc95b3e870ca73f8b7d65d8f7722cb972fd
28889 F20110318_AABZCS mabiso_a_Page_001.jp2
72da4e1dc9f7b3a30089876df69b6185
0415b4c17d894da5660b111c9817c044ef98bac1
26986 F20110318_AABYXM mabiso_a_Page_071.QC.jpg
4fe895bef82565c898140554946f8a1d
2553b95a15bc94a9c6ea3f6455d0aee5b3916bc7
82529 F20110318_AABYWY mabiso_a_Page_062.jpg
2e6f910ca56728d551e3c2f6c89b1c85
b1e778db5fa5992170600cf24154baba889f7c6e
107211 F20110318_AABZDH mabiso_a_Page_017.jp2
b67aa3152aaefedfd18b55ab69b81a2d
c8a36785aae5d9b98dabc2986216ccb373a4ca59
64543 F20110318_AABYYB mabiso_a_Page_079.jpg
201bdf355485ab64086a4ea8e8f3a47b
80de8d8219f9832973f22721144e4b7142b356a7
5486 F20110318_AABZCT mabiso_a_Page_002.jp2
1c76553f261757e55a22120d7a3f6351
b9086fff217bad1b38da2690b35c49d68fd5a701
81686 F20110318_AABYXN mabiso_a_Page_072.jpg
fc692134023b27bc10165e8638d0498b
53b85f307f87d7f1d09cdb8b3700c73133053fd7
25471 F20110318_AABYWZ mabiso_a_Page_062.QC.jpg
038872a0725f9f7b4dbcb630a07621fa
d3870e5f9452f45322d3fba4f74b7aaf256d140d
63051 F20110318_AABZDI mabiso_a_Page_018.jp2
30dce65d07cde69aa29ef4f5e10a8afc
acb9b68edcd2fb865642bcc77845aec7aa9fbb67
21135 F20110318_AABYYC mabiso_a_Page_079.QC.jpg
6beeb28e798ac177acc5b9f7577159dd
037f48537cd5cebb7655602b2c2448c373b9ddae
4105 F20110318_AABZCU mabiso_a_Page_003.jp2
88b7867b168fe9bd66f69695fcdabcef
4ec5d027a194ef49d75e9435bf08675c9af9231d
24039 F20110318_AABYXO mabiso_a_Page_072.QC.jpg
1a2891ca47890f0032363b513f704317
2dacb4bdda923c114ae257ab63b760587253793b
1047294 F20110318_AABZDJ mabiso_a_Page_019.jp2
6e5b918ff66e934b197978bb43032cfc
567958ca66962bb876a03797027c75f00c43e9c6
60277 F20110318_AABYYD mabiso_a_Page_080.jpg
7819e27c672812d9f6834590bc88e7af
85b52ef2b97cfb937a2c636786b10738491f7b6e
88398 F20110318_AABZCV mabiso_a_Page_004.jp2
2ca99a5c7722137b11594d76d96b5d60
2f0ade2d2a3ec631aec880b572e7e745dbe4ac41
1051976 F20110318_AABZDK mabiso_a_Page_020.jp2
497ad8e451d45a3ba795036eafc0fdda
bbce645ef8615d8a5774de8db8abea1d311de815
18754 F20110318_AABYYE mabiso_a_Page_080.QC.jpg
34a353e46c634f31012997b38ccbdf56
99365f4fa9dd5107c76ba5e86aedc08d9209cd9f
F20110318_AABZCW mabiso_a_Page_005.jp2
ab7e87d27beeda4d1305f4f629fb04eb
f246987c4bcfb8d2dda53989fd1b11b9591a39c8
89866 F20110318_AABYXP mabiso_a_Page_073.jpg
d73a136fd08f6644a18c4444016f756c
0db984028aa7152e9caa84e2be086a53e11b5c9a
1051983 F20110318_AABZDL mabiso_a_Page_021.jp2
47357d2e16351d92780fc1d10832533c
6cda7c74e42f9f1ab710607d26562c0b55a04ae9
48551 F20110318_AABYYF mabiso_a_Page_081.jpg
3d2960d30ee521766b83aeb6cea70d90
2e544e6310c75dedb9df069ad02887af730aaed0
1051914 F20110318_AABZCX mabiso_a_Page_006.jp2
6c05e275e34ca15d3ad1d6c3357dddef
2bcf81add9f56d900867aaf7be003f25dd341548
26168 F20110318_AABYXQ mabiso_a_Page_073.QC.jpg
1f741ba14bdb01287beff0f0caa0a0dd
25a34d888a1f3636fb79754e71d96bf46931edb7
1051965 F20110318_AABZDM mabiso_a_Page_022.jp2
937198e0d0e380219a604273ed1e4f3c
b63064d440373a4d6072f74f4b813736ae47b9eb
16312 F20110318_AABYYG mabiso_a_Page_081.QC.jpg
f5c8db97596edcb2bc5815ffa48003b7
ba3ca4e6db2265061f4efbafab4f1a5a3928b7d3
155526 F20110318_AABZCY mabiso_a_Page_007.jp2
5333ca3451183d44fa86e4f56005dc54
0fb76e16092eae78a9b4a9b5f12bf52701d0a0ee
26926 F20110318_AABYXR mabiso_a_Page_074.jpg
4008a9649d0264027f7a804e968f5b11
b1e2047ce48adc5ce240733be638ad12d7792c93
1051889 F20110318_AABZEA mabiso_a_Page_037.jp2
df2468aaa09819dce796bc900f32d390
4412c97f234cfab22e7ab94682b287d6886bb918
6419 F20110318_AABYAL mabiso_a_Page_024thm.jpg
c6606eec047801cab8e016799c24ea55
d1771061a121680231bc012d92c08cbbece994af
1051970 F20110318_AABZDN mabiso_a_Page_023.jp2
70cbe0ae0578107d561a4e4ab1ba27ee
60e2d83928fc1b9f81e1fc5c676663d42591ab74
71317 F20110318_AABYYH mabiso_a_Page_082.jpg
87ece08014aa34ab19d7405ca28a6bb8
4cc9da6a43dad6ecb3df5eafd86fa5982f5d1d38
8910 F20110318_AABYXS mabiso_a_Page_074.QC.jpg
70b86ddf2e5db8bb454b4bfd5593d26d
24a885710d2581c291b45ffbcd6277e258c12dce
1051923 F20110318_AABZEB mabiso_a_Page_038.jp2
e861092164e752e4a6da2b7a9a1150c2
0895a1631e6fa6ccb51ed9aa3c63ee13c723a0d1
52293 F20110318_AABYAM mabiso_a_Page_113.pro
98d3da7a65d22a839671b3d27e0e45b4
1b4110e549a0318e3e84cbda5960ba36d7ae8013
1051985 F20110318_AABZDO mabiso_a_Page_025.jp2
3deb394df7949df838741c13fc129d3c
f7b56ad648c3c872b3cde0ead76c75e190899e11
21264 F20110318_AABYYI mabiso_a_Page_082.QC.jpg
41346af60efd44ec52db40d702029e47
18a13aa11679bded6566dea19bec49d021f573b7
1051986 F20110318_AABZCZ mabiso_a_Page_008.jp2
42b411d86f451abd5374db54b409e009
693ec2645a4dc9bbf11bb80dd18e159e041f0849
70910 F20110318_AABYXT mabiso_a_Page_075.jpg
57c9994cb57455e8a6235a0a1d2c184b
cdad249f3eebfee0654070762282228f92d0f5d0
F20110318_AABYBA mabiso_a_Page_029.tif
23763e1738cdcbdde46bf08a6324d858
6194140f472175ec6c72c8e1f2bf8c6f0bba0681
49413 F20110318_AABZEC mabiso_a_Page_039.jp2
583e6536dce19e39f8e9a530a57b0706
24d9c916925716bf90f5a19c01a129ff1ecdfda9
6167 F20110318_AABYAN mabiso_a_Page_066thm.jpg
c648bf87e562fc550466c55267761844
e805dda6ad75ab6215332c74aa77623d0c7912fb
1051984 F20110318_AABZDP mabiso_a_Page_026.jp2
1232ee32075bc195ba449718fcb9a0ae
e89c0c0be6f6d0f5fee03cf8e90ea18fac2a2277
95456 F20110318_AABYYJ mabiso_a_Page_083.jpg
0ba6d89f49b9be2819428899cc447918
88ef4d5337ccb1b19c2e55c3660b2727bfccdffa
F20110318_AABYXU mabiso_a_Page_075.QC.jpg
beb7a5efd89c09c7b08fef5aaf4b3a35
96d337e5d53617cd3409470bab5f5ab1b9ebdf3a
2025 F20110318_AABYBB mabiso_a_Page_037.txt
100c1a0cf7979e22243c1cf735a143bf
8d836396f0af25b0f0dfa4d0652863ca796c34a6
1025349 F20110318_AABZED mabiso_a_Page_040.jp2
1e1456ef1a79e5ca7bccc35fcc3e644e
9854d5e4a238794b76cb5cbdae1d609da4a6ea6d
46307 F20110318_AABYAO mabiso_a_Page_064.pro
16110c6900153560f458c2472660d911
ada06e00df9a43f34629714a858b1bad3562e778
1051982 F20110318_AABZDQ mabiso_a_Page_027.jp2
e01869850d31e1e3ed52a177b411d6ad
afcdfcbd7116360fe2e8bfcba8ddebd143f7fb0c
26297 F20110318_AABYYK mabiso_a_Page_083.QC.jpg
e1ec1ea7333e561cf6c8fa7739f4a0d1
7afe0878239d78f1cfdf205044f8b354b83b6fc7
70062 F20110318_AABYXV mabiso_a_Page_076.jpg
2f622dbef1f41a7a7d864d42ce59acbe
0c3dcfc8a4a8d1d6535ecb1ed959d9477d62c7bc
992887 F20110318_AABYBC mabiso_a_Page_047.jp2
631d6d0cad0b3f58cf68ee31fd984bf7
1a92f5504d0316878dcc44605d1792c06d4476ed
F20110318_AABZEE mabiso_a_Page_041.jp2
b47f8e894bc5936c564e74970e8128ee
be21683a9aef7a9b978dc19616996edee2d1a3fa
84613 F20110318_AABYAP mabiso_a_Page_015.jpg
fa53a6b787b7c2aa9f2a60bb0ed36349
8662422e04c5cea764f393c839db9bad154e2d99
1051932 F20110318_AABZDR mabiso_a_Page_028.jp2
7c0dbc4c9c12211b65b594ec41a0cda8
c90fa1e7669806d2401ff1ad736cf8bfd66751fb
79396 F20110318_AABYYL mabiso_a_Page_084.jpg
3d3d9a284da39d5e9c7033ce3c806ec9
4030c98f5f03ad6575fd51743ace2a8fef62ba5d
21395 F20110318_AABYXW mabiso_a_Page_076.QC.jpg
9e802958da0ad4a10283cab67038f90b
44fdf4b4cee5a42b07dadddd3dc5ec9709d136d9
89001 F20110318_AABYBD mabiso_a_Page_071.jpg
20cbb53f9ee393339ec047327b7bc9e5
5202b320817303deb6e62edb20fe1395d7edf6d3
1051966 F20110318_AABZEF mabiso_a_Page_042.jp2
c96e576634a84beab35df5ff5eb1404e
58474867a3b14504de8c4c64a72bfe3ab097b8eb
43834 F20110318_AABYAQ mabiso_a_Page_099.jp2
e9b021c87d65d2d9729510781897f583
6abe0fe0c9866f869718b5e0b6266573c6956381
83109 F20110318_AABYZA mabiso_a_Page_092.jpg
274822e230e0f0b851a788628444be42
3642d2734642a1a024cd1589e2f25a6d020cd15c
897124 F20110318_AABZDS mabiso_a_Page_029.jp2
935b323a8cdbb7babf73a277c24f1146
cf347dce062d99a78e08c80601f4dd8b05cdbcea
23552 F20110318_AABYYM mabiso_a_Page_084.QC.jpg
83d261f27651f1ebacceb0bf994d56e3
de2881011411433e3942ee1c646fb6778f38ac2e
55119 F20110318_AABYXX mabiso_a_Page_077.jpg
282fb07b8577c87113346f9c5963dfc9
86fd5bb1ecc35d8b910fa15a9d2c6d12e3c5d34b
68261 F20110318_AABYBE mabiso_a_Page_005.jpg
a3d69f5a8a3ccfff07a9d89ef8340816
353dffe142ec1f400c7c0cb394fc5e0db8d9a7ec
95384 F20110318_AABZEG mabiso_a_Page_043.jp2
18981b1bdb72115b2146570410f15508
4820caa9af4fac50aecd45e1b1579f7d460ae170
F20110318_AABYAR mabiso_a_Page_012.tif
8c5440127cbbaee8f0d5d9a4e5b75c88
27221de80647cd8925625b3c9720e7952ad7a136
26228 F20110318_AABYZB mabiso_a_Page_092.QC.jpg
e3f88b4ee46e31c009227bf70b00759c
2fc0b85ba0f20ba52d4532f9d725a7609bc4ecd1
67582 F20110318_AABZDT mabiso_a_Page_030.jp2
13ce62f71bd510ab7ed7e7285d608244
3678557259287c82d314c70727d477caae3ffbee
83626 F20110318_AABYYN mabiso_a_Page_085.jpg
7e41cee97778ea8f2a09753bd537e01e
a39c3d1f7f9958951bf3b03546162732e4c6e517
17764 F20110318_AABYXY mabiso_a_Page_077.QC.jpg
831aa1408088fe413ce6ca49ba686b02
7a37a20312adf7c2a73774bdd0d2ee1f0cff0e5a
78717 F20110318_AABYBF mabiso_a_Page_063.jpg
185dc4e0384cedc78eef9064fb576dc5
7a74dfb002025bd952edfeb81b7e9be9d5232826
892050 F20110318_AABZEH mabiso_a_Page_044.jp2
9238ba5045f36edae403d52258bda6e2
6663c65d5593050ae718f777cddfb149711ffe48
6650 F20110318_AABYAS mabiso_a_Page_022thm.jpg
f03dc7cac131fc086e70b8a3eb4f4670
8264e1966b671b0251369483eadf8dc9e3c28e26
87226 F20110318_AABYZC mabiso_a_Page_093.jpg
c11198a567a1925d94543288755d8d3e
037f1532bd3e341e9b688eb127caa398a18ebd76
602060 F20110318_AABZDU mabiso_a_Page_031.jp2
44177eaeec4bc9dde682ded5578d3799
b08d962430f36fc2af89352608940c4370a878bb
23973 F20110318_AABYYO mabiso_a_Page_085.QC.jpg
53a81e1694b05b4965f5b33a71f7f7b7
67427e2d7b03cc8571e823de2c13c06dd56e8315
74269 F20110318_AABYXZ mabiso_a_Page_078.jpg
b5d0437d0405b5f78dc05b389dfaeacb
0089f76ac7f579b7bb189f4f75df0247a561af96
1736 F20110318_AABYBG mabiso_a_Page_081.txt
8731790389d9eea64a5b0564598cd863
115895572cb10aa8b8b0e102541ed82179a9d56a
1049958 F20110318_AABZEI mabiso_a_Page_045.jp2
b189e219647916d6dde3673baefcbc16
66147ddce5514f7afdc065ee6f2ca5a62abf2726
6831 F20110318_AABYAT mabiso_a_Page_137thm.jpg
cb877dfd288441101e8cfc5086175046
81bae35cd235b1baab9f5c8a9d122f9e8556fa00
27516 F20110318_AABYZD mabiso_a_Page_093.QC.jpg
b4cb3afe2fbfc56f7a26b0b24d65c5dd
90c309f1b785b3f247ebc841dacf434ac22d953f
895854 F20110318_AABZDV mabiso_a_Page_032.jp2
33c635ec8f071d224462afdf4e8d5415
2c9796ce8aaf2424812fe94e4c57d34d1ed1f8e2
83702 F20110318_AABYYP mabiso_a_Page_086.jpg
5e3812c4c29c23ad079490daa83077ec
f199f53ec8985223bc688f53b1e4edc1bcb5e6b1
F20110318_AABYBH mabiso_a_Page_032.tif
ecedc3ca0305e0c054506819f6a147ee
6cbd8d643c6ea94482ff99adc82bf9d3cc4af5dc
78645 F20110318_AABZEJ mabiso_a_Page_046.jp2
0a233dc19b69676e38dd5bbe08672047
c67821e9c3bc64587c57b1ad84226c17238b562d
2235 F20110318_AABYAU mabiso_a_Page_100.txt
3c7edcc91404ff3deb4c7a3b1d2463a6
11b58df375f7e73ef58859a8f38c5fe52dd39b51
82502 F20110318_AABYZE mabiso_a_Page_094.jpg
1af22b6ce734611fde9b29e1e91319de
e0acbca13acd8c56fa5e608e5839b6e4e0897c76
51713 F20110318_AABZDW mabiso_a_Page_033.jp2
1cd1ab3db4501d07a2a41b9cc9ee5adf
307f353ec6cc957922a8b8b026064ef4b8f1f339
F20110318_AABYBI mabiso_a_Page_140.tif
351717db5aa662297e55871d19547b95
b692f66a90440a7d014ca764dfa76f18c90e6c16
1051963 F20110318_AABZEK mabiso_a_Page_048.jp2
b033eb46a6c79ccb58723a43c52299d5
dfd5f3b9021bf29d6d3cc97d2cf488af2fa4e8ae
1520 F20110318_AABYAV mabiso_a_Page_111.QC.jpg
34b44b6670524f14c1fed5c71769e765
e810449e9cfa261824e2c1e93b8fbf05bbb6b7a4
25698 F20110318_AABYZF mabiso_a_Page_094.QC.jpg
2697dedf7a8a294bab4fceddc5a2bfd2
aafd49c2280d521d019388bb346ca60cabe1f5b1
910711 F20110318_AABZDX mabiso_a_Page_034.jp2
7867d52e88b5a8332f119a2c6007c901
d7709122941810bbafd73dec523f3c2696d41adf
85858 F20110318_AABYYQ mabiso_a_Page_087.jpg
2ca9ac4ceb115c42157e25d8c04a0ec9
032845c18151d0c92a8d9c2cf9312f168ed95de1
108918 F20110318_AABYBJ mabiso_a_Page_095.jp2
293c48f0b7bc645ee9e7450b16a8e4f0
380ba2f5ce4dc9bef1c3284adf0245c6192d3891
F20110318_AABZEL mabiso_a_Page_049.jp2
21acf5c7c9024339f46ec32553655671
9de0954e2a7f4ac0490bb91ef0614be23f6aaad7
103113 F20110318_AABZFA mabiso_a_Page_065.jp2
f921ee29cc5a9dab3e6dcbc2a0e4a029
0ff6baf0fd120396f758824bf4b417b6f3e3a866
5434 F20110318_AABYAW mabiso_a_Page_044thm.jpg
673317292492ea3645b734a7457621db
ca1e276e61041077ccf2e5855f3eb6131b6d613b
84165 F20110318_AABYZG mabiso_a_Page_095.jpg
885c9476fa56dd359a496ec996e0bd78
e0900f91ee3f0310d389545273f0ad210fab5029
1051971 F20110318_AABZDY mabiso_a_Page_035.jp2
f482e1f4a0a6819c5933ea0fea0e0736
3c0c98484cab499310a3b89355b3882883df07ca
24810 F20110318_AABYYR mabiso_a_Page_087.QC.jpg
650f239dcff146eeaf79d88ab7b094f7
8649f096300dc862164dfe2163a320a9b17a20ec
22469 F20110318_AABYBK mabiso_a_Page_013.QC.jpg
4e6549a97e5dab0a8f1915dc6cd66aa3
c08681f30adbc9b2da35e5bb949bad523575909b
1051899 F20110318_AABZEM mabiso_a_Page_050.jp2
5fe0cde9cfd5926a02db9143e8b40dea
e554f72220579578892f0aff49e5a851856bd355
820121 F20110318_AABZFB mabiso_a_Page_067.jp2
6d4349d863004fc86779b2c420b9b941
d189991ec24cb22917275ebb3f02813cffdfb068
20195 F20110318_AABYAX mabiso_a_Page_067.QC.jpg
0041dc5e6e31f99d9670dbfbb75f5411
53ea550c91d30ee9a9e97dad47710a4e7051e6f0
26004 F20110318_AABYZH mabiso_a_Page_095.QC.jpg
2d0345498874ff10a90a6cf6e26bb1f4
ba485d8e2a9c8056d8f893e5814b5a28bbf02e83
F20110318_AABZDZ mabiso_a_Page_036.jp2
840cbea957e8a2373e22efc990230b53
5e2c9642c9a83a7125ed6656199b4558aa04c99c
85453 F20110318_AABYYS mabiso_a_Page_088.jpg
f629f6b6d3073c8c6a0b9df5dc033c72
b394e9fee7f606fd0e0a5273a2001e4996390d31
F20110318_AABYBL mabiso_a_Page_070.tif
6a13293fb3abdd6c3de982ab218be09d
25bf37edb5ccc409f97079098a5ef801aaa6d4fe
1020178 F20110318_AABZEN mabiso_a_Page_052.jp2
b661005f70effe608e6152ef9df09def
07cbb330b812e58a3822171901ace6623e1470b8
61689 F20110318_AABZFC mabiso_a_Page_068.jp2
3fe39a76acf395a0d4efe4a6964ffae5
4e2b7eb1dc6f01ac45c749b0194eb7f4075f31b4
80098 F20110318_AABYZI mabiso_a_Page_096.jpg
d8ea4274ea6197a4bfc1201e4e8441e2
2e668e60c40433ff5f50decc6589ef3b54d622dc
26330 F20110318_AABYYT mabiso_a_Page_088.QC.jpg
7b921dc89461de1638738a89c8fb95ec
bbbfee114a2b18eec1b5306b006f8ff9ac4900df
901135 F20110318_AABYCA mabiso_a_Page_076.jp2
6d5e9114db2f6c3f8640da3855d23b7f
369e83ec7dd26b8147ec576bec4bf317dce023dc
2533 F20110318_AABYBM mabiso_a_Page_139.txt
bf1b9b6bd405bcade917ce57e246ceea
a075f90d010ea1be10e14583147dc84d115378ff
969895 F20110318_AABZEO mabiso_a_Page_053.jp2
f5174190fc93c7d78dc632e316a2222c
84b2615207c8b2367441c1c0b77154f8f9ba9464
61770 F20110318_AABZFD mabiso_a_Page_069.jp2
00737c3f3dc79cac7993e61f3dd14001
8d957c4b8b1973ad5ed7b68db7f414e839f0ca7b
554554 F20110318_AABYAY mabiso_a_Page_131.jp2
d0123d33f566b18a13d95eec671036b4
066bb8112e59378fa9a990bb9d95771e8dbc1a86
24320 F20110318_AABYZJ mabiso_a_Page_096.QC.jpg
ddbdd46583f60c39d700dde115fdd4ce
f1eb0eec2e88b0f2e7e692c4a183535c21116eac
86797 F20110318_AABYYU mabiso_a_Page_089.jpg
13c6744a2c2948b661c11b2588c9ffc1
199aff6290ca908e0f527a5e7fae64c36186b896
87911 F20110318_AABYCB mabiso_a_Page_012.jp2
402f164a91b8b90fc3a45923c86eba3c
afc69e5ce5c8fc3167ba46ee2ee072cdb68b7286
F20110318_AABYBN mabiso_a_Page_110.tif
e66dfc89a97c1af2259c26cce4790e77
74d8f489122ad1ae67b246a308f0a450c66b5bc0
787323 F20110318_AABZEP mabiso_a_Page_054.jp2
a73552291a90b347f4b2557696cf8bc4
dcce07a14400d550eef37a32db1304ac0e561b7a
104768 F20110318_AABZFE mabiso_a_Page_070.jp2
910bb6440d610facda0c99c51ab03f84
0381f36b8d9feac71fc3bf65d3a895eb58bd8092
60152 F20110318_AABYAZ mabiso_a_Page_046.jpg
6ecfc0c2b4e68043cafe96394f118191
1170791ee69d4e987829a4d0203d2797bbb91d7f
85623 F20110318_AABYZK mabiso_a_Page_097.jpg
b5fb9cecd85120d1b0d96ec1810f2930
3a9153ae30e1fc612b2a195f5934899963b79035
25600 F20110318_AABYYV mabiso_a_Page_089.QC.jpg
04e0f34ad3244987ecbc7b7b7f8cf33f
4e13753a11fdd6a17345e3f02026f04574b16869
16795 F20110318_AABYCC mabiso_a_Page_134.pro
64b47c4f5a90d2c276e139901f9cfba8
8fd6cb2874ee85ce7a2fe0b4e8d47ebe0bb1d191
23175 F20110318_AABYBO mabiso_a_Page_047.QC.jpg
cd80db3db6c8712e7433bd7d9ec61992
071af971312bbab258faf1a2a36aa8f4b271abe1
94892 F20110318_AABZEQ mabiso_a_Page_055.jp2
e27c4b15af9dfe290dd5858151aed536
5853251b06deac5702faeadfc1d0acff97e80b40
1051969 F20110318_AABZFF mabiso_a_Page_071.jp2
d00bac4ce232cdeaaae4059905f4e6f0
dac31352afdad0cd200459061ef3fd3f23eb7728
71724 F20110318_AABZER mabiso_a_Page_056.jp2
4bf68a6c31b6e8e548fff49203970b2b
8c9936bb7794580e74e6fdaa2e4ab60c4923086f
27251 F20110318_AABYZL mabiso_a_Page_097.QC.jpg
0db9d38731e395764f53ade0dbec344e
dfcca1f6cf5435b8537f36dbdffef03bcb8ae549
43705 F20110318_AABYYW mabiso_a_Page_090.jpg
3ea3a412921fa392b49560483deec8f6
189fab48915f59a443ac62573c663f0a24918b8f
54024 F20110318_AABYCD mabiso_a_Page_113.jpg
a41faf7d693637d8c3874a406a8afe9f
6bf182f1fcfbc775f4b2075a39dbf9bb285b5f58
F20110318_AABYBP mabiso_a_Page_081.tif
471586ce54cf8372031e301fc61f6880
b138ed15e293f79a73bba83af80429a2e62b8976
108481 F20110318_AABZFG mabiso_a_Page_072.jp2
37ad7d116f88a12b97ee9cc50101d2f6
6e44486580db2b94ef094625c89e91371bcd5da1
837202 F20110318_AABZES mabiso_a_Page_057.jp2
e9721780b5a3f034091efb1c79b1bf38
c4e13899b92825fe9edd714052e3f7e3024e2d6d
74857 F20110318_AABYZM mabiso_a_Page_098.jpg
b5458131bd2a2571fa8fb65798398c4f
f7fe04d35316b2d300d304e8d579d32e680d09bf
13934 F20110318_AABYYX mabiso_a_Page_090.QC.jpg
2bdc575bfd4d88188a7bb36ae84cf054
77e7c8987d93dc994a2b7b336be59d2e58e5fcaa
85017 F20110318_AABYCE mabiso_a_Page_049.jpg
22d4d5d14f91ad1a7b11340c6e53b048
eebc14b7e90605b4a0213ac148ecc7c2cbadd510
3047 F20110318_AABYBQ mabiso_a_Page_087.txt
125eda5629107d816ce63aa4e5a2d9b2
1d17ba23e2708233eff20a8defce5ee2be95026e
117052 F20110318_AABZFH mabiso_a_Page_073.jp2
def81a9c42662d1d02eb21017ebc358e
7fd69cac93b918b3cfb9e89ee695d027c8ed8351
978104 F20110318_AABZET mabiso_a_Page_058.jp2
be8fe483d2edb1f4d40e0e14972829e9
dffe6f306eab442b22cd960c34eb25d70d537a40
23149 F20110318_AABYZN mabiso_a_Page_098.QC.jpg
a42f7693ae78e8245cc8d1f933f64391
bff8c6dea1fd30bd7cb5ed3b8f4f7a0d8d99efae
69070 F20110318_AABYYY mabiso_a_Page_091.jpg
5015087ff4e164edf9ef714d7eba86ef
3d6274d729ae219ade9aeea62b364fe30608d068
63312 F20110318_AABYCF mabiso_a_Page_067.jpg
a3ba193b3fde917d03bcdacb4ad7a17a
6b56d87fedf78edf46cb65dfb6eebb3b44693609
894 F20110318_AABYBR mabiso_a_Page_116.txt
6dd85aa815c47c8550e22e957bd118e8
6a7e2a43958eaab480a3a4ebea5f7d1861e18c7c
34845 F20110318_AABZFI mabiso_a_Page_074.jp2
d6d8f4b61e2b5ce814c83448ba3a06a5
c9915418c0169eb82106521062e93e9504876bef
19378 F20110318_AABZEU mabiso_a_Page_059.jp2
22e9091acdf87c8d906b31f4c285b1d2
6e23a02926393ac5fb5e015af1d364dbf1d0a257
34204 F20110318_AABYZO mabiso_a_Page_099.jpg
eed9fafc62a0a697c82223940e0ec64e
21b2d9f599ac843a36a38aa082f02ba614bcca41
21399 F20110318_AABYYZ mabiso_a_Page_091.QC.jpg
684539c295293584dc1a9a25c97ea4e3
d815ac9f000454d2bce8745f0f9d34d2bee498cf
78474 F20110318_AABYCG mabiso_a_Page_080.jp2
9bd42e339cefad05c3ece849a7a265f5
c93434ab45ec1735b63e77a31c6ca050d1ce5fc8
2054 F20110318_AABYBS mabiso_a_Page_093.txt
1d7a9a83b3791f9fb32476800c278c5a
562819f0f818f27b787d78440a0d5b76cd0d3586
933604 F20110318_AABZFJ mabiso_a_Page_075.jp2
88681a8599c13e524b5a4af1fd309ffa
59af4fc7afb565e635aa639476958845daf0f43d
67303 F20110318_AABZEV mabiso_a_Page_060.jp2
4ba2ad7f92b8c4c30ad2edd42b9a7456
b6cd84fe275c631f853e2ca2b25d73322b3de6c8
10592 F20110318_AABYZP mabiso_a_Page_099.QC.jpg
2e763089e6e7dbf883101dcd60660ece
d19b942b7e515c73f1a6ed0af93af102afb135b0
F20110318_AABYCH mabiso_a_Page_028.tif
45a84f90b72a7da1a9003858f17fead8
05a3923ad3c7047b8cc8fd4c211020cfe9cfcd2d
5826 F20110318_AABYBT mabiso_a_Page_136thm.jpg
5acbae092bd08cacb951bdf0be0a536f
b6197d319bd431cbf8d56ae214c5e78746074d51
70979 F20110318_AABZFK mabiso_a_Page_077.jp2
062d56aca9ab3ba5910c5559451140d9
b3c09241c473afcf9950e9a0c8b0ff7c8b8a61c4
1016753 F20110318_AABZEW mabiso_a_Page_061.jp2
1b85d658260cc1cc109a083ab83b7e5e
20c417f43ba76635031ee716a647c96272d59f2c
87809 F20110318_AABYZQ mabiso_a_Page_100.jpg
ea2eef9f12e22b8a8398878ea42d2283
af0bb29ed4eb18f50053e705c538f12aacb2a758
1051929 F20110318_AABYCI mabiso_a_Page_024.jp2
7543b2f269695db1eb9eda04fb9d7a70
542d06df72bc645a423bb4d2920836a62c75b55e
5995 F20110318_AABYBU mabiso_a_Page_055thm.jpg
c847695f77b1775d0e66b334366bdb6e
bb400770de69505aefb5d92c21605be2c6379547
935347 F20110318_AABZFL mabiso_a_Page_078.jp2
c3a946cc391a3a276021bb8ebe9bb20e
767557e630b9438be573e7f981e0e0b1b196dc82
105749 F20110318_AABZEX mabiso_a_Page_062.jp2
548b6109e288122130e01874b5796253
3462e1a5e4846d0c7b415b7ca13074f91ef07320
29634 F20110318_AABYCJ mabiso_a_Page_141.jp2
e0e8572882a8456719320a46efb23dad
10eb9d3bb460e612dbc9ff501c17e25e860bac84
F20110318_AABYBV mabiso_a_Page_002.tif
68c3a7db56dbfe43ee57b124cd216cc4
43c1b83b31239c0e21844d4abd7d9afdbe26a1f9
106972 F20110318_AABZGA mabiso_a_Page_094.jp2
e53b3374bbf6ada8ab6f9a041a0bf729
ab2b5a0d9c4dcc7a2d513908f9c1bc9aac61a6ca
81940 F20110318_AABZFM mabiso_a_Page_079.jp2
ae0c868b0166721246470808fd837ee4
3c8427973cd7cd509632730badea7d5a9edc7e97
102645 F20110318_AABZEY mabiso_a_Page_063.jp2
db16d62f33e45fb64a680c9978a85c7f
d12f630685de5b10e4afcb1097f297074415bc74
23074 F20110318_AABYZR mabiso_a_Page_100.QC.jpg
f88c9d1a37461721bda544cccfd7819c
71abe6f6ebfb73ecdca518a6328bd09056b1fe26
F20110318_AABYCK mabiso_a_Page_126.tif
48fe8d48e36a216d21ac31ac35dbb148
3be937770645a1c9a8ed41a057899645df7c0c06
5801 F20110318_AABYBW mabiso_a_Page_072thm.jpg
f55945e0076325db110d404f3616a174
9ab648f1532e130335c547882f32cefffc763303
103990 F20110318_AABZGB mabiso_a_Page_096.jp2
894f7948a53f1427164d9ba253196fe4
782416db32bf7c3c07fb99fbb015317449329505
63249 F20110318_AABZFN mabiso_a_Page_081.jp2
e313bd24a4e403b5beca9cc06983d19f
baf55fa65e54dd240834cabd3618d8e49c6dbaad
F20110318_AABZEZ mabiso_a_Page_064.jp2
515d2723de1425c7877df73bf808133a
484fa9bd401efa7dd753f5cdee4d1fad9489bfd9
108247 F20110318_AABYZS mabiso_a_Page_101.jpg
e91062e518032c544823bf172c5e28fd
114253ccbfcbbbf1ebb5365ba66c11063fae0821
153478 F20110318_AABYCL mabiso_a_Page_105.jp2
0602caa85f318cd25f4860c38fb8b0e0
9dad660927c46dcf527f36dc2b78db898af4946e
F20110318_AABYBX mabiso_a_Page_066.jp2
b7c83b684658c6f8afe0c471924ec690
f1fa6d0d5fb6697398d783916841d0e0eea0a9a9
111457 F20110318_AABZGC mabiso_a_Page_097.jp2
4edda2750cc073d1069a0607dad01c6c
9c1979ff3d737a28bd259ff2e56e609de6af6c29
91423 F20110318_AABZFO mabiso_a_Page_082.jp2
bddd23416ebfa9186ffb18913ff597dd
8d66cff4190de0440eace5ae9c930affccb0945b
28202 F20110318_AABYZT mabiso_a_Page_101.QC.jpg
88e43257fb79a7474b208fc2ea68beee
d334aaba90ffa0c26306bd6f67cebae0fff963c8
28366 F20110318_AABYDA mabiso_a_Page_018.pro
c861b238812d681073ada480ba7f4e07
a8272c3d5595ebe2e2d66cda7ab96191098f27f4
49354 F20110318_AABYCM mabiso_a_Page_117.pro
c19f343e658b48b9801b325c5a52dd01
5814d3c2e377389c38e9ccbea7263bc2edbaa727
1133 F20110318_AABYBY mabiso_a_Page_018.txt
3e1290f651478d58ebfcea9aed55768f
8d65c1f8fbbde8b8dc7e239be68993e2b923bc24
95714 F20110318_AABZGD mabiso_a_Page_098.jp2
76445a11b1f5d6f6c31748f24b5a978a
67b8cb772f8a75b8a5b2875c7be53291f9697641
127704 F20110318_AABZFP mabiso_a_Page_083.jp2
97c2589682f5702fbc392ba32a4ffcd9
cde758ef3472858d7ad12f90efaa508d997b9d1f
F20110318_AABYDB mabiso_a_Page_116.tif
72748a66b2cb3ae955d67b2c4047a625
35e799676ed0ee43b194e57ef4af9c2fe4de198f
2865 F20110318_AABYCN mabiso_a_Page_104.txt
2944306c9955541a1b41df715ff39b7d
c5abce8446d73066917b55eab37e869bfa95b383
77119 F20110318_AABYZU mabiso_a_Page_102.jpg
e670d9c474af1768ae90ecc4b555d964
6e52bf19f43b955d01bed8e1cd9c24b2c78176c0
112699 F20110318_AABZGE mabiso_a_Page_100.jp2
d53af1d0b1976af73c3fd9b0f726e7a1
960ffaca20858504b1a4bceb5a4040be58236099
107713 F20110318_AABZFQ mabiso_a_Page_084.jp2
19c9ce386173ebf568bbe616f65d172b
bb9b9bea401ad9e55df418eccb5822af69421a4a
1134 F20110318_AABYDC mabiso_a_Page_111.pro
1e12c5f3ba959b3ae4e4abb175df3792
61991151199a2eac771602e6d15f039c931bd2ed
1535 F20110318_AABYCO mabiso_a_Page_118.txt
29fd3f0ec0419adaab8f9dce56dd969c
1ef48b7b402bb26392c6d66c7fe18963aefa7564
70000 F20110318_AABYBZ mabiso_a_Page_005.pro
8484f45dfc8728db8d8255761fd7e9a2
9ca71e59fe2fc278cdf125553c5f2f144a8d95ae
20202 F20110318_AABYZV mabiso_a_Page_102.QC.jpg
13db5fa50055012efa8920a694798c9a
c9e9415e2709e8a4013ba8aaff262cd6e30f4354
140750 F20110318_AABZGF mabiso_a_Page_101.jp2
bfaf78b6868e03a140dea59082fde695
dd02ee8aa99c7cc62cb9279cf5a1bceaa0a99e24
110672 F20110318_AABZFR mabiso_a_Page_085.jp2
f062fedc24dfeee894ee3994851fab9c
f1ebec24f5383e9db15cf3d51a91fbbcffbdf0d9
114416 F20110318_AABYDD mabiso_a_Page_108.jpg
ff7860dbe45a7f68cb58190e6fa9ae96
5211fcb604cc18835473356e265a75550cbb3bd4
1828 F20110318_AABYCP mabiso_a_Page_061.txt
f67c118d3dcfdd87e75a1e75cdeac0c9
be04faff140e9f73433b01f8504459514646c00b
75150 F20110318_AABYZW mabiso_a_Page_103.jpg
83f9a9f0e9a4edf26e6320589fd1d147
fe2ebf6191b3ffdfd7b2e7da84e26646536f4aac
97125 F20110318_AABZGG mabiso_a_Page_102.jp2
41ca604a9d6efda9590d63f2778b3514
f3c2642e515c5c222ed073f0e732cf3b65ef152a
109450 F20110318_AABZFS mabiso_a_Page_086.jp2
a775f0b031fd05ea2023a06619729253
de5f68693079ab00de2a5a87bf4ee04e95e14307
1957 F20110318_AABYDE mabiso_a_Page_016.txt
6ed74c4457ebb43e330da9132ade0a75
a50c546041f464c24fae354cb9facd75a952fe11
10754 F20110318_AABYCQ mabiso_a_Page_132.QC.jpg
e960b37358c54cde9aea58e0f5aad0bb
23da98b8cca8087115266231579c630f27f1af51
20322 F20110318_AABYZX mabiso_a_Page_103.QC.jpg
3056c6d18129c2e5d563350bb3288fc5
d9fb8ad80dde9370f60e56209d0f70e8ea515555
94619 F20110318_AABZGH mabiso_a_Page_103.jp2
65258708865faeae6f25c1f0001d37ae
06d7c9bf05e8e8387e43b3d799b73801daec17bc
113934 F20110318_AABZFT mabiso_a_Page_087.jp2
7b2b0f2dabc19838ad1680d1f1b91749
7a509ffd51583dbbe0f11d4b0efe0af2bd444e94
F20110318_AABYDF mabiso_a_Page_059.tif
8858ef25d1071cdeaa28b94e02d7f881
a65881bf3e243c1e60d83211fee0444a03c3e9bf
45992 F20110318_AABYCR mabiso_a_Page_041.pro
51b4b7d6f1abd7f6137d3a17c625ed39
3556ccb8eef5736e74909beae400a01986d6ae48
111102 F20110318_AABYZY mabiso_a_Page_104.jpg
751874e10b98abe4a0eb85df8607c4c0
affb7511af6b1822c67656a3df9c2be8ca719e1f
F20110318_AABZGI mabiso_a_Page_104.jp2
9ce0720f7791c3179427c60635b1176d
735294cb8b513dcf7d74c077d250e2f79a5fc17c
110567 F20110318_AABZFU mabiso_a_Page_088.jp2
6cb0aa2a0667bc10228a679917ad3568
c748cf965151cff4b79fe960446310c2813c23bf
59774 F20110318_AABYDG mabiso_a_Page_110.pro
70e0d2f93ef3b5d81babfc556d21a8dc
6289b38ab0f0a012e9c60d24bf360d9d9d3a46a1
F20110318_AABYCS mabiso_a_Page_064.tif
bd0918c2260978d757a76f6e43109497
43dd51a7694eb5f1c98c24ff32eac64be132f23b
28015 F20110318_AABYZZ mabiso_a_Page_104.QC.jpg
716cab7f4719a1b1f077bad5bdff1e18
e2cfe05cfc3b8a98312aa774dccd321e3e08a998
153887 F20110318_AABZGJ mabiso_a_Page_106.jp2
7ae0cb2115e5b61bddf091ba03bdc563
c2f308c1c451532bbb6eac38407efbe2f11ab4c0
1051957 F20110318_AABZFV mabiso_a_Page_089.jp2
93cccf2e5d30718404171896cee4aa26
01a94a091fe92bb8e0a769face062ba2dc98f115
1054428 F20110318_AABYCT mabiso_a_Page_068.tif
e0f651b78852d031955e7d7b453e95c8
b079f3a83663725eaaa389f105cd022f95e08c4a
6665 F20110318_AABYDH mabiso_a_Page_104thm.jpg
b78f211acd0074580339953507a08e74
cfe6342043b7be8d51f3edb818d2536b3664e9ae
F20110318_AABZGK mabiso_a_Page_107.jp2
7c26669a9b0ab0936aa880b83300190f
185e457b669931b5a4abdd2ea597f60ab6a86256
55766 F20110318_AABZFW mabiso_a_Page_090.jp2
5084fc48f499dff1d1db2d29e361c460
9e858934f0ee6bb0f9b21a460c41873567c0a470
2809 F20110318_AABYCU mabiso_a_Page_073.txt
01a0fcbd2b231331b254bdf7c1258fef
a6c00a1a1594d995f43334375d5e20a66210f0fe
69375 F20110318_AABYDI mabiso_a_Page_044.jpg
53b5f0ebecf116a048f7a1c954663d80
4f21b57ffdb1bddb686a5c0b030a2db4f01ceff3
146334 F20110318_AABZGL mabiso_a_Page_108.jp2
846d7844f00281d61bedd7a044135013
e8086c4f02fb32d204885067c0b6cecebe43eff8
87373 F20110318_AABZFX mabiso_a_Page_091.jp2
6b863394ced10553479fafd4ea8907e0
b8ef9ed3786d26dead3501f6d2c1af0dacbb0b2f
F20110318_AABYCV mabiso_a_Page_090.tif
f86d8644ee426d51e301fc7d453068dc
ece56d04b5948f0b3094fc0c9fd06e2436064d5c
5749 F20110318_AABYDJ mabiso_a_Page_034thm.jpg
5b0b97a9d9b0332a095f7e6951dcbcef
9aaa645412659978e50abb4d1cbbf31058fc8032
76238 F20110318_AABZHA mabiso_a_Page_124.jp2
bddecfb09da23373e3aa6074a8c635be
1c8bb68286eba29cab2e88edfdef52a28c8cf49f
164900 F20110318_AABZGM mabiso_a_Page_109.jp2
4bd7006d91ead29f564cb6c00c756717
421479fe2160370216ab8653ec8e8e5f2867ce2e
107777 F20110318_AABZFY mabiso_a_Page_092.jp2
0d868fa015fed71a7711dfafc133f0a4
7e01cac002aa5e80a0b84ea8326aa2a7e482ff11
27099 F20110318_AABYCW mabiso_a_Page_037.QC.jpg
199475750961681e4eb9a53a34697dd3
ac67cf29ec5028193871529ae1bf46b99189308c
9353 F20110318_AABYDK mabiso_a_Page_051.jp2
0bf2abc21add5759b260950c297bec4b
edfdd7dec23cc787e468ef9c6ee24342af78affb
7095 F20110318_AABZHB mabiso_a_Page_125.jp2
8d458d88bd342256301e8410d9341cf6
676ff61c78d636905298213ec28c70e811c05f23
120061 F20110318_AABZGN mabiso_a_Page_110.jp2
52eb83cb1a441dc21f9c8b64b7324829
bec497b19c30aba9a6bbb9172bae9e9f4cc536e7
1051938 F20110318_AABZFZ mabiso_a_Page_093.jp2
5bbb992b4924179ac0d578de343abec7
06159453454c56f75f606143e942214908fd41e8
357 F20110318_AABYCX mabiso_a_Page_003thm.jpg
218dcd9f267384c91d1805bd09402b85
3ecb157591adda03fb9af47232b66d3b298b505d
5512 F20110318_AABYDL mabiso_a_Page_075thm.jpg
30be37f968021e919abfe05e3561ecc5
17113e6ca4777725bc563baf00ecba8372d3e820
140095 F20110318_AABZHC mabiso_a_Page_126.jp2
0ee42e7d3d79ca531fdaa66ea8b45c5b
f8368588301c509f8b7a3ef0e82a4a0476134b27
5669 F20110318_AABZGO mabiso_a_Page_111.jp2
2204bf5a49a82cd19384106d1e499706
2bb12f434d2aa1f3c930b646308086fad00e7d1b
26738 F20110318_AABYCY mabiso_a_Page_021.QC.jpg
58d92927b79a2eaaf4b25850b89cbe38
25dc5b63aa78a3dae257d1415c0222eca1bce5d8
56542 F20110318_AABYDM mabiso_a_Page_119.jp2
61d211d37ccfab9072fb60d4a4c5e274
7143a0ec8210889ce27b12edc46ebeac1cbffd4e
66732 F20110318_AABZHD mabiso_a_Page_127.jp2
0c7ac2121afbd245e4c1c9735de6b9f6
b4dde623d5039eb066584134d6b11b329571189a
74687 F20110318_AABZGP mabiso_a_Page_112.jp2
07a040508eabc1a2dc82eb08469f941b
e238aee28df57f8d574a857443e3bd6d0ba5f8ad
1931 F20110318_AABYCZ mabiso_a_Page_024.txt
daa93feb4f8d868999b15b3084aba615
0cc74aa3c0e336d692a2d82fc0954f2aa17524c7
F20110318_AABYEB mabiso_a_Page_001.tif
e5072100b2bd866cb5e3f32e3ef0316c
1879d5b848616bef2c7a29f7686ca314f299cef7
6690 F20110318_AABYDN mabiso_a_Page_108thm.jpg
ce4c231845eca8c3fda7cf200f753822
a8be5efb9b8dc204f784663fc25a3a42aa479796
795543 F20110318_AABZHE mabiso_a_Page_128.jp2
3122894762c00809c07ba342d89f31a4
cf1fcef5c7e234535ddcda75b41f378d7bc7d568
75408 F20110318_AABZGQ mabiso_a_Page_113.jp2
9da923ace7d8d9a266e863b479031232
ed33302d08e4e69c43df43609c2c70a6a8f873d7
F20110318_AABYEC mabiso_a_Page_003.tif
74beeb5a3ce03a4b5a6dcb5885bd7acb
411b54d0c57a665ee7877e3f9f37f01b7c2cc8b3
F20110318_AABYDO mabiso_a_Page_135.tif
25cee9931d19126a9514c293c7ca7250
862d01ad1abf36db8c43157d0ec05e2a3fe26609
575306 F20110318_AABZHF mabiso_a_Page_129.jp2
91560636cc702eac10d08ee9016676b0
f45088acc34e1a67264782f2df03c228d21df126
97375 F20110318_AABZGR mabiso_a_Page_114.jp2
c2173972750cf034516eed59526cf603
11280d92532304307eb832ee01f6a7b177abcdb2
F20110318_AABYED mabiso_a_Page_004.tif
d754a3b7207a35566fdcedd368ef21f2
b02b010a4f994ddf587ea906239b0a7903722fd3
27913 F20110318_AABYDP mabiso_a_Page_022.QC.jpg
2b5ffceff9e3283119e0b84d6a3c7204
f4b457c4b7a8332afc6faf976c47215bb4bafc9f
767701 F20110318_AABZHG mabiso_a_Page_130.jp2
d40cea5939a307aafc3f5b0d88857fb8
4a216d78200a7727e57f6b1c26434371d8b3e42c
102398 F20110318_AABZGS mabiso_a_Page_115.jp2
584d94ae06a21cfd60b2c74f27480c6f
f8c5807d7235ca9be13d16668a9bcf5e71cb1492
F20110318_AABYEE mabiso_a_Page_005.tif
ea84accf8381e6c8d2cd70c6b4e9bb04
d1976b052b59194317d5d1bea61d0b283a5b0b89
2335 F20110318_AABYDQ mabiso_a_Page_069thm.jpg
2ccfaca04182aa9b2e71d705f00084e4
78cb6d72519f2d3ad182a08ac9763eac29a60c27
39392 F20110318_AABZHH mabiso_a_Page_132.jp2
fe838bbaa1255dd8c1f7320c12a6d91a
cbb2b46cc170792546672fbd55e8c79cefbe31df
38204 F20110318_AABZGT mabiso_a_Page_116.jp2
d51fb6cf04fa899b1ebdd4e15fadfe00
c250817e5d396dc1cd956aef0eff41470e39300b
F20110318_AABYEF mabiso_a_Page_006.tif
07adc7c15923d770104a49c1a21b3351
e488af71eb979b78a3952c40aa3c0ef6272f48a6
4324 F20110318_AABYDR mabiso_a_Page_083.txt
5ea28fbfec96c9874e134a05c4b6a815
a9c48fcd708a61228920c1e8954d246ba39bcdd3
41110 F20110318_AABZHI mabiso_a_Page_133.jp2
d28a875086247dfa68bbbe5080f53b04
c2bedb63f498504e707efda4332b557ecf579f5c
94471 F20110318_AABZGU mabiso_a_Page_117.jp2
ac4c3e092abfb827a1bb633bb3349000
9950655178cd17078196caaa064940a618245321
F20110318_AABYEG mabiso_a_Page_007.tif
2ce1fa4007e204cba05aef84cf7ab3c1
3243f8a7b8a53ee4962b5909c193605457ff8021
2013 F20110318_AABYDS mabiso_a_Page_009.txt
559e6d501672bcb1404f6fea442b316d
d136eb0d4cfb3ca3f95c47865ec2eb12cd085192
506403 F20110318_AABZHJ mabiso_a_Page_134.jp2
376188a75a9d7ea0e47090a7e634769c
e49eda88345b75e262add3b4ccfc8f30bedabde8
57834 F20110318_AABZGV mabiso_a_Page_118.jp2
32c1347f5ef7fddbf8f137547edd03ff
87fb2f92b0711d9bae26b9c8958a78b2fbf89741
F20110318_AABYEH mabiso_a_Page_008.tif
f4707eb88a5ed417d1b90f843a9e85f2
d216ecad267247ce49edfe5d43559058d6377a16
25686 F20110318_AABYDT mabiso_a_Page_086.QC.jpg
ff687566c3f7615542001e0180e26b39
f33e976266e147ec1338734594aab94ab33a5940
249392 F20110318_AABZHK mabiso_a_Page_135.jp2
0b7fcc9f5a5570ef592501ce04f8e836
a76fbad61801bc24c15b5fb285b6bdd291147082
111595 F20110318_AABZGW mabiso_a_Page_120.jp2
c5a75ef6c6c72d0ee67df6dcd19408d6
1497656799eea80719f89568bea59d5e25d11abc
F20110318_AABYEI mabiso_a_Page_009.tif
f681e73c9f5f71eb11b31ce29cc61a11
fa2006f65b447e973853cfc237d069e5afb937d5
82575 F20110318_AABYDU mabiso_a_Page_035.jpg
883321062c997bb0d094461d975d47f9
7d83f29f292e90fd14879c1d4e40fa992d67eeef
118763 F20110318_AABZHL mabiso_a_Page_136.jp2
ab4e7e98a089d68ac4a47687114bec26
3574ce7c8e8a97999850d8832007c92292cb8f0b
474031 F20110318_AABZGX mabiso_a_Page_121.jp2
237fbf9911907d8855fa148ff48a1706
ca817acb8297656d33cb664a26f1acdaf4fd1dca
F20110318_AABYEJ mabiso_a_Page_010.tif
1f464f793e33926941ea795e2013de0c
4e22a3ef94a3bc08934c18ba1c60170b267afe36
F20110318_AABYDV mabiso_a_Page_080.tif
b31a78dda37fed2ba72c88b7c9dba1ea
eddf90b4fb021e1a9cd13dec2693fe55bdb181ac
5091 F20110318_AABZIA mabiso_a_Page_012thm.jpg
ac4e105088817bd04777bf53125a5b5d
d1f05402315f89e4ebd244983dd6698a335e465f
136048 F20110318_AABZHM mabiso_a_Page_137.jp2
55928a9acc23d25abf20b6d3efd6f504
44b5e64a8575affeed3225fee34fc54fbc315b6f
465416 F20110318_AABZGY mabiso_a_Page_122.jp2
40011ce79ef8e2cb8545b1d8cac264f6
15767d778c429d718a7c04378cefd686771ad121
F20110318_AABYEK mabiso_a_Page_011.tif
f0970d7682ffd1129a1f93f2987c255d
95fbdb9dc73b3575799eed8ef910be5001a5124e
29998 F20110318_AABYDW mabiso_a_Page_069.jpg
0a4eee366824b3c10cf678e5b62ba9e4
84e99382e89534c39cfaf6d5bca465e190e46b4a
5464 F20110318_AABZIB mabiso_a_Page_013thm.jpg
22a5930f7e5e70f35900758999cebd29
19e5d3037365cd21fc1078a5a0a11af95b728e47
1051943 F20110318_AABZHN mabiso_a_Page_138.jp2
82b33b7972773fc153c8f49913696e6f
091b8164d05e4e9340fb144f25c4f6aeb0ebc74f
85186 F20110318_AABZGZ mabiso_a_Page_123.jp2
f6c8fa20daaecc80f2553bdc636c700b
1999f74f4d2ee40bcc48c88e4b7ba4fd6299d831
F20110318_AABYEL mabiso_a_Page_013.tif
3bd09929c2db1f523098d7cb4facef2b
73f34f325e394246204db6b0a694a0b3244d6005
1842 F20110318_AABYDX mabiso_a_Page_065.txt
0f41d4cc7371840266481d31ead45eb5
2174adc00448c10e01232b92f04724fca98ccefa
F20110318_AABZIC mabiso_a_Page_014thm.jpg
5100b05f1900b418a0e686ad953660bd
47882b3d101b1787a50702bf4f3a88c0acc0291e
1051977 F20110318_AABZHO mabiso_a_Page_139.jp2
659f4b662759f9be744edb351c448a22
24b6d077482e5ffd1f67b57b9bf01b0edb0c0469
F20110318_AABYFA mabiso_a_Page_030.tif
6d615a4e0525ecd94ae8f25c7197d174
73e4a5f68b1e44a7bdbe22f276733a87d4ab0fc5
F20110318_AABYEM mabiso_a_Page_014.tif
555cfedc2594e2f3ecedaf5d1afd6f6d
759ce9647841397d40632d89373dd23d6aa08900
227888 F20110318_AABYDY UFE0011841_00001.xml
633c8b9e82576988d684b801a2e6df2e
4d52cb6a99c52471a7396a57dd312ba1fe551b55
6369 F20110318_AABZID mabiso_a_Page_015thm.jpg
8e45d7a33592a9323020cee0b22f8c17
9dde06f0eeffcf4304c1146b270cfa9d45399242
1051973 F20110318_AABZHP mabiso_a_Page_140.jp2
2d6b21281075e1bb0f28b631f1a05c5a
a1a93009299597c0d6a7f0fbd2f8e3a9abb3c5b0
F20110318_AABYFB mabiso_a_Page_031.tif
46c05c0abfc4ec6e5dad3dd2cb78e610
cc55808c0e99a7ef4852e788894b9733c61cb35d
F20110318_AABYEN mabiso_a_Page_015.tif
2002e3ea30e9c30f6b44fdb0f66330f2
80af0896f4b85775b53f2f84f82b3d98d09fe9fd
5861 F20110318_AABZIE mabiso_a_Page_016thm.jpg
c329dc0493fcd87ab937c0e922b55914
bd9d711f688a84ef74bb2b5ae5035c69431b5c42
1854 F20110318_AABZHQ mabiso_a_Page_001thm.jpg
60912df8143e5467f512b277562d1ac3
34a6993e056312b3468c865be052567df03dcc4b
F20110318_AABYFC mabiso_a_Page_033.tif
26365c981465d241f350dc845e583f13
2e9992537bcf3e49099b5b6295172865d1ae86d0
F20110318_AABYEO mabiso_a_Page_016.tif
c04ad69bdbac241fcfa2376756299c3f
3a7c677a5a1a955dffb2f3bb69bef2c74fd50879
5815 F20110318_AABZIF mabiso_a_Page_017thm.jpg
416797636a985feb1a4caeb4d922073e
357651bf2d5a5119b54afd8d44f1e9288d24d1e5
496 F20110318_AABZHR mabiso_a_Page_002thm.jpg
9c8eafbe0bca2de65957fd8b585a21de
b8a6111a26bf8a3ce7478a03ca0eb67f326d9afa
F20110318_AABYFD mabiso_a_Page_034.tif
2f6a617f586b0aefcc5bbbdfb2481945
3402499c7773ebb079e38d501727468dbb5f0f81
F20110318_AABYEP mabiso_a_Page_017.tif
2429ec09080caccb4a99c6c4b492c624
3b7144e0a234ef308da9b967ab9ea50843e7a888
3651 F20110318_AABZIG mabiso_a_Page_018thm.jpg
819b366fb75ce760d797b34c4354cdee
f323fb035f10cfaa4fa987adebfcf6c2b979c036
5266 F20110318_AABZHS mabiso_a_Page_004thm.jpg
dff6fd353225c60259a22f6c8bb188f2
7075e411f996ee51d3b0d641e8d473daddeaa063
F20110318_AABYFE mabiso_a_Page_035.tif
06c703ef9a799f30ce519c96c0235dd7
68554e5a5750cd83b75e8d928d21da2d20fd34e1
F20110318_AABYEQ mabiso_a_Page_018.tif
4e6018e65a8173b858afdc7fcf43aecd
720cca9bcb1c0b6b1ad95ec9d98ea9972f4c6efc
F20110318_AABZIH mabiso_a_Page_019thm.jpg
9f0bf81cd8cd8043cb6c26433388df99
adcaf8a31efff544c4ecdd376742d950da994199
3939 F20110318_AABZHT mabiso_a_Page_005thm.jpg
d9228351a795b64c42b4a02384c06c49
95c924b0f47ec18b11d0074b80a81eb933df398d
F20110318_AABYFF mabiso_a_Page_036.tif
4459b372de2cd5302414fc9f595b54e0
492629cd2f7eece1d6128ebfb5c3f45a2336a68c
F20110318_AABYER mabiso_a_Page_019.tif
99559ffb6a33e4c9e0a5ff8220d8acf0
c6e69882a591c8ebaf430777d89a37782b0f172c
6564 F20110318_AABZII mabiso_a_Page_020thm.jpg
86f58c8037be69424ea25e3ceeac6dcc
4c7bcec647e069c6e1900e90c3b29c1f6a3ba242
5306 F20110318_AABZHU mabiso_a_Page_006thm.jpg
c7a2f89d5a0ce5c27514333b5b225e7a
0b3270c21bb0bd914df5d2e83e5eadc8ad2414e9
F20110318_AABYFG mabiso_a_Page_037.tif
dd815f76f1979b97d6d3405af6ab0b5a
115ec6699848cf708e1b4e709b622c6dc73a1de1
F20110318_AABYES mabiso_a_Page_020.tif
3a564ab7a4b99dfc116149ac483da12a
cfc8889af7eb2160f41e086e4699bb652afe1428
6551 F20110318_AABZIJ mabiso_a_Page_021thm.jpg
8dcfc51e26ff4300f6bc9325f7d88a31
484e1822ef7b75fe573c796fa1961dad6ac91e65
1035 F20110318_AABZHV mabiso_a_Page_007thm.jpg
68b40f98409bb8e46aa585e67ff53dbd
625cdab67abb5e5404bda43e0ff13102fa761353
F20110318_AABYFH mabiso_a_Page_038.tif
107a450af5f474eac80a36e3783fda5e
ec8118e3b5a988bb0bef0c1decec710d2d27d911
F20110318_AABYET mabiso_a_Page_021.tif
5c19f1b3201591a6f68d96328929e7ca
c6118f416be1fb947e1f5890fa79e3a1ff44ab99
6517 F20110318_AABZIK mabiso_a_Page_023thm.jpg
80e5b3db5d4c62b3304f13c20686d325
41335087a9b7dc3e49b061b5dd4b8c02a67e0d8f
5502 F20110318_AABZHW mabiso_a_Page_008thm.jpg
c5921ce6de1314dd82a7d84ab2ce69c9
d6445e43f2aed00d76a6cd261ca4797971c280c1
F20110318_AABYFI mabiso_a_Page_039.tif
3102d696da4e8ced6693dfe41f8e2d52
e6f9421c79b87e13b515f0a5ffc4ef943dc29b44
F20110318_AABYEU mabiso_a_Page_022.tif
ac620f1cce869c226acac710649b3cca
961d3b8f2eae5d9a12b5f29af8265a018ff9d768
6534 F20110318_AABZIL mabiso_a_Page_025thm.jpg
a9b25ef4ef77f072b232a4e283d27708
8a240b66eed4b3d36898fe78f55dacb5322f655f
5153 F20110318_AABZHX mabiso_a_Page_009thm.jpg
e59bc876c5900a6451fc60fa70916cf4
2faa8bcd735bff9b46f1e26683f2a941eb217448
F20110318_AABYFJ mabiso_a_Page_040.tif
950b91f132b83d9bc9d0174f37f1d70b
71a26136f43027da8f96fd10af9b55e32b35608a
F20110318_AABYEV mabiso_a_Page_023.tif
dd505067c860e1396985d7958c7ae509
962507a811a7f71343a4dca0e9946029b70b5f18
6136 F20110318_AABZJA mabiso_a_Page_041thm.jpg
843606d6fd6184d39ca27c0f4a3f8854
7a58abf647e66525baeb168c4640459faa724d1f
6555 F20110318_AABZIM mabiso_a_Page_026thm.jpg
35a6f55fe8d1157ec17072f36c39be0b
6feafb6a871a69e363f23499f38105ba18a9b209
3515 F20110318_AABZHY mabiso_a_Page_010thm.jpg
c9e9f3b8f48040e322482f0d53860896
5aa7201cb204db9667a687cac887ee806dcd11f2
F20110318_AABYFK mabiso_a_Page_041.tif
d5dabad28a3cad3b327717b66cc31ff7
387a8b2fc56a5d974cf4771146a363967b890f8b
F20110318_AABYEW mabiso_a_Page_024.tif
4634d3519411633f31888b568dda561f
ad6e4a53df41664e063d288079ecaef4851ea44e
6561 F20110318_AABZJB mabiso_a_Page_042thm.jpg
ea1f50eb56f3db191b7ac327138fe6d8
12a7476b6c2e2dc9a9163aa8b8a334c57586bc3f
6394 F20110318_AABZIN mabiso_a_Page_027thm.jpg
6e5bef59979820d3bcc779a063b5ec49
84611918797bb56a54cf4f86add18f001e461e0b
5319 F20110318_AABZHZ mabiso_a_Page_011thm.jpg
ab9655b9fc3777374da3d7d88ba1d866
1c1a2db0f2700cdd941461ade1bcf04b166b19c8
F20110318_AABYFL mabiso_a_Page_042.tif
4c05ddf82cc0b5aaa4470e61abca9f1f
48c6fbe89f5a9c15f0ab501ccfc94d223356d4a9
F20110318_AABYEX mabiso_a_Page_025.tif
a09e2602895413b5993f7c301c3ee226
cd3e3a9cd79acc8ad938d03e2a677dad95d3990e
5796 F20110318_AABZJC mabiso_a_Page_043thm.jpg
04c022b743c70af7c013ea749953e459
36ffe38ac1fe0391ca469e49a66ef4c95504d6fe
6524 F20110318_AABZIO mabiso_a_Page_028thm.jpg
7c6dc21fcb51e863662e7fa63170495f
e287d064b7f001af05d889d0a8113ee664efc9cf
F20110318_AABYFM mabiso_a_Page_043.tif
ed602c1efadd6eb6d2e66c83fc24fa11
84b9548950cfaa16ce12844b3a50293777fd6357
F20110318_AABYEY mabiso_a_Page_026.tif
d8edda997031f338464fd102c56c52cf
4da58aa536506332b3d5375cefc9994d1eca81d1
F20110318_AABYGA mabiso_a_Page_057.tif
361c1f49b844806b70c0a5abd9c38c8e
155840fafc4cd241d7702a972e06b1c31030b6e7
6200 F20110318_AABZJD mabiso_a_Page_045thm.jpg
7066dc971fe69a1f1403ed9143fe3321
2b01c63795141261929e6dc6351a7006fc7a96ba
F20110318_AABZIP mabiso_a_Page_029thm.jpg
dac61a84d37b8d86170c1a97b1172101
f7f204434cdc11c30af54a419638f5a0f08b381c
F20110318_AABYFN mabiso_a_Page_044.tif
5ae91a97a8693f6b4680c9d39461611b
224585cb5406eb744e60d97b3568cbe1afe45166
F20110318_AABYEZ mabiso_a_Page_027.tif
e8827ef043f22c087f8976eae2362293
7c25e55732f4f936db875ff1472eaaf3f1ca84f4
F20110318_AABYGB mabiso_a_Page_058.tif
7863ea93128ee2269786625448685d8c
967922dfd6b2ccb57660fa28e6fd126e35c0e986
4853 F20110318_AABZJE mabiso_a_Page_046thm.jpg
32a0c083e45a1a7b628fc478819cf8c6
698d146f681abbb21398073ffd9b1ba336093f31
3927 F20110318_AABZIQ mabiso_a_Page_030thm.jpg
84735ed801ab684982d2f63e6bcd17b4
1037b87c1d980c46730b9827e5544e500b8d4835
F20110318_AABYFO mabiso_a_Page_045.tif
841c22f0cc307a2772cafa4832f84004
ab4190cfc6b02efc4a03ebd839c106e64da829bb
F20110318_AABYGC mabiso_a_Page_060.tif
c184f2a392a7694dd92f97346dfb935c
585dddf497057d417b5a30b61ebfc24a78580e64
6099 F20110318_AABZJF mabiso_a_Page_047thm.jpg
c5162c26da8022c283b992471d17ade7
0237eff593962a3edd03e123d73d6125d9e0d9a7
5152 F20110318_AABZIR mabiso_a_Page_031thm.jpg
e9fa42906cba9349bf7c22703290f58e
81e3e81b7ae31011014efbac32268ce684b280f2
F20110318_AABYFP mabiso_a_Page_046.tif
96a34ffa939df50bdfb685396ceb6a53
71a7cb329cb371d33a7e2fff585e78623c2a4262
F20110318_AABYGD mabiso_a_Page_061.tif
e5dbdb6c750942c373ca386bbca73947
ef90220c8ef447c32d9a4a86fee44da27f8ae8a4
6546 F20110318_AABZJG mabiso_a_Page_048thm.jpg
4f1161ab77e6bba3858b2425a22f0fc2
fc854d64654f4afcd385de9254d19b30eb0922a8
6146 F20110318_AABZIS mabiso_a_Page_032thm.jpg
7ace0d93961e2c7e1873a188320d5a24
5c486d172ee8d464b064d75945278cf5660a0b09
F20110318_AABYFQ mabiso_a_Page_047.tif
fca6ee203cdc4d450754cdd5226cb21e
cb4d607477231f9e8d44bb942361a6c4a30fa15e
F20110318_AABYGE mabiso_a_Page_062.tif
35d79648b9b09004186c1211467c8cc4
6e28b24a724e54f11421085a7e81c1147ee46bd7
6425 F20110318_AABZJH mabiso_a_Page_049thm.jpg
ca5a27f502e1b812efaa70579345e67d
a1cfd07c669b77beed858de626ba9479f3cb130e
4123 F20110318_AABZIT mabiso_a_Page_033thm.jpg
59945f72d44d9ebc9fd669817230e91f
3f8e8d4e7f960ea7d25a9fc92b45e0822de14de5
F20110318_AABYFR mabiso_a_Page_048.tif
557c19db602401cac6456b0f4549d625
c29a624a01e81d86a57a8d07b622c599a5219629
F20110318_AABYGF mabiso_a_Page_063.tif
33eccd400b5b29582d337fae373a5465
025899409708b52f511cdfebc3d19127e63134b3
6552 F20110318_AABZJI mabiso_a_Page_050thm.jpg
dd6720ee8d86a585d2c3c82d2133ea8c
f2940c4a5cbb43c2aa38ac01c7b72161679522f0
6407 F20110318_AABZIU mabiso_a_Page_035thm.jpg
0c38ab2cc89c1f637ec822ec9e61e8e5
1710e6d1afc5cf7c6ec11674456bf9ef6dceecf6
F20110318_AABYFS mabiso_a_Page_049.tif
262cdabb87b43811ebf2a9d5a9c5a0c0
5f7485c35ff5dfcd8129d8b7e4b4495101b07317
F20110318_AABYGG mabiso_a_Page_065.tif
4f50c723e97d5e8cbbcf8ec576833784
82bf6e98c396008cecc0edbefc0a0c45cfc39282
715 F20110318_AABZJJ mabiso_a_Page_051thm.jpg
f78612423ff15f068e60327a2c1b90f8
f3509e6ae632101a7db1d13d93ef9b062fba7991
6627 F20110318_AABZIV mabiso_a_Page_036thm.jpg
62b5f6bf2fa3f6f7e77aa06d4870efe2
eb3243708dc4393ae59b4f1ffb15434d0439aed1
F20110318_AABYFT mabiso_a_Page_050.tif
12d0caac4e3b92cce66260303372b67e
b085965b8a680065a774e0d2c8f206a79af2b538
F20110318_AABYGH mabiso_a_Page_066.tif
f419ba24dbd82ba0e6b89c449d9cf0f9
9471f9fe4ca7764059778603669308d3d58918fb
5965 F20110318_AABZJK mabiso_a_Page_052thm.jpg
05eaf68079a4799cd84e755332eb7254
c2562a6f076e1c5f3d89ab3352f23e9b8cce02f2
6485 F20110318_AABZIW mabiso_a_Page_037thm.jpg
c18302527a95470f82ab4d14427e4bd3
79c9ab4121aaea51c6bedea8619b5c635c77c8b2
F20110318_AABYFU mabiso_a_Page_051.tif
17cc620b95adc7f1b12999b1afb3cf1a
8c7d39be059f43b73bb51c0001ce61a2b57fa28a
F20110318_AABYGI mabiso_a_Page_067.tif
585df49be783dc6f6a4f6c7114caaf14
a42eb6ee32e9fdd9afe5447ec6f75d72ed04734a
6484 F20110318_AABZKA mabiso_a_Page_071thm.jpg
895fa66f91c5c4fb765a12a5aed48c96
9562236435065b39e4397128cf11cf44bb5a0a0f
5917 F20110318_AABZJL mabiso_a_Page_053thm.jpg
58342f3bce7576697dc1293e3e9064ca
5965dcef9df09b23570820afecd4fea19da980c8
6584 F20110318_AABZIX mabiso_a_Page_038thm.jpg
73ed31f57989a983fb46f12aefd1526c
4facae597c357c079f27d9151d625212d9455550
F20110318_AABYFV mabiso_a_Page_052.tif
9e38d380892fd5b03972720aad113976
78fc799458a4232455e218d635a80bd11570db2e
F20110318_AABYGJ mabiso_a_Page_069.tif
45845383c2792e41b72aea9592e8334a
5233d51f2c3dafa19d224a61f0da309b6269925e
5176 F20110318_AABZJM mabiso_a_Page_054thm.jpg
15322b0e679e6ba51cf7e8ada87df616
a399ebb006e4e2972b04d9f269cde05f7397f5cd
2997 F20110318_AABZIY mabiso_a_Page_039thm.jpg
832ae8d9d8b637b0d041ecffeb8685c5
9b6b8e2c41e4e2ada893c96338ddf8de46dd5f14
F20110318_AABYFW mabiso_a_Page_053.tif
ddb6e873d2c7cd9d8ec995e7b5425c08
28bae859570802eb3471e4df953389750bfcdae4
F20110318_AABYGK mabiso_a_Page_071.tif
12eeb87b141f694726868af58ba021df
e4a1e29b67be232576fc3938d938cbc3a7788d4f
6198 F20110318_AABZKB mabiso_a_Page_073thm.jpg
0c4a1869a7c7564c957ee7af1b612a42
ccb063eaa9e3e810a7cb3c4c7941ee611bbfaa92
4669 F20110318_AABZJN mabiso_a_Page_056thm.jpg
73e3f7b279c7628c404770df32126cb7
9be8453d239b1a6368deba543f9217f765df22a6
5884 F20110318_AABZIZ mabiso_a_Page_040thm.jpg
a69af80459c1dfbfe7b263c2722205af
0e29eaa2314ac869e5e2e0fc9d3f75f5d3c12d9a
F20110318_AABYFX mabiso_a_Page_054.tif
9c1fe7b92ff99e43d8b521eb3c54c672
c5b331444e31c10d10290b078a888d8d72c2d432
F20110318_AABYGL mabiso_a_Page_072.tif
881d95903865d6997605d8c39673bdc8
33b6c465580e3835c616506d3aca6fc98b7e4591
F20110318_AABZKC mabiso_a_Page_074thm.jpg
9f6bf3a25df51dc86d755fbe3311e3a2
66366c23ce4e1ff4c333258922ff26806fc607c6
5252 F20110318_AABZJO mabiso_a_Page_057thm.jpg
a0e6fffdc71feb29e9386af6a2f9c1dc
5a7d8dc2911cf55ebc915070d4ef4fb44c6890a1
F20110318_AABYFY mabiso_a_Page_055.tif
9ec7f2951f9cdecc911dd22a643d4bad
c03ed4867ad75e1953c3cf5330404a2c7fd390db
F20110318_AABYHA mabiso_a_Page_089.tif
ddcfaf33ad0c8786a7f5e5d81d2b136f
bbf9579ace75f31ce83a6c2eeffe16fd41464a38
F20110318_AABYGM mabiso_a_Page_073.tif
b9485522474b4d02def8a45695fce314
d34c1823d7f8cd2f397bad0f3693329aa0c86aa0
5620 F20110318_AABZKD mabiso_a_Page_076thm.jpg
3d03151723b919cb1bfb13a2269b021d
c1e910b6ff05416733880caf3f67cbb485a85fde
6065 F20110318_AABZJP mabiso_a_Page_058thm.jpg
22ce2fa0e25bb7a782a1d0a122545d9d
e7b043cbd8ffe62c789e4b4d917e5450f32dc18b
F20110318_AABYFZ mabiso_a_Page_056.tif
3851789cb57ab4a160b532127c415472
2c9eb28aa528453bb042c1d4e8d6390524d62eca
F20110318_AABYHB mabiso_a_Page_091.tif
34b006169dc6d23b8c80f7696b27b27a
2d823b12574b7b2e2ed1124e17eb1c8ee3f2c88a
F20110318_AABYGN mabiso_a_Page_074.tif
ab06a0914c3833775cd7a091d9dcec21
d12866c5217e3b07eb9407fc637e9713d9568b85
4863 F20110318_AABZKE mabiso_a_Page_077thm.jpg
aca234ac5e47ffa09281a74f1f9aa86f
3e7a1ba5ce7dcd5c8cdaf06cc6b4e067e4385f92
1406 F20110318_AABZJQ mabiso_a_Page_059thm.jpg
c333a8d958158606b22cefa94254fcf8
8848964b77cad9f63d877d8f61361f54c27ebb53
F20110318_AABYHC mabiso_a_Page_092.tif
3d445633379840ea7b4575698c0145fe
9750dd08ee05df91c48b15f0802ac169c3d95c53
F20110318_AABYGO mabiso_a_Page_075.tif
313bf6934d2f5d5490b848ca55725192
6a03dea16c144ffcf8199f72fcebfa3cfed7db7c
5958 F20110318_AABZKF mabiso_a_Page_078thm.jpg
c98a599733f2f1831f6650e1ca944f91
63ef86e43211e445dde351ff816cc3a820043fcb
F20110318_AABZJR mabiso_a_Page_060thm.jpg
0018eca78bdc298bd39e0ee388428814
50766fca2556803e3e7151185bdb6b6a5ea1ece4
F20110318_AABYHD mabiso_a_Page_093.tif
048cf0fee01e50f3f95620cd5b8a87ea
3b7ff2200ba18e43e51d970580bc112be8050eb4
F20110318_AABYGP mabiso_a_Page_076.tif
e61448ffe0b75d583044a1e6bcdf9718
d78c87c5f5f44de8cef73fe75560a3e2e4da8e4c
5297 F20110318_AABZKG mabiso_a_Page_079thm.jpg
5f292549590f2a8c6c5dd23a94cb06c3
1b390759ba45172b6a44c334d96d0db9489c595c
5923 F20110318_AABZJS mabiso_a_Page_061thm.jpg
842c02c60c069c9ef790374dd46793e5
fe24474ae6bbb0701db65242fb2f954341cbd6f4
F20110318_AABYHE mabiso_a_Page_094.tif
0aadf20dd616a656ac7f9c822c387744
dfc8e9f531d21e1bc76747e22324e67f464d8a25
F20110318_AABYGQ mabiso_a_Page_077.tif
d9265b95a1bd7c6970654766f5ec4cfa
728f65b9de0d8778062743ac713f9b8206d86d81



PAGE 1

ESTIMATING CONSUMERS WILLINGNESS-TO-PAY FOR COUNTRY-OF-ORIGIN LABELS IN FRESH APPLES AND TOMATOES: A DOUBLE-HURDLE PROBIT ANALYSIS OF U.S. DATA USING FACTOR SCORES By ATHUR MABISO A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Athur Mabiso

PAGE 3

To my parents

PAGE 4

iv ACKNOWLEDGMENTS Firstly, I thank Almighty God for the oppor tunity and blessing of education. I consider the successful completion of this work a gift from God and I am truly grateful. To my family and friends who provided unquantifiable support and encouragement I express my gratitude. I sincerely thank the International Agricultu ral Trade and Policy Center, the Florida Agricultural Experiment Station and the Food and Resource Economics Department for providing the necessary funding for this res earch and my masters studies at the University of Florida. This thesis would not have been without their financial support. My utmost thanks to my thesis committee, Drs. James Sterns, Lisa House, Allen Wysocki, and John VanSickle, who on numerous occasions offered their time, solid advice, guidance, and instru ction in the production of this work. Special acknowledgement goes to my committee chai r Dr. James Sterns whose guidance and confidence to let me try different approaches made the process of writing this thesis an adventurous learning experience. Also, special thanks go to Dr. Lisa House for being available to share her knowledge and insights on data analysis. For assistance with certain aspects of statistics and equivalency testing, I thank Mr. Carlos Jauregui. Finally, to my fellow gr aduate students, the FRE-GSO, BGSO, ASU, OCS, MANRRS, Florida Rotaract Club and the rest of the alphabet soup; I am thankful. They made my graduate studies experien ce at Florida forever memorable.

PAGE 5

v TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x ABSTRACT....................................................................................................................... xi CHAPTER 1 INTRODUCTION........................................................................................................1 Background...................................................................................................................1 Problematic Situation....................................................................................................3 Problem Statement........................................................................................................4 Testable Hypotheses.....................................................................................................4 Research Objectives......................................................................................................5 Specific Objectives.......................................................................................................5 Outline of Thesis...........................................................................................................5 2 LITERATURE REVIEW.............................................................................................7 The History of Country-of-Origin Labeling.................................................................7 Previous Studies............................................................................................................8 Food Labeling in Fresh Apples and Tomatoes......................................................8 Country-of-Origin Labeling Studies....................................................................10 Experimental Auctions...............................................................................................13 3 RESEARCH METHODS AND DATA.....................................................................18 Study Design and Data...............................................................................................18 Auction Bid Data Collection......................................................................................23 Candy Bar Auction..............................................................................................23 Apple and Tomato Auctions................................................................................25 Written Questionnaire.................................................................................................26 4 THEORETICAL FRAMEWORK AND ANALYTIC TOOLS.................................28

PAGE 6

vi The Psychology and Theory of WTP Decisions.........................................................28 Theoretical Framework...............................................................................................31 Analytic Tools............................................................................................................33 The Double Hurdle Model...................................................................................33 Factor Analysis (Principle Component Analysis)...............................................35 5 AUCTION BID ANALYSIS AND RESULTS..........................................................40 6 EMPIRICAL SPECIFICA TION AND RESULTS....................................................48 Empirical Specification..............................................................................................48 Description of Variables......................................................................................49 Organization of Model Estimations.....................................................................54 Models without Factor scores.....................................................................................54 Apple Model (Model 1).......................................................................................54 Tomatoes Model (Model 2).................................................................................58 Combined Apples and To matoes Model (Model 3)............................................60 7 FACTOR ANALYSIS RESULTS.............................................................................63 8 ECONOMETRIC MODELING WITH FACTOR SCORES.....................................70 Apples Model (Model 4)............................................................................................70 Tomatoes Model (Model 5)........................................................................................72 Combined Apples and To matoes Model (Model 6)...................................................75 9 CONCLUSIONS AND IMPLICATIONS.................................................................79 Summary.....................................................................................................................79 Implications................................................................................................................81 Areas for Further Research.........................................................................................85 APPENDIX A EXPERIMENTAL AUCTIONS INSTRUCTIONS...................................................88 B WRITTEN QUESTIONNAIRE.................................................................................99 C BID SHEETS............................................................................................................109 D EQUIVALENCY TESTING AND SAS CODE......................................................111 E DESCRIPTION OF VARIABLES...........................................................................114 F FACTOR ANALYSIS RESULTS...........................................................................115 G CHI-SQUARE SPECIFICATION TEST.................................................................121 H PROGRESSION OF MEAN BIDS BY LOCATION..............................................122

PAGE 7

vii LIST OF REFERENCES.................................................................................................124 BIOGRAPHICAL SKETCH...........................................................................................129

PAGE 8

viii LIST OF TABLES Table page 3-1. Data distribution across location..............................................................................19 3-2. Demographic profile of respondents........................................................................21 5-1. Average WTP for apples (n = 136) and tomatoes (n = 175)....................................42 5-2. Average WTP for apples (n = 108) and tomatoes (n = 126): Sampling only those consumers who were WTP more than $0.00............................................................43 5-3. Mean WTP across location......................................................................................44 6-1. Apples probit model without factor scores (Model 1) ..............................................56 6-2. Apples truncated tobit m odel without factor scores (Model 1) ................................57 6-3. Tomatoes probit model without factor scores (Model 2) .........................................58 6-4. Tomatoes truncated tobit model without factor scores (Model 2) ...........................59 6-5. Combined apples and tomatoes probit model without factor scores (Model 3) .......60 6-6. Combined apples and tomatoes truncat ed tobit model without factor scores (Model 3) ............................................................................................................................... 61 7-1. Rotated component matrix(a) for f ood quality proxy vari ables-apple data.............64 7-2. Rotated component matrix(a) for f ood preference proxy variables-apple data........65 7-3. Rotated component matrix(a) for food quality proxy variables-tomato data...........66 7-4. Rotated component matrix(a) for food preference proxy variables-tomato data.....67 7-5. Rotated component matrix(a) for food qua lity proxy variables-combined apple and tomato data set..........................................................................................................68 7-6. Rotated component matrix(a) for food preference proxy variables-tomato data.....69 8-1. Apples probit model with factor scores (Model 4) ...................................................71

PAGE 9

ix 8-2. Apples truncated tobit model with factor scores (Model 4) .....................................71 8-3. Tomatoes probit model with factor scores (Model 5) ..............................................72 8-4. Tomatoes truncated tobit model with factor scores (Model 5) .................................73 8-5. Combined apples and tomatoes probit model with factor scores (Model 6) ............75 8-6. Combined apples and tomatoes trun cated tobit model with factor scores (Model 6) 77 9-1. Comparison of mean bids: U.S.A. Grown versus Other Country labels..................86 D-1. SAS equivalency testing results.............................................................................111 E-1. A detailed description of variab les used in the econometric models.....................114 F-1. Correlation matrix for fo od safety proxy variables................................................115 F-2. KMO and Bartlett's test fo r food safety proxy variables........................................116 F-3. Total variance explained in food safety factor scores............................................116 F-4. Total variance explained in food quality factor scores-apple data.........................116 F-5. Total variance explained in food pr eferences factor scores-apple data..................117 F-6. Total variance explained in food quality factor scores-tomato data......................118 F-7. Total variance explained in food pr eferences factor scores-tomato data...............118 F-8. Total variance explained in food quality factor scores-combined apple and tomato data.........................................................................................................................11 9 F-9. Total variance explained in food prefer ences factor scores-combined apple and tomato data.............................................................................................................120 G-1. Chi-square values used in the specification test of each model.............................121

PAGE 10

x LIST OF FIGURES Figure page 3-1. Sample distribution across locations........................................................................19 3-2. Comparison of gender between surv ey data and U.S. Census data.........................20 5-1. Line graph showing the pr ogression of combined bids in both tomato and apple auctions.....................................................................................................................41 F-1. Food quality factors scree plot-apple data..............................................................116 F-2. Food preferences factors scree plot-apple data......................................................117 F-3. Food quality factors scree plot-tomato data...........................................................117 F-4. Food preferences factors scree plot-tomato data....................................................118 F-5. Food quality factors scree plot-combined apple and tomato data..........................119 F-6. Food preferences factors scree pl ot-combined apple and tomato data...................119 H-1. Progression of mean bids in Gainesville, FL.........................................................122 H-2. Progression of mean bids in Lansing, MI..............................................................122 H-3. Progression of mean bids in Atlanta, GA...............................................................123

PAGE 11

xi Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science ESTIMATING CONSUMERS WILLINGNESS-TO-PAY FOR COUNTRY-OF-ORIGIN LABELS IN FRESH APPLES AND TOMATOES: A DOUBLE-HURDLE PROBIT ANALYSIS OF U.S. DATA USING FACTOR SCORES By Athur Mabiso August 2005 Chair: James. A. Sterns Major Department: Food and Resource Economics In September 2006, Mandatory Country-of-O rigin Labeling (MCOOL) policy is set to take effect in accordance with the Om nibus Appropriations bill passed by Congress in January 2003. U.S. producers and marketers of fresh apples and tomatoes will be subject to this law while contending with increasi ng import competition for the domestic market. This thesis explores U.S. consumers willingness to pay (WTP) for Country-of-Origin Labeling (COOL) in fresh apples and tomatoes, particularly the label U.S.A. Grown, in order to inform policymakers and industry part icipants of some of the implications of COOL in the U.S. market. By analyzing how much U.S. consumers will pay for the label U.S.A. Grown in fresh apples and tomatoes, and establishing the major determinants of the WTP, the thesis provides useful empi rical evidence to give good reason for MCOOL or at the very least voluntary COOL, if costs of labeling ar e less than the price premiums. U.S. consumers are found to be willing to pay approximately $0.49 more for a pound of apples labeled U.S.A. Grown and $0.48 more for a pound of tomatoes labeled U.S.A.

PAGE 12

xii Grown in comparison to unlabeled yet identic al fresh apples and tomatoes, respectively. Of the consumers survey ed, 79% are willing to pa y a premium for fresh apples labeled U.S.A. Grown while 72% will pay a premium for fresh tomatoes labeled U.S.A. Grown. Applying the double hurdle probit model, we developed six variants of the sa me model specification. Results show that consumer food quality perceptions and loca tion are statistically significant explanatory variables for determining the amount consumers are willing to pay for the U.S. labeled produce. Some demographics and psychographics also showed a significant relationship with WTP, though with less magnitude. Data were collected using a written questionnaire a nd a Vickrey auction (fifth-bid sealed-price). The written questionnaire collected data on consumer demographics and perceptions of food safety, food quality, and food preferences among other variables. The Vickrey auction determined the actual amount cons umers are willing to pay to exchange identi cal unlabeled fresh apples and tomatoes for those labeled U.S.A. Grown. The information in this thesis can be used by U.S. producers and market ers of fresh apples and fresh tomatoes to develop formidable marketing strategies in their efforts to boost demand for U.S. produce in the face of rising import competition. It is also informative to policymakers who are in the process of making laws on COOL and tr aceability of agricu ltural food products.

PAGE 13

1 CHAPTER 1 INTRODUCTION Background As consumer demand for agricultural food-products becomes more complex and dynamic, food labeling is taking an increasi ngly important role in the food marketing system (McCluskey and Loureiro, 2003) Consumers are constantly obtaining different kinds of information about food-product attr ibutes via food labels and their purchasing decisions are influenced by these. Theo retically, consumers demand food-product attributes (e.g. food quality or taste) not the food-product per se and the food-product is considered to be merely a bundle of these indivi dual attributes that give rise to utility. Thus purchasing decisions made by cons umers are based on specific food-product attributes embodied in a food-product (Lancaster, 1966) This is important if one is to rec ognize the significance of food labeling. Food labels present information about specific food-product attributes, which potentially can affect consumer willingness to pay (WTP) a nd in turn the effective demand for a foodproduct. Some studies have found consumers to be willing to pay premiums for ecolabels, organic labels, or count ry of origin labels (COOL) (Loureiro et al., 2001; Burton et al., 2001; Umberger et al., 2002 respectively) Alternatively, these can be viewed as premiums for desired food-produc t attributes, which the labels make claim to. Thus, in the case of eco-labels the desired attri bute would be Environmentally-friendly, organically produced in the case of organic labels, non-GM food in non-GM labels,

PAGE 14

2 and U.S.A. Grown, in the case of Country of Origin Labeling (COOL). This alternative view is more consistent with Lancaster. Though the studies have shown consumer s to be willing to pay premiums for labels, which indicate specific food-product attr ibutes, it is not necessarily the case that all labels will command a price premium fr om consumers. Different consumers respond differently to different labels; some willing to pay and others not. Even more so, those who are willing to pay will pay different amounts. Producers and marketers alike are aware of this complex mesh of possibilities, and know that labels making claim to a key product attribute can ultimately determine if the product will be purchased and/or how much is purchased and/or at what price. In the U.S. fresh produce industry there is t opical interest in th ese complexities of food labeling, particularly COOL. This is la rgely due to COOL legislation in the 2002 Farm Security and Rural Investment Act. Sub title D of this composite act specifies that currently market actors can voluntarily label their products with COOL so as to inform the shopper at the final point of purchase. Guidelines for voluntary COOL, which were issued by the USDA in October 2002, apply to fresh meats (beef, pork, lamb and fish) as well as fresh peanuts, fruits and vege tables the so-called covered products (VanSickle, 2003) These products were selected to be covered by the policy because proponents asserted that these are food products most prone to food safety and health problems. Though currently voluntary, the law on COOL is set to change to mandatory country of origin labeling (MCOOL); unless new proposed bills, such as the Meat Promotion Bill of 2005 are passed. The cha nge to MCOOL was initially set to be effective on the 30th of September 2004 but this was postponed by two years with the

PAGE 15

3 passing of the Omnibus Appropriations bill in January 2003. The imminent change to MCOOL brings to the forefront several issues of debate, surrounding the justification of MCOOL policy. One of these major issues co ncerns consumers WTP for COOL. There is little empirical information on how much consumers are willing to pay for COOL in different products. Information on this woul d be important in poi nting policy and/or decision makers in a particular direction in as far as implications of MCOOL are concerned (Menkhaus, 2001) This study seeks to fill part of this in formation gap by analyzing if U.S. primary shoppers are willing to pay a premium for fr esh apples and tomatoes labeled U.S.A. Grown. In addition, the study examines whethe r premiums for U.S.A. Grown labels in fresh apples and tomatoes are product specific or not, then ascertai ns what the prominent factors affecting the WTP for the U.S.A. Grown labels are. Thus, the study gives several insights about COOL in the produce sector. Problematic Situation In 2006, MCOOL in fresh fruits and vegetabl es is expected to take effect, in accordance with the revised Farm Bill of 2002. Predicted consumer response to MCOOL, however, is lacking. U.S. producers of appl es and tomatoes are uncertain if MCOOL policy, which producers and packers expect to be very costly to implement, will have a negative impact on consumer purchases and i ndustry profits. They do not know what to expect in terms of how much consumers will pay for COOL, if indeed they are willing to pay. In addition it is unclear what the key factors affecting the consumers WTP for COOL are. Food safety and food quality concer ns are some of the factors hypothesized to be major drivers of WTP for COOL. However, debate remains on the issue due to lack of empirical information.

PAGE 16

4 Access to such information would be useful as U.S. producers strive to market more competitive produce relative to import substitutes. Several players within the marketing channel are also interested in obtaining information on whether the WTP for U.S. labels is product specific or not. Th is is particularly so, given rising import competition and speculation on the prospects of differentiating domestic produce on the basis of U.S.A. Grown labe ls. Moreover, the potential of segmenting the market on the basis of COOL is a reason w hy market players are interest ed in knowing more about the consumers WTP for COOL. Problem Statement It is not known if U.S. primary s hoppers who purchase and consume fresh tomatoes and apples are willing to pay a premium for produce labeled U.S.A. Grown over identical produce w hose country of origin is not specified. Furthermore, it is uncertain which key factors determine their WTP for fresh apples and tomatoes labeled U.S.A. Grown This leaves agribusinesse s and policymakers to decision-making based on asymmetric information about COOL policy in the fresh apple a nd tomato industries. Testable Hypotheses The primary hypotheses that this rese arch will test are as follows: 1. If fresh apples and tomatoes are labe led U.S.A. Grown, then U.S. primary shoppers will be willing to pay more money for them as compared to what they are willing to pay for identical unlabeled pr oduce whose country of origin is unknown. 2. If consumers are willing to pay a premiu m for fresh apples and fresh tomatoes labeled U.S.A. Grown then the premiu ms will be product specific and unequal. 3. Consumer perceptions about food quality, food preferences and food safety are key factors of WTP for COOL in fresh apples and tomatoes.

PAGE 17

5 Research Objectives The overall objective of this research is to determine if a differential premium exists between U.S. primary shoppers WTP for fresh apples labeled U.S.A. Grown and fresh apples without a label of origin. Th e same is sought for in the case of fresh tomatoes. Underlying determinants of the diffe rential premiums (if they exist) will also be ascertained in both cases. Specific Objectives 1. To determine if U.S. primary shoppers wi ll be willing to pay a premium for fresh apples and fresh tomatoes labeled U.S.A. Grown by calculating the mean premiums (PMTi) recorded from each product sample and testing if they are significantly different from zero 2. To ascertain if WTP premiums for COOL are product specific and unequal. This will be done by performing the z-test for independent samples followed by a premium equivalency hypothesis test for two unrelated samples. 3. To assess if perceptions about food quali ty, food preferences and food safety are key factors of WTP for COOL in fresh apples and tomatoes The last objective will be achieved by performing a double hurdle probit analysis. In order to test the hypothesis that percep tions about food quality and food preferences are key factors of WTP for CO OL in fresh apples and tomatoes, principle component factor analysis will be done and resulting fa ctor scores will be used as explanatory variables in the double hurdle probit analys is. The significance of each factor score variable will be assessed, and so will the respective marginal effects. Outline of Thesis This chapter introduced the thesis, by giving an overview of the problematic situation to be addressed and the objectives central to th e study. In Chapter 2, I begin with a discussion of the hist ory of COOL in the U.S. followed by a review of the literature on food labeling and COOL. Relevant previous studies are referenced and a

PAGE 18

6 section on experimental auctions is also provi ded. In the third chapter, a description of the data set and how it was collected is pres ented. The chapter includes a summary of the auction bid procedures and the survey ques tionnaire used in the study. Chapter 4 opens with a section on the psychology and theory of WTP decisions, before presenting the theoretical framework and the analytic tools used in this thesis. Chapter 5 is a presentation of the aucti on-bids price analysis and results, which mainly entail univariate statistical analyses while Chapter 6 is a presentation of the double hurdle probit specification and estima tions without factor scores. Chapter 7 follows, showing the results from the factor an alysis performed in the process of deriving factor scores to be incorporated in subsequent double hurdle estimations. These estimations are presented in Chapter 8 befo re Chapter 9 concludes the thesis with a summarized discussion of findings and their im plications. Areas for further research are also made reference to in the final chapter of the thesis.

PAGE 19

7 CHAPTER 2 LITERATURE REVIEW The History of Country-of-Origin Labeling The rudiments of COOL of fresh produce in the United States can be traced back to the Agricultural Marketing Act of 1946 (United States Statutes, 1946) In this act, the consumers right to knowledge or inform ation about the agricu ltural produce s/he consumes is fostered. Emphasis is, however, on information that pertains to livestock and in most part, the act deals with various asp ects of food distribution and consumer welfare (e.g. food safety, informational labeling and ge neral packaging standards), not COOL in fresh produce marketing. Nonetheless, these are legal issues that unde rpin COOL of fresh produce and thus unsurprisingly form the f oundations of todays all encompassing COOL policy and the proposed 2006 MCOOL policy. Until 1998, not many amendments pertaining to COOL are seen in the 1946 Act. In July of 1998 amendments to the act were appe nded to the 1999 fiscal year appropriations bill by the Senate, only to be removed by the House-Senate conferees prior to its passage (Schupp and Gillespie, 2001) It is in 2002 that one can ac tually see serious amendments to the act that legislatively introduce COOL at the federal level. This set the stage for a debate on COOL, pitting voluntary COOL vers us MCOOL in the on-going deliberation about the appropriate policy dire ction to take on the matter. At the state level, the hist ory of COOL presents a so mewhat homogeneous profile. In most states no law on COOL existed. It is in the private sectors of the industry that COOL occurred and only on a voluntary basis. However, in Florida a statute was passed

PAGE 20

8 in 1979 making it compulsory for re tailers to label fruit and ve getables country of origin (Florida Statute, 1979) Similarly, in the state of Main e, a COOL law was passed in 1989 making it a requirement for any vegetable or fr uit produced outside the United States to have COOL if the origin is deemed to have lower pesticide residue standards and regulations compared to those of the United States (Maine Statute, 1989) This was later modified to MCOOL of all imported produce, i rrespective of standard s and regulations in 1999 (Maine Statute, 1999). Overall, there is limited pr evalence of legislative acts on COOL at the state level. However, in the two instances that it exists, the main premise for its existence is food safety concerns and minimum acceptable quality standards. This same premise stands today as one of the fundamental arguments for MCOOL. The other is consumer freedom of informed choice. All these come under th e overarching principle of the consumers right to information and al leviation of informational asymmetry in the food markets (Golan et al., 2000) Previous Studies Food Labeling in Fresh Apples and Tomatoes In the last decade several studies have b een carried out on food labeling in apples and very few in the case of tomatoes. Many of these have looked at consumer behavior and consumer perceptions in the contex t of WTP for food labeling. For example, Blend and Van Ravenswaay (1999) studied eco-labeling in apples and how eco-labels influence consumer behavior by affecting their percep tions about food-product attributes. As their study confirmed, changing the labeling on th e apples and ina dvertently the ecoinformation provided to the consumers at the point of purchase, changed consumer buying behavior. Consumers were willing to try eco-labeled apples as well as pay a premium for them if made av ailable in the marketplace. Sp ecifically, the study showed

PAGE 21

9 that at the time, more than a third of su rveyed consumers were willing to pay a $0.40 premium for the eco-labeled apples. In a more recent paper, Loureiro et al. (2001) studied WTP for organic labels and eco-labels in apples. Using contingent va luation methods (CVM), they estimated a maximum-likelihood multinomial logit model which included food safety, perceived produce quality and environmental concerns together with demographic variables as independent variables. Results are that f ood safety and produce quality concerns were significant determinants of WTP for the labels. The same was ascertained for environmental concerns and the presence of children under the age of 18 in the household, which was found to increase the probab ility of an U.S. consumer to purchase organic labeled apples. Regard ing regular unlabeled apples, they established that all significant independent variab les (i.e. food safety concerns, quality perceptions, environmental preferences, pr esence of children in the hous ehold, income level) had a negative impact on the likelihood to purchas e with the exception of household size, which had a positive effect on the WTP. Generally, U.S. consumers were found to be willing to pay for organic labeled and eco-label ed apples, with those that are eco-labeled ranking as a second choice while regu lar unlabeled apples ranked last. In the case of tomatoes, not many studies dealing with WTP for labeling have been conducted. However, Akgungor et al. (1999) analyzed WTP for reduced pesticide use labeling in tomatoes, though in the context of metropolitan areas in Turkey. Using CVM, the study found that consumers were willing to pay a 2 percent premium for the labels. In addition, demographics such as income, ge nder and household size were significant determinants positively increasing the proba bility of consumers WTP for labels.

PAGE 22

10 However, age was not significant. Though this study clearly suggested a causal relationship between WTP for labels in tomato es and demographics, this was specific to Turkey. The scenario in the United States may be different. In the United States, a relatively older study by Eastwood et al. (1987) explored some of these aspects in the produce industr y. By surveying households in Knox County, TN, they found that just over one half of the consumers fr om Knoxville, TN were likely to be concerned about the origin of th eir tomatoes. In addition, housewives and professionally occupied responde nts were most likely to be willing to pay a premium for locally produced tomatoes over out-of-state tomatoes. A probit estimation was used to ascertain these results. In the case of appl es, broccoli and cabbage, it did not seem to matter where the produce came from. This sugge sts that different pr oduce yield different consumer responses when it comes to WTP for origin labels. Thompson (2003) looked at retail demand for di fferentiated fresh tomatoes and particularly greenhouse tomatoes, which consumers identify th rough labels. Using descriptive statistics he esta blished that the average market price of greenhouse labeled tomatoes was higher than that for field-gr own tomatoes in the cities of Albany, NY, Chicago, IL, Dallas, TX, Los Angeles, CA, a nd Miami, FL. Only in Atlanta, GA was the average market price for greenhouse toma toes lower ($1.56 compared to $1.99). In addition, he presented empirical evidence showing that greenhouse and on-the-vine tomatoes experienced market-share growth at the expense of field-grown tomatoes, despite a reduction in market prices of field-grown tomatoes. Country-of-Origin Labeling Studies Schupp and Gillespie (2001) are part of the first group of researchers to have turned the focus onto COOL in ag ricultural products and specifically beef. By sampling

PAGE 23

11 households in Louisiana in order to establ ish if consumers support MCOOL of beef in grocery stores and restaurant s, they found that on averag e 90.3 percent of surveyed consumers were willing to support MCOOL. Als o, they looked at factors influencing the support for MCOOL by estimating a probit m odel and found the food safety concern variable to be significant and positively aff ecting the probability of a consumer to support MCOOL. Consumer preference for locally pr oduced beef also turned out to be a statistically significant independent variable positively affecting the likelihood to support MCOOL. In another paper, Loureiro and Umberger (2002) researched COOL when they explored consumer WTP for a MCOOL program in general as well as WTP for steak and beef hamburgers labeled U.S. Certified beef. By survey-sampling consumers in different grocery store locations in Colora do, they established that consumers were willing to pay for a MCOOL program. They then found that consumers who had completed a high level of education and had a hi gh income level were less likely to pay a premium for COOL or U.S. Certified labe ls in beef. This disproved their initial hypothesis that a more educated and wealthie r consumer would pay attention to COOL and be more likely to pay a premium for it. Moreover they showed that female consumers are most likely to pay a premium for C OOL and are more s upportive of MCOOL. All these results pertained to the U.S. certified beef hamburger. In the case of U.S. certified steak, slightly different results we re obtained with the presence of children in the household being the only significant de mographic factor of the likelihood of WTP. Overa ll, the findings suggested that WTP is not only based on variables such as food safety concern, food quality and demographics but also on the

PAGE 24

12 product itself (i.e. different products would have different factors influencing the consumers WTP under consideration.) Loureiro and Umberger (2004) used experimental aucti ons to solicit information on U.S. consumers WTP. Again they looked at beef and the reason being that most of the debate pertaining to MCOOL has centered on be ef owing to its link to BSE food safety concerns which have, in the re cent past, plagued the industry in isolated circumstances. According to these authors, COOL in beef is a less important factor of consumers WTP as compared to food safety inspection labels product quality labels (tenderness) or traceability of the beef. This presents mixe d results on COOL and suggests a weaker link between COOL and food safety concer ns than previously thought to be. Similar studies in Europe, though using stat ed choice data, have revealed similar responses. In Belgium Verbeke and Ward (2003) conducted a survey seeking to determine those informational labeling cues on be ef that really attract consumers interest (attention and importance ratings) as part of consumer risk aversion associated with food safety. By analyzing a data set of 278 obser vations and using ordered probit models the authors found that COOL is of moderate interest to consum ers, while traceability is of low interest. Instead, food quality concerns rank very high and advertising/publicity of labeling policy positively affects consumers in terest in labels pert aining to quality and origin. Roosen et al. (2003) however, found that European consumers are more concerned about COOL. English, French and German co nsumers were surveyed and analyses found that French and German consumers are more concerned about the origins of their beef than about product attributes (e.g. quality). Br itish consumers were shown to have a high

PAGE 25

13 regard for COOL but not to the same exte nt. Interestingly en ough, high WTP premiums were recorded for hormone-free labels on beef, this being the underlying unobservable variable postulated to be the cause of the high regard for COOL in Europe. While all these studies sugge st a relationship between consumers WTP for COOL and food safety concerns, food quality and demogr aphics, it is evident that other variables may have a bearing on the nature of the re lationship, e.g., the types of products under consideration or the consumers geographi c location. With respect to fruits and vegetables the situation is unclear, since the majority of previous research has focused on the beef sector. Though some work has been done with apples, none is specific to the question on COOL and little has to do with tomatoes as the product of interest. This thesis delves into these specific issues to address COOL in fresh apples and tomatoes. Experimental Auctions Experimental techniques are not widely us ed in economic research because it is difficult to incorporate experime ntal controls when dealing w ith people as the subject of study. Constraining the human subject to a cont rol is exacting if not impossible, without breaching the necessary unbiasedness and validity in the research. Nevertheless, of late the use of experimental t echniques to research consum er WTP and product value has gained much popularity and has ha d satisfactory degrees of success (Maynard et al., 2003, Alfnes and Rikertsen 2003, Shogren et al., 2002, Rousu et al., 2002, Fox et al., 1996, and Hayes et al., 1995) In the case of WTP for COOL studies, a few researchers have used experimental auctions to collect data. Others have conti nued to base their findings solely on Stated Choice (SC) surveys and CVM, and yet questions have been raised about their validity. List and Gallet (2001) List (2001) and List and Shogren (1998) for example, point out

PAGE 26

14 that respondents of SC and CVM surv eys tend to exaggerate their WTP. Harrison and Rutstrom (2005) further augment this view when they summarize 39 CVM based studies and find that hypothetical bias is sign ificant in 31 of the 39 studies. This has led some researchers to move away from CVM and towards a variant of CVM, where calibration is performed (e.g. by us ing statistical bias functions, uncertainty scale adjustment or cheap talk). Othe rs have tended towards combining these calibration methods with the use of experi mental auctions instead of CVM. These methods are believed to do away with most hypothetical bias associated with SC and CVM surveys. Based on this, this study used a combination of experimental auctions, cheap talk and a written questi onnaire in an effort to minimize hypothetical bias in the data. However, neither calibration of CVM nor experimental auctions are without problems. For instance, research has found that sometimes hypothetical bias remains even after calibration of CVM. For example, Aadland and Caplan (2003a) and Samnaliev et al. (2003) show that using cheap talk does not always eliminate hypot hetical bias and neither does using an uncerta inty scale or a not sure/dont know option in the questionnaire. With respect to experimental auctions the sealed-bid, second-price (Vickrey) auction has received wide recognition in term s of dealing with hypothetical bias. Named after William Vickrey who pioneered auction re search in 1961, the auction is desirable for its truth revealing traits in the consume rs bid. Since the auction is set up in such a way that winning bidders pay the second highest bid price for the product being auctioned, the bidder is encouraged to bid trut hfully as his/her weakly dominant strategy.

PAGE 27

15 According to Lusk et al., (2004, p.1) this is the incentive co mpatibility of the Vickrey second bid auction, meaning that it inherently provides individuals with incentives for truthful bidding while simultaneously puni shing untruthful bidding. If a consumer underbids, s/he runs the risk of losing th e auction, whereby s/he could have won at his/her true WTP value. As for the consumer that overbids, s/he runs the risk of winning but at a much higher price than his/her true bid. Variants of the second bid auction have also been popular because they have simila r incentive compatibility features (e.g. the Vickrey fifthpriced sealed bid variant used in this study). Comparing the Vickrey auction to anothe r incentive compatible method, the Becker DeGroot Marschak (BDM) mechanism, Lusk et al. (2004) show that the Vickrey auction is more punitive to untruthful bidding in the case of high value bidders. However, when low value bidders are involved, the BDM is fou nd to be more punitive. Lusk et al. thus recommend the application of the Vickrey auctio n when research interest is in high value individuals. Besides incentive compatibility other a dvantages of using the Vickrey auction second bid price or variants thereof are noted basically that respondents actually engage in a transaction and have the opportunity to s ee the product of interest. This adds to the validity of research results as it closely mirrors the s hopping experience within which purchasing decisions are made in the real world. However, there are a few challenges associ ated with the Vickrey auction and other experimental auction methods. Menkhaus lists these as disadvantages in his commentary of 2001. First he mentions that the bidding process in auctions do not naturally mimic how consumers reveal prefer ences in grocery stores (Menkhaus, 2001, p. 2) Instead of

PAGE 28

16 contending with administered pricing, which is the case in grocery stores, consumers have to bid their own prices. In addition, there are no competing substitutes in the experiments, which is the case in grocery st ores. Thus the WTP decision-making process is arguably different. Despite these assertions, the Vickrey auction has proved to be formidable especially when cheap talk is incorporated in order to frame a grocery store mindset for the consumers participating in the auction. In such instan ces, it has shown effectiveness in revealing WTP. For example, in a Vickrey auction-based study, Hurley et al. (2004) apply this method when they study 329 cons umers in Ames, IA; Manhattan, KS; Raleigh, NC; Burlington, VT; Iowa Falls, IA and Corval lis, OR and effectively establish WTP in pork. Researching the WTP for environmental a ttributes embedded in ten different pork loin chop packages, they inco rporate cheap talk as well as provide some form of substitutes. They use several related environm ental attributes in their experiment, thus providing consumers with varyi ng levels of environmental fr iendliness to contend with. Admittedly there is essentially one product (por k chops) in the experiment and this does not cover a wide enough spectrum of substitutes as would be the case in grocery stores. Nevertheless, statistically valid findings show that 62 percent of the consumers are WTP a premium for the most environmentally friendl y pork in the experiment with more than 40 percent WTP a dollar or more. These findings are evaluated and found to be hypothetical bias-free. In another Vickrey auction based study, Roosen et al. (1998) measured the WTP for the removal of insecticide use in appl e production. They analyzed a sample of 54 primary shoppers in a Midwestern univers ity town, using nonparametric statistical

PAGE 29

17 analyses and the double hurdle model. Findings revealed that the consumers are willing to pay $0.34 for apples not treated with Azinphos-methy (APM) pesticide and $0.43 for a similar bag of apples not treated with any kind of neuroactive insecticides (NAI). The double hurdle analysis reveals that concern about pesticid e use in food production is a variable with a positive effect on the probabi lity to bid greater than $0.00 for the apples with reduced pesticide use. The same is found for household income. Concern about food prices is found to be insignificant in influe ncing the probability to purchase the reduced pesticide use apples. Roosen et al. also establish that the presence of children below the age of five in a household has mixed effects. If a household has a child below the age of five, then the probability of being willing to pay more than $0.00 is reduced. However, for those households that are willing to pay, they will pay a higher premium if they have children below the age of five. Overall the findings add to the growing list of studies that use the Vickery auction (second bid price or variants thereof) to su ccessfully elicit consum er WTP for different product attributes without hypothetical bias. This study follows suit and uses the fifthprice sealed bid variant given that there was a large number of participants per group that was surveyed in each auction site. This ensures incentive compatibility even with the relatively large groups (as will be explained in Chapter 3).

PAGE 30

18 CHAPTER 3 RESEARCH METHODS AND DATA This chapter documents how the data set was collected and gives a description of the data. Study Design and Data To estimate WTP for COOL in apples and tomatoes a sample of U.S. consumers was recruited for each product. Data were collected in December 2003 and January 2004 using a Vickrey experimental auction (fifth-priced seal ed-bid) followed by a written questionnaire. The Vickrey auction solicited data on WTP premiums then the written questionnaire solicited data on numerous va riables including demographics, food safety concerns, food quality concerns and food preferences. Data were collected in Gainesville, FL, Atlanta, GA and Lansing, MI. Respondent s were randomly recruited through local civic organizations, ranging from faith ba sed organizations to Parents Teachers Associations (PTA) at schools. The su rvey was conducted in each respective organizations facilities and compensation for use of these facilities was made. In total 335 primary shoppers were sampled and the chart below shows the breakdown of these across locations.

PAGE 31

19 72 127 136 0 20 40 60 80 100 120 140number of respondents Lansing, MIGainesville, FLAtlanta, GASample Distribution Across Locations Figure 3-1. Sample distri bution across locations Of the 335 observations sampled, 311 were useable for analysis; 175 from the tomato auctions and 136 from the apple au ctions. The 24 observations deleted were unusable due to missing data. The table belo w shows how the data distributed between the two products vis--vis location. Table 3-1. Data distribution across location Product Location Observations Deleted Observations Retained Percent Cumulative Percent GNV, FL 3 67 38.3 38.3 LAN, MI 2 49 28.0 66.3 ATL, GA 9 59 33.7 100.0 TOMATOES Total (n) 14 175 100.0 GNV, FL 1 56 41.2 41.2 LAN, MI 4 17 12.5 53.7 ATL, GA 5 63 46.3 100 APPLES Total (n) 10 136 100.0

PAGE 32

20 In general, fewer respondents were r ecruited in Lansing, MI, making the data unevenly distributed across loca tions, particularly in the ca se of the apple data. The useable data were such that the responde nts ages ranged from 25 to 65 years and included primary shoppers only. Here, a primar y shopper was defined as an individual responsible for at least 50 percent of food purchases in the household. Comparing the sample data to the U. S. population census revealed several disparities. For instance, a larger proporti on of the sample was female, as shown in Figure 3-2 below. This was nonetheless expect ed considering that most primary shoppers are female and that this was the target population. Research prot ocol had specifically asked for primary shoppers only. 9.6% 90.4% 13.1% 86.9% 49.1% 50.9% 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% Apples dataTomatoes dataU.S. CensusComparison of Gender Between Survey data and U.S. Census Data Male Female Figure 3-2. Comparison of gender between survey data and U.S. Census data For the apple data, up to 90 percent of the responde nts were female and 86 percent in the case of tomatoes data. This is notably different from 50.9 percent females in the U.S. census data. In addition, a few ot her demographics did not mirror those of the

PAGE 33

21 national census data. These observations call fo r caution in the interpretation of results from this study, as extrapol ative generalizations based on the study may be erroneous. The table below shows more details on the co mparability of the survey data to the national census data. Table 3-2. Demographic profile of respondents Sample Average (%) Category U.S. Census Average (%) Apples Tomatoes Age 25-34 27 11.0 9.1 35-44 31 39.0 48.9 45-54 26 36.8 34.6 55-65 16 13.2 7.4 Race White 75 84.6 90.9 Black or African U.S. 12 7.4 4.6 Asian 4 2.2 1.7 Other 9 5.8 2.8 Ethnicity Hispanic 12 6.6 2.3 Income <$15,000 15.2 3.0 1.7 $15,000 to $24,999 13.2 5.1 4.5 $25,000 to $34,999 12.3 7.4 8.0 $35,000 to $49,999 15.1 11.0 13.1 $50,000 to $74,999 18.3 19.1 27.8

PAGE 34

22 Table 3-2. Continued. Sample Average (%) Category U.S. Census Average (%) Apples Tomatoes $75,000 to $99,000 11.0 22.0 13.6 $100,000 or above 14.1 32.4 31.3 Education Bachelors Degree or Higher 24 64.7 67.0 Some College 27 28.6 23.9 High School Diploma (or equivalent) 29 5.9 8.5 Less than High School 20 0.8 0.6 Major disparities are found in the ethnicity, the highest level of education attained and the pretax household income variables. In proportional terms, fewer Hispanics participated in the survey (6.6 percent and 2.3 percent in the apples and tomatoes data respectively, as compared to 12 percent in the 2000 census). The greatest disparity however, showed up in the highest level of e ducation attained where the majority of the sample had attained a much higher level of education than the general U.S. population. More than 60 percent of the respondents in both the apples and tomatoes data had attained at least a Bachelors degree. More over, the survey captured household primary shoppers that were more affluent relative to the census data. Nearly a third of the respondents had a pretax household annual in come greater than $100,000 as compared to only 14 percent for the census data (U.S. Census Bureau, 2000) Negligible disparities are also noted in terms of age and race with more respondents falling in the 35 to 54 age ranges and more Caucasians being sampled. Ev en though the data used in the study were

PAGE 35

23 somewhat deviant from the U.S. population th ey were useable because the deviation was initially expected, since the target population of the study was U.S. primary shoppers and not the general U.S. population. The demographi c profile of U.S. primary shoppers does differ slightly from the census data profile. Al so, some similarities between the data and the U.S. census data do exist. Auction Bid Data Collection The following section describes the manner in which the auction bid data were collected. Respondents participated in two auctions which were fashioned as random fifth-price sealed-bid (Vickrey ) auctions. The first auction was a preliminary run used merely to familiarize the respondents to the auction bidding process. Here the respondents bid for a large Snickers candy bar. Candy Bar Auction Initially, respondents were assigned to a seat with a corresponding I.D. numbered envelope containing unfilled bid sheets, ques tionnaire, consent form and instructions. Each respondent was asked not to communicate with others participating in the survey, thus all of the consumers bids and questi onnaire responses were made independently. The respondents were then given $10 in cas h for their voluntary pa rticipation in the survey. This became entirely theirs and available for use if they so chose, in the event that they won an auction. They were also endowed with a small Snickers candy bar and asked not to consume it. Each respondent voluntarily signed the consent form, which was then collected by one of the survey monitors. Follo wing this, instructions on the whole auction procedure were read aloud to the respondents by the lead monitor, including a cheap talk script before the bidding process began (see Appendix A for the detailed instructions which include th e cheap talk script). The purpose of reading aloud the

PAGE 36

24 cheap talk script to the respondents was to further defray any potential of hypothetical bias. According to List (2001) incorporating a cheap talk script in an experimental auction can effectively improve the quality of data in terms of the accuracy of WTP measures elicited. Respondents were asked to raise their hand if th ey had any questions after instructions were read out to them. Fu rthermore, they were told they could raise their hand to ask questions for clar ification at any point in time. In the first auction, respondents bid the am ount they were willing to exchange their small Snickers bar for a larger Snickers bar. All auction bids in the survey elicited WTP premiums for exchange. A total of five bi dding rounds were conducted for the Snickers candy bar auction, giving the respondents ampl e bidding practice and familiarity with the process. At the end of the auction a binding round was randomly selected and the winners of that round (i.e. the four hi ghest bidders) had to pay the 5th highest price of the randomly selected round in order to exchange their small Snickers candy bar for a larger Snickers candy bar. Since the size of groups participating in the auctions was large (approximately 25 consumers in each group) the 5th price auction was appropriately chosen instead of the commonly used 2nd price. This better e ngaged low-end bidders who could have been disengaged by a 2nd price, if the high-end bi dders bid too high for the low-end bidders. Disengaging low end bi dders would have resulted in a less representative estimate of the mean WTP. Using the 5th price ensured incentive compatibility, which encouraged truthful bidding, and in turn better revealed the consumers demand. In addition, more than one top-bidder could be a winner (i.e. the top four bidders could each win). This too, ensu red incentive compatibility and maintained the truth-revealing trait of the Vickrey auction by eliminating the chance of artificial

PAGE 37

25 scarcity, which usually arises if only one bi dder can win. Similarly, the random selection of the winning bid and winning round ensured truthful bidding by eliminating wealth effects that could arise if a bidder wins an earlier round. After the winning bids were established and exchanges done with the winne rs, the respondents were told they were free to consume their candy. Apple and Tomato Auctions Since there were two products of main inte rest (i.e. fresh apples and tomatoes), respondents were either exposed to the a pple-auction procedures or tomato-auction procedures after completing the candy bar auction. Exactly which one the respondents were exposed to depended on the survey site. For each location (e.g. Atlanta), there had been several survey sites (i.e. civic organi zations) where the auctions were conducted. At each site random selection was done to estab lish whether the apple auction or the tomato auction procedures would be administered. On ly in a couple of instances where the sites surveyed had a large number of voluntary respondents was there need to randomly divided respondents into two separate groups. In such case each randomly created group was ushered into a different r oom right at the onset and thus later exposed to either the apple or the tomato auction procedures afte r their candy auction was over. The different auction procedures for the second auction had been so designed in order to create two samples (one for each product; see Appendix A for details). The manner in which the apple (tomato) auction was conducted was similar to the candy bar auction. Respondents were first endowed with a pound of unlabeled gala apples (off-the-vine tomatoes). Again, as in the candy bar auction, th ey were not allowed to immediately consume their apples (tomatoe s) as they would have the opportunity to exchange them for identical apples (tomatoes) with the label U.S.A. Grown (i.e. if their

PAGE 38

26 bid won the auction). Considerable efforts were made to ensure that the unlabeled apples (tomatoes) bestowed to the respondents had iden tical visible attributes with those labeled U.S.A. Grown. Visible attributes here refers to the size, color, blemishes and variety of the product. Thus the only visible difference was labeling. Prior to bidding, respondents were given the opportunity to visually insp ect the apples (tomat oes) labeled U.S.A. Grown allowing them to compare them with their own unlabeled ones. Thereafter, they were asked to make their private sealed bid. In contrast to five rounds of bidding in the candy bar auction, four rounds of bidding we re conducted in the apples (tomatoes) auction. After each round, survey monitors would colle ct the sealed bids, sort and rank them then post the top five bids in rank order on a board in the front of the room, thus making the bids visible to all partic ipants. This was done to give bid price information feedback to the bidders. At the end of all four rounds, one round was randomly selected as binding, thus establishing the price of exch ange and the winning bidders. Once the whole auction was over and exchanges with winners were done, respondents were told they were free to consume their apples (tomatoes). Written Questionnaire After participating in either the apple or tomato auction, the respondents were asked to complete a questionnaire which solicited information about their buying behavior and stated preferences for fresh produce and labeling. The questionnaire took an average of fifteen minutes to complete, with most questions being based on a nine point Likert-scale. These questions were designed to elicit the levels of attribute-related preferences for each respondent. The de tailed questionnaire is shown in Appendix B Not all questions in the questionnaire were used in this study. Only those pertaining to the

PAGE 39

27 variables that were included in the double hurdle model (later elaborated in the analysis section of this thesis) were used. A 100 pe rcent response rate fo r the questionnaire was achieved since all volunteer participants gave responses a nd none declined to participate after coming to the survey site. However, not all of the respondents answered every question thus giving a completi on rate of just over 80 percen t. Most of the respondents that did not complete the questionnaire did not respond to the question on their pre-tax household income level. Generally consumers may be reluctant to divulge their incomes as this is a somewhat personal question. Neve rtheless, most questions were answered and the auctions data collection went smoothly in all survey sites to provide the necessary data analyzed in this study.

PAGE 40

28 CHAPTER 4 THEORETICAL FRAMEWORK AND ANALYTIC TOOLS The Psychology and Theory of WTP Decisions Consumer cognitive theory in the context of neoclassical consumer theory forms the theoretical framework for this study. Consumer cognitive theory proposes that consumer behavior in food markets is invariab ly affected by a multitude of factors, which can be viewed as either internal or external to the consumer (Hansen, 1972) Those factors considered to be exte rnal are stimuli derived from the product itself, as well as information gained by the consumer about the product and various other environmental sources that may influence the decision to consume. Theoretically, these are all measurable by the researcher and are exogenous to the WTP modeling framework. Practically, not all fact ors are measurable, at least not di rectly and some factors may be endogenous. Similarly, not all internal factors are di rectly measurable. The internal factors include demographics, tastes and preference s (both congenital and learned), perceptions about the product and intrinsic randomne ss owing to asymmetric processing of information by the consumer. Several authors (e.g. Kreps, 1988; Solomon, 2004) consider the randomness factor to be non-ex istent as per the random utility model. According to this school of thought, the indivi dual consumer is assumed to be rational and capable of perfect discrimination. Thus the consumers behavior is inherently deterministic and any randomness factor in the WTP modeling framework can only result from the researchers failure to accurately capture or measure the explanatory variables

PAGE 41

29 influencing the consumers decision making phenomenon. This view has wide acceptance, especially with econometri cians and is adopted in this study. Also of theoretical interest is the dist inction between the decision to purchase and the decision on how much to consume (comm only referred to as th e participation and consumption decisions, respectively). Cragg (1971) makes the case that these are two separate decisions made by the consumer when buying a product. The first is whether to buy or not to buy the product and the second is how much to buy if indeed the consumer chooses to buy. Both decisions may or may not be determined by the same explanatory variables. This forms the basis for using double-hurdle econometric models to explain consumers WTP and consumer demand (e.g. Maynard et al., 2003; Dong and Gould, 1999; Gao et al., 1995 ). Another important aspect of theory, which lends itself to the work of Lancaster (1966) is that pertaining to the trichotomy of product at tributes, namely search, experience and credence attributes. As Lancaster articulated, consumers demand attributes that provide utili ty to them subject to economic constraints. They do not demand the products in and of themselves. Following from this theoretic assertion, Nelson (1970) and Darby and Karni (1973) developed the trichotomy as a means of classifying product attributes, which are considered to be the focal points of demand. Criterion used to classify attributes was the pr e-cost and post-cost associated with quality detection for each attribute (i.e. the economic cost incurred by consumers when identifying the level of quality of the produc t attribute and the economic cost incurred by consumers after dete cting the quality.)

PAGE 42

30 Search attributes were defi ned as those with a low precost of detection and zero post-cost such that consumers would be willing to shop ar ound (hence the name search) in order to find th e best quality. A key distinction of the search attribute was that it was observable and measurable by th e consumer prior to purchase, e.g. product color. Thus, no economic cost can be incurred after detection. Andersen and Philipsen (1998) elaborate on this when they note that the trichotomy is dependent on the level of uncertainty that consumers will face when pr ovided information about the attribute. In the case of search attributes, there is zero unc ertainty as the attribute is observable and measurable by the consumer. In contrast, experience attr ibutes cannot be deciphered by the consumer prior to purchase and consumption (e.g. taste). For this reason, there is a high level of uncertainty (and/or pre-cost); however, once the particip ation and consumption decisions are made, the level of quality of the attribute is dete cted. The consumer can then choose if they want to engage in repeated buys or not. If the quality was low, it is likely that repeated buys will not occur and the high economic co st of continued consumption of a low quality product is averted. Nevert heless, a post-cost and uncertainty still exists as there is a possibility that additional units of th e same product may have higher quality which could yield greater marginal utility. In the case of credence, both the pre-cost and post-cost are high because at any stage, the consumer is unable to measure or observe the level of quality. An example of a credence attribute is country of origin. As Andersen and Philipsen assert, the consumer must rely on a third party for detection purposes, e.g. a product certifying body. Credence is closely linked to beliefs and trust. For instance, the consum er can only believe and trust

PAGE 43

31 that a tomato s/he bought was produc ed in the U.S. There is no way that s/he can test this and in most cases a third party such as the US DA can act as a detection arbiter. In such a case, the consumers level of trust in the US DA then has an effect on the participation and consumption decisions. Understanding the trichotomy described a bove is important to this study because the focus of this study is on a credence at tribute (COOL). Model specification draws from this understanding as f actors that affect consumer WTP for these attributes will differ due to the different nature of each type of attribute. Theoretical Framework Drawing from the aforementioned theories this study assumes that the individual consumer can attain utility from a specific produc t attribute, in this case COOL in apples or tomatoes (U.S.A. Grown). This utility is a function of: (i) consumer characteristics that influence consumer choice and (ii) the co st that the consumer will pay in order to obtain the attribute. Thus, 0 i ic U U (4-1) where i is a combination of consumer characteristics and ic the cost that the consumer will pay to obtain the attribute. Utility gained from the attribute is zero when the consumer is not willing to pay anything to obtain the attribute otherwise it is greater than ze ro. The case of disutility is disregarded (i.e.0 U ) because a rational consumer who does not make mistakes is assumed, whereby the buying decision ultimatel y must yield positive utility. In addition, a consistent consumer is assumed, such that it is not possible for him/her to be willing and unwilling to pay for the at tribute at the same time (i.e the consumer is a perfect

PAGE 44

32 judge of his/her utility func tion and remains in his/her chosen state of WTP.) This assumption allows for mathematical cons istency, without which the theoretical framework cannot hold. The utility function is unobservable and cannot be measured by the researcher; however, a proxy measure of u tility can be estimated by th e WTP. Similarly, it is assumed that not all consumer characteristics are directly observa ble and quantifiable, e.g. consumers perceptions about food quali ty or consumers feelings about food preferences. These are, instead, latent c onstructs whose phenomena are observed via other directly quantifiable proxy variables. T hus, the utility function is deconstructed in similar fashion to Adamowicz et al (1998) with the only difference being that this study proposes a directly observable deterministic parti an indirectly observable deterministic parti and a stochastic error termi The error term is assumed to be independent and identically distributed with a mean of zero and a constant variance. i i i i i ic V c U (4-2) It is postulated that the vari ance of the indirectly observable i can be better estimated by way of a factor analysis of the directly observable and quantifiable proxiesi rather than by using an individual obs ervable proxy variable Mathematically, i i (4-3) where is a vector of factor loadings Thus the WTP decision can then be framed in likelihood terms as i i i iV PMT Pr 0 Pr (4-4)

PAGE 45

33 even though i is unobservable directly. The a bove forms the basic theoretical framework for the double hurdle model estimati on with factor scores, which is proposed in this study. Analytic Tools In analyzing the consumer WTP for U.S .A. Grown labeling, various analytic tools are used in this study. In order to as certain the mean WTP, summary statistics are computed and parametric hypothesis testi ng is performed to assess significance. Equivalency testing is also performed to a ssess price premium equivalency between the mean price premium consumers are willing to pay for U.S.A. Grown labeling in fresh apples and that in fresh toma toes. As earlier alluded to, th e double hurdle probit model is applied to estimate the effects of various factors on consumer WTP. In the process, factor analysis is performed on a set of variable s that, in general, co ncern food safety, food quality and food preference subjec t areas, to derive factor scor es that are incorporated as independent variables in the model estima tions. The following sections present these analytic tools. The Double Hurdle Model The double hurdle model is a product of work by Cragg (1971) after the realization that the Tobit model originally developed by Tobin (1958) was inadequate for a complete analysis of censored or truncated data. Cragg noted that the Tobit model was restrictive as it used the same explanatory variables to estimate the dichotomous choice variable and the quantitative extent of choice variable (i.e the participation and consumption variables respectively). In contrast the double hurdle m odel that he proposed segregated the two, making it possible for each one to have differe nt explanatory variab les and error terms.

PAGE 46

34 By assuming a random utility function that explains the latent dependent utility variable expressed firstly as a likelihood func tion of the WTP and then as the actual WTP amount, this double hurdle model is adap ted here to analyze WTP for COOL. Thus, i i iu Z WTP (4-5) for the participation equation denoting the di chotomous willing to pay or not willing to pay part of the framework. Then, i i iX PMT (4-6) for the quantitative consumption part of the framework. In (4-5) the variable iWTP is the consumer willingness to pay assuming 0 if not and 1 if willing to pay. This dependent variable represents the underlying utility associated with the participation decision; essentially whether or not the consumer derives utility from the attribute. In (4-6)* iPMTis the actual premium that consumers are willing to pay for the apples or tomatoes with COOL, if in (4-5) iWTPwas equal to 1. This represents the magnitude of the latent utility associated with the COOL attribute. iZ and iX are predictor vectors while and are parameter vectors to be estimated for the respective predictor vectors. iZ and iX can potentially be identical and include reduced variables in th e form of factor scores deri ved from factor analysis. If iZ and iXare equal and and are also equal then the tobit model results instead of a

PAGE 47

35 truncated tobit. iu and i are random error terms, normally independently distributed. ) 1 0 ( ~NID uiand ). 0 ( ~2 NIDi Theoretically the underlying utility which is non-measurable can also be expressed as ) (* * i i iZ X f U where iU is the individual consum ers utility. Equations (4-5) and (46) are estimated separately with the (4-5) being estimated first because its results are used in the estimation of the second (i.e. in estim ation of the censoring rule). A probit model can be estimated for the first equation using the maximum likelihood function: ) ( ) / ( ) / ( ) | 0 Pr(* * * i i i i i iZ X X X Z PMT (4-7) Then the second equation can be estimated using, 2 2 2 * *2 ) ( } 2 / ) ( exp{ ) 0 , | ( i i i i i i iZ X PMT PMT X Z PMT f (4-8) Where signifies the standard normal cumulative density function. Factor Analysis (Principle Component Analysis) Factor analysis is a data reduction tool whose main objective is to define an underlying/latent structur e of data and reveal interrelati ons between correlated variables (Hair et al., 2003) It can be used as an intermediate analytic tool to generate a reduced form of data for further analysis. Essentially there are two ways that results of a factor analysis can be used in subsequent data an alyses. The first is by analyzing the factor matrix to select a surrogate variable, whereby that variable with the highest factor loading is chosen to represent the w hole set of variables in the an alysis. This method is less commonly used as it involves omission of t hose variables found to have lower loadings but potentially pertinent to the subsequent analyses; hence it is not used in this study.

PAGE 48

36 The other scenario involves generation of an entirely new yet parsimonious set of variables created from summated scales or f actor scores. In the case of summated scales, a single composite measure is created when se veral original variable s with high loadings are combined to come up with either th e total or (more commonly used) the average score of the variables for use as a replacement variable. In contrast, when using factor scores a composite measure incorporating the va riance of all variable s under analysis is computed for every case/respondent. This co mposite measure portrays the extent to which each cases responses are correlated to the factors derived in the factor analysis. For each factor generated, a factor score will be generated. Basically, factor scores differ from summated scales in that they are genera ted for all variables not just those with high factor loadings. Furthermore, factor scores can be or thogonally generated thus doing away with multicollinearity pr oblems that may arise in a subsequent regression analysis or variants of a regression analysis. There are two classes of factor analysis namely exploratory factor analysis and confirmatory factor analysis (Thompson, 2004). In the case of explor atory factor analysis, the researcher simply performs the analysis to unveil stru cture among variables and take the results as generated. The researcher doe s not start off with any preconceptions about the number of underlying factors or their natu re. In addition the researcher can use the results in subsequent analyses Conversely, confirmatory fact or analysis begins with a preconceived notion by the researcher about th e structure of the data, with this being founded on theory or previous research. This a pr iori notion sets limits to the analysis in terms of component estimation or the number of factors to be ex tracted. Moreover the analysis adopts the arrangement of a stru ctural equation model (SEM), incorporating

PAGE 49

37 more complex systems of correlations, which are based on prior information (Hair, et al. 2003) Given limited research on the latent cons tructs of most variables in market research and the tautological complexities of SEMs, the majority of factor analyses tend to follow the exploratory perspective. This is the same perspective taken by this study. Prior to performing a factor analysis, th e researcher must perform a correlation analysis of the variables. Usually, this entails inspection of the Pearson correlation matrix together with the respective significance levels of these correlations. If correlation coefficients are greater than .6 and si gnificant at 0.01, then this is regarded meritorious. However, given the large number of variables that ar e often included in factor analyses, correlations of this magnitude are rare; hence lower levels of correlations are considered to be acceptable if the Ka izer-Meyer-Olkin (KMO ) measure of sampling adequacy and the Bartletts test of sphericity are satisfactory. According to Hair et al. (2003) the KMO measure of sampling adequacy assesses the factorability of the vari ables by quantifying the level of inter-correlations among the variables. It tests if each pa rtial correlation is small and if it is significant, to give a summarized index of the individual correla tion coefficients. The KMO measure ranges from 0 to 1; with 0 signifying absolute lack of correlation (i.e. in appropriate for factor analysis) and 1 signifying perfect correlation. The Bartlett test of sphericity is somewhat similar, providing the statistical probability that the correlation ma trix, as a whole, is correlated to the individual variables. Esse ntially, the Bartlett test assesses if the correlation matrix is an identity matrix, which if so would make factor analysis inappropriate. These measures are used to dete rmine if factor analysis is appropriate in this study.

PAGE 50

38 Factor scores are derived and the scree pl ot decision rule is employed to determine how many factors to extract, because the variable s used are less than 20. In addition, as a cautionary measure, the acceptable extraction ra te is set at approximately 70 percent or greater. Usually, the latent root criterion would be used to determine how many factors to extract, with each factor having a latent root /eigenvalue of greater than 1. The rationale behind this is that each variable contributes a value of 1 to the total eigenvalue, thus each factor would have to contribute more vari ance than a single variable if it is to be considered significant. However, this is most applicable to factor an alyses involving 20 to 50 variables (Hair et al., 2003) If less than 20 variables ar e involved, then this decision rule is likely to extract too few factors; hence it is not used in this study. In addition, Varimax rotations are used in this study to facilitate naming of the derived factors. Fundamentally, a factor rota tion is an adjustment of the factor axes, which results in the transformation of the deri ved factors through the redistribution of the variance extracted. The total variance extract ed is maintained and does not change, and only the position of each factor in the rotated plane is shifted; thereby making it easier to interpret associations and name the factors. Different forms of factor rotations can be performed, but the Varimax is widely used b ecause it is a 90 degree rotation that leads to orthogonal factors. This eliminates chances of multicollinearity w ithout complicating the interpretation of each factor. Loadings tend to wards 0 or with 0 signifying a lack of association and signifying hi gh degrees of association. The signs on the loadings only serve directional purpose, to indicate the natu re of the relationship and not the magnitude. Various other rotations can be performed in fact or analysis but these are irrelevant to this

PAGE 51

39 study and beyond the scope of this thesis. On ly the Varimax rotation is of significance here.

PAGE 52

40 CHAPTER 5 AUCTION BID ANALYSIS AND RESULTS The manner in which the collected data were analyzed can be broken into two categories: i) univariate-parametric statisti cal analyses of the WTP bids and ii) the econometric analyses of WTP bi ds. The latter entailed a f actor analysis followed by a double hurdle probit model analysis. This chapte r presents the former (i.e. the univariate parametric statistical analysis of the auction bids). Both the apple and tomato auction bids were initially evaluated separately. As previously alluded to, four rounds of biddi ng had occurred in each auction. The amount that each participant bid was the premium s/ he was willing to pay in order to exchange one pound of unlabeled fresh apples (tomatoes) that s/he was initially endowed with, for one pound of fresh apples (tomatoes) labele d U.S.A. Grown. The bidding progression for all rounds in all locations are combined and outlined in the line graph in Figure 5-1. Also, the analysis of bidding progressions by location was done and results of this are shown in Appendix H In the fresh apples line graph, in Figure 5-1, the average bid for the first round was considerably lower than that of all four rounds combined. Ot herwise the subsequent bids converged at the approximate average of $0.48. With respect to tomatoes, the last bid is the one that was relatively deviant. The firs t three rounds of bidding had a mean premium of approximately $0.44 while the last average bi d jumped to $0.53. This invariably raised the mean bid, making the average for al l four bids to be $0.46. A look at Appendix H reveals that the unprecedented deviation may have been due to a spike recorded in the

PAGE 53

41 fourth bid in the Gainesville, FL data. Nevertheless, increases from the third to the fourth bid were also recorded in the other locations. Progression of Mean bidsAll Locations 0.30 0.35 0.40 0.45 0.50 0.55Round1Round2Round3Round4Round of bidding $/Lb Tomatoes (n=175) Apples (m=136) Figure 5-1. Line graph showi ng the progression of combined bids in both tomato and apple auctions The mean of the four bids as well as the mean of the last two bids (i.e. PMTi, which is the variable used as the quantitative WTP variable in subsequent econometric analyses) are shown in Table 5-1. The m ean of the last two bids (PMTi) is used as the dependent variable in the econometric analys is instead of the mean of a ll four bids because the last two bids are considered to be better estimates of the WTP. This is because in the first few rounds of bidding, consumers learn the nature of their WTP for the product attribute at auction as value formation takes place. Thus bids often change substantially from one round to the next before stabi lizing at the true WTP value after the first few rounds have passed. According to Cox et al. (1985) and Fox et al. in Caswell (1995) this is why the Vickrey auction with multiple rounds is bett er suited to obtain true WTP because it accommodates for value formation. Shogren et al. (2000) add that the first rounds of

PAGE 54

42 bidding in a Vickrey auction may reveal less accurate WTP measures because the value formation process may take place during th ese rounds, especially if a new unfamiliar product is at auction. They refer to this as the lab novelty effect. Table 5-1. Average WTP for apples (n = 136) and tomatoes (n = 175) Mean Standard Deviation All four bids 3rd and 4th round bids (PMTi) All four bids 3rd and 4th round bids Apples $0.48 $0.49 $0.55 $0.58 Tomatoes $0.46 $0.48 $0.53 $0.55 As shown in Table 5-1, the mean WTP for COOL in apples and tomatoes is approximately $0.49 and $0.48 respectively. The standard deviations in both cases are high, 0.58 and 0.55 respectively. This high le vel of dispersion shows that different consumers have distinctly different levels of WTP (i.e., it can be viewed as an indicator of consumer surplus). Univariate hypothesis testing of the mean WTP (i.e. PMTi) proves the bids to be significantly greater than zero. Thus: H0: PMTi = 0 Ha: PMTi > 0 (One-tailed Test) 0.05 is chosen as the significance level ( = 0.05) For fresh apples: 85 9 136 58 0 0 49 0 ) 0 ( se t Therefore, I reject the nu ll hypothesis that the WTP for COOL in fresh apples is equal to 0 since the t-value yi elds a probability value < 0.001 Similarly for fresh tomatoes:

PAGE 55

43 58 11 176 55 0 0 48 0 ) 0 ( se t which also yields a probability value < 0.001. Therefore the null hypothesis th at the WTP for COOL in fr esh tomatoes equals 0 is rejected. On calculating the means for only those c onsumers who were willing to pay more than $0.00 for either fresh apples or fresh to matoes labeled U.S.A. Grown, the expected increase was registered as shown in Table 52. This calculation was done to give insight on the existing differentials between the sub-sa mple of those willing to pay and the whole sample. Table 5-2. Average WTP for apples (n = 108) and tomatoes (n = 126): Sampling only those consumers who were WTP more than $0.00 Mean Standard Deviation All four bids 3rd and 4th round bids All four bids 3rd and 4th round bids Apples $0.60 $0.61 $0.56 $0.59 Tomatoes $0.64 $0.68 $0.53 $0.54 Overall, 79 percent of the consumers were willing to pay more than $0.00 for fresh apples labeled U.S.A. Grown while 72 per cent were willing to pay a premium in the case of fresh tomatoes labeled U.S.A. Grow n. Detailed analysis by location revealed that in general, consumers in Lansing were w illing to pay substantially less for the label U.S.A. Grown irrespective of the product under consideration. As shown in table 5-3, it was only the consumers in Gainesville who seemed to have a high WTP for COOL in fresh tomatoes, while those in Atlanta ha d a high WTP for COOL in fresh apples. Using the findings on the price premium means (PMTi) for fresh apples and tomatoes, a comparison was made to test if these means were statistically different in

PAGE 56

44 order to begin testing of the second hypothe sis in this study. As noted earlier, the hypothesis is: 2. If consumers are willing to pay a premiu m for fresh apples and fresh tomatoes labeled U.S.A. Grown then the premiu ms will be product specific and unequal. Table 5-3. Mean WTP across location Gainesville, FL ($/Lb) Lansing, MI ($/Lb) Atlanta, GA ($/Lb) Apples 0.41 0.18 0.64 Tomatoes 0.78 0.20 0.39 In addition, it was imperative do so because subsequent econometric analyses entailed combining the two data sets with the premiums as the dependent variable. Combining the two would only be statistica lly plausible if the two data sets PMTi were independent of the product treatment that th e respondents received. In other words, the premium that consumers were willing to pa y could not be a function of whether the product they bid for was an apple or a tomato. If it was, then the PMTi would be regarded as two different variables making it implausi ble to combine them. Thus the z-test for independent samples was performed: H0: PMTapples = PMTtomatoes Ha: PMTapples PMTtomatoes Choosing a significance level of 0.05 ( = 0.05) Criterion: Reject H0 if z < -1.96 or z > 1.96, where

PAGE 57

45 tomatoes tomatoes apples apples tomatoes applesn n PMT PMT z2 2 Calculation: 154 0 176 303 0 136 337 0 48 0 49 0 z Since z = 0.154 < 1.96 we fail to reject the null hypothesis that the two means are not different. This analysis of the data sugge sts premium equivalency in fresh apples and tomatoes labeled Grown in the U.S. Nevertheless, due to the inherent logical asymmetry between the null and a lternative hypotheses, we cannot stat istically confirm equivalence. To confirm equivalence, the following two-sa mple t-test for equivalence is performed: First, I assume that th e distributions of PMTi observations follow the basic parametric model i.e. they are normal w ith a common variance and potentially unequal expected values. Mathematically, m i N Xi,..., 1 ) ( ~2 n j N Yj,..., 1 ) ( ~2 with , IR IR2 where X is the premium bid for apples and Y is that for tomatoes; m is the sample size of the apple data and n is the sample size of the tomato data, while is the mean PMTi for apples and that for tomatoes. 2is the common variance of the PMTi which is a positive real number. This is also assumed for both previous hypothe sis tests and all othe r analyses in this study. I follow Wellek (2003) and define equivalence of th e apples treatment to the tomatoes treatment by the condition that the difference between the standardized

PAGE 58

46 premium means falls in a sufficiently narrow interval ( 1, 2) in the neighborhood of zero. Thus I then formulate th e hypothesis testing problem as H0 : 1) ( or 2) ( versus Ha : 2 1) ( ( 1, 2>0) where is the mean PMTi for apples and that for tomatoes; is the standard error of the difference in PMTi. A critical region, which is synonymous to the confidence interval in the traditional t-test for independent samples, is then se lected based on the researchers discretion. Discretion here would depend on how narrow the desired equivalence interval is to be. Here, I choose an alpha level of 0.05 and cal culate the region using the SAS program (see Appendix D ) because this alpha level corresponds to the narrowest interval, compared to other commonly used critical regi ons for this kind of hypothesis test. The test statistic is calculated as 2 1 1 2 1 2. 2 n j j m i iY Y X X Y X n m n m mn T where Xi is the WTP for apples and Yi is that for the tomatoes; m is the sample size of the apples data (136) and n that for tomatoes ( 175). Thus we would re ject the null hypothesis of non-equivalence of the premium for apples to the one for tomatoes if the calculated Tstatistic falls within the critical region. The resulting T-value from the data is 0. 045. This signifies premium equivalence since it falls within the critical region of -2.72 and 7.04. The null hypothesis of nonequivalence of the premium for apples to the one for tomatoes is rejected. See Appendix D for detailed results from the SAS program. It is concluded that the mean premium that

PAGE 59

47 consumers are willing to pay for apples labe led U.S.A. Grown over unlabeled apples is equivalent to the mean premium they are wi lling to pay for tomatoes labeled U.S.A. Grown over unlabeled tomatoes. The detailed im plication of this finding is discussed in Chapter 9.

PAGE 60

48 CHAPTER 6 EMPIRICAL SPECIFICA TION AND RESULTS Empirical Specification To address the last hypothesis of this st udy, the factors influencing consumers WTP for the label U.S.A. Grown were an alyzed using the doubl e hurdle probit model specification in (6-1). Additional consideration of demographic variables was also included in the analysis. i i i i i i i i i i i i i i i i i i i iu Qual Pfr Safe Trust PC Expose Inc Loc Edu Gender Age WTP 20 19 18 16 15 14 13 11 9 7 10 6 5 4 3 2 1 (6-1) where WTPi is the dichotomous willingness to pay (i.e. participation dependent variable), expressed as a probability For the second hurdle PMTi replaces WTPi, where PMTi is the quantitative willingness to pay (i.e. consumption dependent variable) and i takes the place of iu as the error term. Age = Age of respondent Gender = Gender of the respondent Edu = Highest level of edu cation completed by respondent Loc= Location (Atlanta, Ga inesville, or Lansing) Inc = Income group Expose = Self rating on exposure to food sa fety information in fresh fruit and vegetables

PAGE 61

49 PC = Presence of children unde r age of 16 in the household Trust = Extent of respondents trust in information about food production obtained from U.S. Government Agencies (e.g. USDA, FDA, EPA, etc.) Safe = Perceptions about food safety Pfr = Food preferences factor scores Qual = Food quality factor scores Description of Variables A detailed description of the variables is provided in Appendix E Age was measured in terms of the number of year s that the respondent had, while gender was measured as a dichotomous variable, denoting the respondents sex. In all of the models estimated in this study, the male sex was dropped and used as the base1 variable. Females were expected to be willing to pay more for COOL since they are considered to be more concerned about details or food safety, wh ich would normally induce a desire to know the products origin. Similarly, older consumers were expected to be more willing to pay for COOL for the same reasons. The highest level of education completed by the respondent is a variable measuring how well-educated a respondent is, based on the formal education system. Given the ordinal nature of the edu cation variable, 3 dummy vari ables (EDU1, EDU2 and EDU3) had to be created, while taking into account th e distribution of the education variable in the samples analyzed; (most respondents were relatively well-educated thus necessitating the creation of fewer dummy variables). Th e University postgraduate degree (EDU3) dummy variable was dropped and became the base in all model estimations. 1 In all of the models estimated in this thesis, th e same base was used; this is the case for all dummy variables i.e. GENDER, EDU, LOC, INC and PC.

PAGE 62

50 Three dummy variables were also create d for the respondents location/state of residence given that location is a nominal variable (LOC1, LOC2 and LOC3). In this case, Atlanta, GA (LOC 3) was used as a base by dropping it out of the estimations. Prior to model estimation, it was anticipated that consumers from Gainesville, FL would be more willing to pay for COOL because mandato ry COOL has been prevalent in Florida for the past 26 years, at the state level and it would be expected that Floridian consumers are accustomed to COOL. In addition, it was expected that consumers from Lansing, MI would be willing to pay for CO OL particularly in the case of apples, since Michigan apples may compete with apples from other countries. The premise here was that Michigan consumers would want to suppor t local producers and since COOL would enable identification of product origin, cons umers would then be willing to pay for COOL in apples. Income is another ordinal variable, whic h was included in the model specification. In this case, 4 dummy variables (INC1... INC4) were created based on the income distribution of the samples analyzed; most respondents had relatively high incomes. The highest income bracket (INC4) $100,000 a nd above was dropped, thus making it the base. Prior to estimation of the model, cons umers with higher incomes were thought to be more willing to pay for COOL, based on th e notion that labeled produce would be regarded as a type of luxury good when comp ared to unlabeled produce. In addition, high income consumers were expected to have a higher marginal propensity to spend. Economic theory shows the high income c onsumers usually have a higher marginal propensity to spend, thus it would be more likely to find high income consumers who are willing to pay for COOL.

PAGE 63

51 The presence of children under the age of 16 years in the respondents household was another demographic included in the mode l. Four dummy variables were created and the highest or more children (PC4) wa s dropped to make it the base variable. Anticipated results were that consumers w ith more children would be more concerned about food safety and the origin of the food they feed to th eir children. Thus, they would be more willing to pay a premium for COOL. Inclusion of all these demographic va riables was grounded on economic theory. Consumer demand theory maintains that cons umers make expenditure choices with the objective of maximizing utility subject to ec onomic constraints. Th ese constraints arise from scarcity, which inevitably leads to the economic problem of choice, while the objective utility function stems from endogenous preferences/desires, which may differ from one individual to the other. Hence, di fferent consumers may obtain different levels of utility from the same m easure of the same product attr ibute. In other words, the endogenous preferences determine the unique level of consumer-surplus that each consumer may attain from a product attribute, if purchased at a given market price; consumer surplus arises from the difference between WTP and the markets equilibrium price. As outlined in the theoretical framework in Chapter 4, demographics influence the nature of an individuals endogenous preferences, which de termine the utility s/he can obtain from a particular product attribute. This is why, for example, older consumers will normally demand different product attributes in comparison to younger consumers (e.g. more health-oriented produc t attributes) and females will normally demand different attributes compared to males.

PAGE 64

52 Therefore, demographics (such as age, ge nder, income, level of education, location and number of children in the household) we re appropriately incl uded as explanatory variables in the model specification. In additi on, the choice of these explanatory variables followed several previous studies, which have hypothesized them to be major drivers of WTP for COOL (e.g. Schupp and Gillespie, 2001; Lour eiro and Umberger, 2002; Lusk et al., 2003) Based on consumer demand theory and cons umer cognitive theory, psychographics were included as explanatory variables in the specified m odel. Also, previous studies have included similar psychographics in their analyses of WTP for COOL (e.g. Loureiro and Umberger, 2002; Umberger et al., 2003) EXPOSE is a psychographic variable that was included in the model specification, and it is an ordinal Likert-s cale rating, indicating the level of exposure to information on food safety in fruits and vegetables that respondents had previously received. This vari able was included because theory suggests that the level of information that a consum er has been exposed to, affects his/her WTP decisions. Perceptions about trust, food safe ty, food preferences and food quality were also included as explanatory psychographic vari ables on the same basis. As alluded to in Chapter 4, psychographic variables such as these may have a significant bearing on consumers WTP. Trust was measured by a Likert scale ra ting, indicating how trusting consumers were of information they receive from U.S. agencies (e.g. USDA, FDA, EPA, etc.) about food production. Given that COOL is a cr edence attribute whos e verification and enforcement is in the hands of such agencies, it was necessary to include this variable in

PAGE 65

53 the model specification in order to assess th e impact it may have on WTP for a credence attribute such as COOL. Consumer food safety concerns in fresh fr uits and vegetables were also included, because COOL can be associated with food sa fety. The origin of a product will usually influence whether the product is safe for cons umption, because the products contact with its environment of origin can contaminate the product with harmful bacteria or viruses. Based on this, consumers that think about the food safety, when making purchasing decisions, were expected to be willing to pay more to know the products country of origin. Consumer preferences were also included, and factor scores we re generated from 10 psychographic questions related to food pr eferences, thus creati ng 3 factors to be included in the model specification. The 3 quantitative factors on consumers food preferences (PFR1, PFR2, PFR3) were na med Open to unfamiliar foods, Choosey and Afraid of unfamiliar foods, respectiv ely (see Chapter 7 for details). The expectation was that conservative consumers who prefer to consume local produce or are less adventurous in the kinds of foods they will eat, would be willing to pay a premium for produce labeled U.S.A. Grown. In the case of consumers perceptions about food quality, 2 factors (QUAL1 and QUAL2) were generated from 7 psychogra phic questions related to food quality perceptions and these were named General quality conscious and Natural quality conscious, respectively. It was anticipate d that consumers who were more quality conscious would be willing to pay more fo r the produce labeled U.S.A. Grown, and the premise being that U.S. produce would be rega rded as better quality produce. Consumers

PAGE 66

54 inclined to naturally produced foods were e xpected to be willing to pay even more for COOL, as these consumers are assumed to be relatively more particular about the origin of the produce they consume. Organization of Model Estimations Prior to the estimation of the model in (6-1), a variant thereof was specified and estimated where the factor scores on food preferences and food quality ( Pfr and Qual ) were not incorporated. Instead, singl e variables on food preferences ( DFF ) and food quality ( PAYQ ) replaced them. Both of these were based on consumer Likert-scale ratings (nine-point scale) indicating the consumers le vel of agreement with the statements: I like foods from different c ountries and I am willing to pay somewhat more for a product of better quali ty (refer to questionnaire in Appendix B ). In addition, the model was first estimated se parately for apples and then separately for the tomato data; hence the presentation of the results is organized into the apple and then tomato sections. A third section is then presented, where the estimation was performed for the combined apple and tomato data set. Combining the two data sets was done since the analysis of the bids by consum ers in Chapter 5 indicated that the WTP was not product specific. This made it statistically plausible to combine the two data sets and analyze them as one data set. The results pr esented in this chapter are only for models without factor scores. Models without Factor Scores Apple Model (Model 1) The apple model without factor scores is presented in Tables 6-1 and 6-2 below. Table 6-1 shows the probit part of the estimation, where an approximate chi-square value of 6.6 with 16 degrees of freedom yielded a si gnificance level of 0.98. Thus, this model is

PAGE 67

55 insignificant. Though none of the explanatory va riables are statistica lly significant, the probit model correctly predicted 78.6 pe rcent of the consumers responses. Also, despite the insignificance of the probit estimation, the truncated estimation has been presented in Table 6-2. In this pa rt of the estimation, as well as all other estimations of the second stages of the double hurdle models in this study, the chisquared specification test was performed to ev aluate if the truncated tobit estimation was a better fit than the tobit. The formula used is, tobit truncation probitL L Lln ln ln 22 and the truncated tobit proved to be a superior fit (See Appendix G for details). Age, gender, income and location were all significant at 90 percent confidence level while the rest of the explanatory variab les were insignificant. In terms of age the marginal effects show that for every year that a consumer is younger, they would be willing to pay an extra $0.01 for the U.S.A. Grown labeling. Females were willing to pay approximately $0.35 more than the males and consumers in Lansing MI were willing to pay considerably less than the consumers in Atlanta GA or Gainesville, FL. Regarding the income levels, only consumers who ear ned below $50,000 per year were willing to pay more for the labels, otherw ise all other income groups we re not willing to pay more for COOL.

PAGE 68

56Table 6-1. Apples probit m odel without factor scores (Model 1) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.016 0.014 0.00 -1.1180 0.2637 45.22 ENDER 0.218 0.439 0.06 0.4980 0.6187 0.90 EDU1 0.030 0.367 0.01 0.0820 0.9347 0.35 EDU2 0.128 0.343 0.03 0.3730 0.7088 0.38 LOC1 -0.109 0.348 -0.03 -0.3130 0.7544 0.41 LOC2 0.484 0.482 0.11 1.0050 0.3149 0.13 INC1 0.196 0.377 0.05 0.5200 0.6031 0.26 INC2 -0.020 0.376 -0.01 -0.0530 0.9574 0.19 INC3 0.167 0.369 0.04 0.4540 0.6500 0.22 EXPOSE -0.049 0.215 -0.01 -0.2300 0.8183 1.90 PC1 0.414 0.424 0.10 0.9760 0.3290 0.26 PC2 0.505 0.462 0.12 1.0920 0.2749 0.15 PC3 0.054 0.339 0.01 0.1600 0.8732 0.37 TRUST 0.073 0.094 0.02 0.7700 0.4411 3.69 SAFE 0.027 0.089 0.01 0.2980 0.7659 3.70 DFF -0.017 0.064 0.00 -0.2680 0.7884 6.90 PAYQ 0.117 0.094 0.03 1.2510 0.2110 7.29 Restricted log likelihood value, ln L10 = -69.15 R2 (McFadden, 1973) = .0479 Maximum unrestricted log likelihood value, ln L1 = -65.83 R2 (Estrella, 1998)= .0487 Log likelihood 2 (df=16)= 6.629 (p = 0.9800) % of correct predictions = 78.6 Number of observations = 136

PAGE 69

57Table 6-2. Apples truncated tob it model without factor scores (Model 1) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0317 0.0133 -0.01 -2.3890 0.0169 44.64 GENDER 1.1218 0.6012 0.35 1.8660 0.0621 0.91 EDU1 0.3950 0.3829 0.12 1.0320 0.3023 0.35 EDU2 0.1335 0.3369 0.04 0.3960 0.6919 0.38 LOC1 -0.2111 0.3447 -0.07 -0.6120 0.5403 0.40 LOC2 -1.9665 0.6501 -0.62 -3.0250 0.0025 0.14 INC1 0.6580 0.3314 0.21 1.9860 0.0470 0.27 INC2 -0.1328 0.3572 -0.04 -0.3720 0.7101 0.19 INC3 -0.2104 0.3681 -0.07 -0.5710 0.5677 0.23 EXPOSE -0.2729 0.2062 -0.09 -1.3240 0.1856 1.90 PC1 0.0042 0.3716 0.00 0.0110 0.9909 0.26 PC2 -0.1079 0.4299 -0.03 -0.2510 0.8019 0.17 PC3 0.3531 0.3325 0.11 1.0620 0.2882 0.36 TRUST 0.0403 0.0869 0.01 0.4640 0.6427 3.74 SAFE 0.1053 0.0912 0.03 1.1550 0.2481 3.72 DFF -0.0271 0.0646 -0.01 -0.4190 0.6752 6.88 PAYQ 0.0461 0.0904 0.01 0.5100 0.6101 7.36 Sigma 0.7348 0.0979 7.5060 0.0000 Number of observations = 136 Log likelihood function = -35.50 Observation after truncation = 108 Threshold values for model: Lower = 0 Upper = +

PAGE 70

58Tomatoes Model (Model 2) In the case of the tomatoes data, the prob it estimation without factor scores had a pvalue of 0.11. Only consumers perceptions about food quality and lo cation turned out to be significant independent variables. Consumer s, who indicated in th e questionnaire that they were willing to pay somewhat more fo r a product of better qua lity, actually had a higher probability of being willing to pay a premium for tomatoes labeled U.S.A. Grown, according to auctions data analys is. Also, consumers who were located in Lansing MI were least likely to be willing to pay a premium for the tomatoes labeled U.S.A. Grown in contrast to consumers in Gainesville, FL who we re most likely to pay a premium. Table 6-3 below presen ts the probit model estimation. Table 6-3. Tomatoes probit m odel without factor scores (Model 2) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0098 0.0147 0.00 -0.671 0.5024 44.21 GENDER -0.1092 0.3260 -0.03 -0.335 0.7376 0.87 EDU1 0.0788 0.2903 0.03 0.272 0.7859 0.33 EDU2 0.4440 0.2824 0.14 1.572 0.1159 0.43 LOC1 0.5479 0.2866 0.17 1.912 0.0559 0.38 LOC2 -0.2992 0.2775 -0.10 -1.078 0.2809 0.28 INC1 -0.1154 0.3048 -0.04 -0.379 0.7049 0.27 INC2 -0.2644 0.2994 -0.09 -0.883 0.3771 0.28 INC3 -0.0809 0.3675 -0.03 -0.22 0.8257 0.14 EXPOSE 0.0856 0.1420 0.03 0.603 0.5467 1.91 PC1 -0.3590 0.4147 -0.12 -0.866 0.3867 0.23 PC2 -0.2022 0.3841 -0.07 -0.526 0.5985 0.22 PC3 -0.2482 0.3175 -0.08 -0.782 0.4343 0.37 TRUST 0.0803 0.0697 0.03 1.152 0.2495 3.71 SAFE 0.0282 0.0698 0.01 0.404 0.6859 3.41 DFF -0.0747 0.0559 -0.02 -1.336 0.1815 6.99 PAYQ 0.1534 0.0847 0.05 1.812 0.0700 7.32 Restricted log likelihood value, ln L10 = -104.70 R2 (McFadden, 1973) = 0.1111 Maximum unrestricted log likelihood value, ln L1 = -93.06 R2 (Estrella, 1998)= 0.1315 Log likelihood 2 (df=16)= 23.28 (p = 0.1064) % of correct predictions = 74.2 Number of observations = 175

PAGE 71

59 As seen above, the model had an R2 (McFadden, 1973) of 0.11 which is considered feeble for cross sectional data such as that used in this analysis. Nevertheless, the model had a 74% correct prediction rate. In the truncated part of the model, the consumers food quality perceptions variable was insignificant while some demogra phic variables turned out to be significant. Table 6-4 below has more details on this. Those consumers whose highest level of education completed was some college or less (EDU1) were willing to pay $0.20 less than the base (Postgraduate degree). Also, consumers in Lansing MI were once again shown to be willing to pay less for the labeling (this time $0.27 less). Those consumers, who had indicated that they have seen, read or heard less about food safety in fresh fruits and vegetables were willing to pay $0. 09 less for COOL. Consumers who indicated a high level of trust for U.S. government agen cies such as the USDA, FDA, EPA etc. showed a greater willingness to pay for Grown in the U.S fresh tomatoes. Table 6-4. Tomatoes truncated tobit model without factor scores (Model 2) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0005 0.0111 0.00 -0.041 0.9674 44.26 GENDER 0.2288 0.2918 0.10 0.784 0.4330 0.86 EDU1 -0.4375 0.2336 -0.20 -1.873 0.0611 0.32 EDU2 -0.2519 0.2147 -0.11 -1.173 0.2407 0.46 LOC1 0.5136 0.2220 0.23 2.314 0.0207 0.45 LOC2 -0.6104 0.3421 -0.27 -1.784 0.0744 0.22 INC1 0.2859 0.2492 0.13 1.147 0.2513 0.27 INC2 0.3649 0.2514 0.16 1.451 0.1467 0.27 INC3 0.2968 0.2902 0.13 1.023 0.3064 0.14 EXPOSE -0.2085 0.1266 -0.09 -1.647 0.0996 1.94 PC1 -0.3528 0.3244 -0.16 -1.087 0.2769 0.22 PC2 -0.1270 0.2805 -0.06 -0.453 0.6507 0.22 PC3 -0.2249 0.2537 -0.10 -0.887 0.3753 0.36 TRUST 0.1407 0.0646 0.06 2.176 0.0296 3.86 SAFE 0.0277 0.0558 0.01 0.495 0.6203 3.46 DFF -0.0209 0.0560 -0.01 -0.373 0.7094 6.99 PAYQ 0.0124 0.0768 0.01 0.162 0.8717 7.49 Sigma 0.6382 0.0791 8.064 0.0000 Number of observations = 175 Log likelihood function = -49.70 Observation after truncation = 125 Threshold values for model: Lower = 0 Upper = +

PAGE 72

60 Combined Apples and Tomatoes Model (Model 3) When the data were combined into one data set and the same model estimated, similar results were obtained. The probit part of the estimation was found to be significant with a log likeli hood chi-square value of 24.11 at 16 degrees of freedom, which corresponds to a p-value of 0.087. Food qua lity concerns and age turned out to be the only variables that were significant, at a 0.1 alpha level. Margin al probabilities show that for every year older that the consumer is the probability of being willing to pay a premium for COOL is decreased by 0.6%. Th is is holding all dummy variables at the base; the base being a female with a univers ity postgraduate degree education, household income of greater than $150,000, with thr ee or more children in the household and located in Atlanta GA. Table 6-5. Combined apples and tomato es probit model without factor scores (Model 3) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0171 0.0098 -0.01 -1.739 0.082 44.65 GENDER -0.0666 0.2476 -0.02 -0.269 0.7878 0.88 EDU1 0.1142 0.2195 0.04 0.52 0.603 0.34 EDU2 0.2848 0.2122 0.10 1.342 0.1796 0.41 LOC1 0.3288 0.2130 0.11 1.544 0.1227 0.40 LOC2 -0.2251 0.2184 -0.08 -1.03 0.3028 0.21 INC1 0.0270 0.2292 0.01 0.118 0.9061 0.27 INC2 -0.1705 0.2269 -0.06 -0.752 0.4522 0.24 INC3 0.0211 0.2520 0.01 0.084 0.9334 0.17 EXPOSE 0.0631 0.1149 0.02 0.549 0.583 1.89 PC1 -0.0303 0.2814 -0.01 -0.108 0.9144 0.24 PC2 0.0165 0.2735 0.01 0.06 0.952 0.19 PC3 -0.1219 0.2236 -0.04 -0.545 0.5856 0.37 TRUST 0.0656 0.0532 0.02 1.232 0.2179 3.53 SAFE 0.0514 0.0530 0.02 0.97 0.3318 3.71 DFF -0.0354 0.0404 -0.01 -0.875 0.3815 6.95 PAYQ 0.1488 0.0592 0.05 2.512 0.012 7.31 Restricted log likelihood value, ln L10 = -175.16 R2 (McFadden, 1973) = .06881 Maximum unrestricted log likelihood value, ln L1 = -163.1073 R2 (Estrella, 1998)= .07717 Log likelihood 2 (df=16)= 24.10601 (p = .8721459E-01) % of correct predictions = 75.2 Number of observations = 311

PAGE 73

61 The truncated tobit part of the analysis is presented in the Table 6-6. According to this model, it is noted that location, level of exposure to information about food safety in fruits and vegetables as well food safety c oncerns are significant variables at a 0.1 significance level. Consumers in Lansing MI we re once more shown to be less willing to pay for COOL with marginal effects at the me ans showing that they were willing to pay $0.49 less than what the consumers in Atlanta GA were willing to pay. Those consumers who rated themselves as having seen or heard or read less about food safety in fresh fruits and vegetables were also found to be willing to pay marginally less for fresh tomatoes labeled U.S.A. Grown. In contrast, those w ho said they thought about food safety when purchasing fresh fruits and vegetables were found to be willing to pay more for produce labeled U.S.A. Grown. This may imply th at these consumers consider U.S. produce safer hence they were willing to pay more for produce with COOL. Table 6-6. Combined apples and tomatoes tr uncated tobit model without factor scores (Model 3 Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0192 0.0123 -0.01 -1.558 0.1193 44.43 GENDER 0.6725 0.4091 0.20 1.644 0.1002 0.88 EDU1 -0.0747 0.2624 -0.02 -0.285 0.7759 0.33 EDU2 -0.0975 0.2540 -0.03 -0.384 0.7010 0.42 LOC1 0.1938 0.2331 0.06 0.832 0.4057 0.42 LOC2 -1.6638 0.5437 -0.49 -3.060 0.0022 0.18 INC1 0.4257 0.2763 0.12 1.541 0.1234 0.27 INC2 0.2584 0.2800 0.08 0.923 0.3561 0.23 INC3 -0.0163 0.3150 0.00 -0.052 0.9586 0.18 EXPOSE -0.2727 0.1542 -0.08 -1.769 0.0769 1.92 PC1 -0.0703 0.3274 -0.02 -0.215 0.8300 0.24 PC2 -0.0024 0.3145 0.00 -0.008 0.9939 0.20 PC3 0.1562 0.2675 0.04 0.584 0.5593 0.36 TRUST 0.0265 0.0646 0.01 0.410 0.6821 3.59 SAFE 0.1814 0.0779 0.05 2.330 0.0198 3.79 DFF -0.0640 0.0585 -0.02 -1.093 0.2742 6.94 PAYQ 0.0280 0.0802 0.01 0.349 0.7271 7.43 Sigma 0.8516 0.1063 8.014 0.0000 Number of observations = 311 Log likelihood function = -102.78 Observation after truncation = 233 Threshold values for model: Lower = 0 Upper = +

PAGE 74

62 Overall, the models without factor scor es all suggested an association between WTP for produce labeled Grown in the U.S. and some of the demographic variables as well as the quality and food safety related vari ables. In all models without factor scores food preferences and presence of children in a household were insignificant implying that these have no effect on the WTP for COOL in fresh apples or tomatoes. Most of the models without factor scores were significant, having a rela tively better predictive power than the nave alternative.

PAGE 75

63 CHAPTER 7 FACTOR ANALYSIS RESULTS The models with factor scores are presente d in Chapter 8, and include factor scores for both food quality and food preference variables derived from principle component factor analyses. Initially, they were also s upposed to include food sa fety factor score(s). However, results from the factor analysis of food-safety-concern variables showed low correlations plus extremely low Kaiser-M eyer-Olkin (KMO) measures of sampling adequacy of less than 0.6 (see Appendix F ). This necessitated the use of a single question from the questionnaire as a food safety variab le in the final model specifications. For those variables for which factor analysis is used to derive factor scores, results are presented in this chapter. Table 7-1 summari zes results of the first factor analysis. The table shows that when food-quality-related vari ables were factor analyzed for the apple data, two factors were extracted. This numbe r of factors extracted was established based on the scree plot analysis decision rule, and in the process a 77 percent extraction rate was achieved as well as a 0.843 KMO meas ure of sampling adequacy. The factor loadings distributed evenly between the tw o factors named general quality conscious and natural quality conscious. As is true in all factor analyses, naming of factors was solely based on the researchers intuition. Thus, the names should be viewed in that context. Factor scores were saved using th e regression method for la ter use in the doublehurdle models presented in Chapter 8.

PAGE 76

64 Table 7-1. Rotated component matrix(a) fo r food quality proxy va riables-apple data Factor Proxy Variable 1 (General Quality conscious) 2 (Natural Quality conscious) I usually aim to eat natural food .227 .813 I am willing to pay somewhat more for a product of better quality .764 .333 Quality is decisive for me in purchasing foods .919 .181 I always aim at the best quality .860 .343 When choosing foods, I try to buy products that do not contain residuals of herbicides and antibiotics .271 .799 I am willing to pay somewhat more for food containing natural ingredients .396 .825 For me, wholesome nutrition begins with the purchase of foods of high quality .704 .462 Table 7-2 shows results of th e second factor analysis. In this case, food preference variables were reduced to yield three factor s and these being extracted also based on a scree plot decision rule. This time, a 69 pe rcent extraction rate was achieved with a high KMO measure of sampling adequacy of 0.86. Scree plots and extraction percentages of all factor analyses in this study are shown in Appendix F Again, names were intuitively assigned to the factors derived. Notably, th e variable, I will eat almost anything, had relatively high loadings on two factors at the same time, i. e. Open to unfamiliar foods and Choosey. This made naming of the factors a little more challenging. In the tomato data, two factors were also derived for the quality variables with 73 percent of the variance in the original vari ables being captured by th e factors. Table 7-3 shows how the variables fact or loadings distributed.

PAGE 77

65 Table 7-2. Rotated component matrix(a) fo r food preference proxy variables-apple data Factor Proxy Variable 1 (Open to unfamiliar foods) 2 (Choosey) 3 (Afraid of unfamiliar foods) I like foods from different countries .712 -.045 -.161 Ethnic food looks too weird to eat -.279 .040 .778 I like to try new ethnic restaurants .832 -.165 -.288 At parties, I will try a new food .852 -.224 -.155 I am very particular about the foods I will eat -.190 765 .135 I am constantly sampling new and different foods .685 -.280 -.382 I don't trust new foods -.285 .283 .700 I will eat almost anything .585 -.654 -.023 If I don't know what is in a food, I won't try it -.080 730 .337 I am afraid to eat things I have never eaten before -.105 .490 645 The variables I usually aim to eat natu ral food and For me wholesome nutrition begins with the purchase of f oods of high quality had more or less the same levels of correlation with each factor, hence these va riables contributed significantly to both factors variances. The interpretation of th is could be that the nutrition and natural product attributes are perceived by consumers as indicators of quality in general and specific quality associated with natural foods. Perhaps no real distinction exists for these variables in relation to these quality factors.

PAGE 78

66 Table 7-3. Rotated component matrix(a) for food quality proxy variables-tomato data Factor Proxy Variable 1 (General Quality conscious) 2 (Natural Quality conscious) I usually aim to eat natural food 439 488 I am willing to pay somewhat more for a product of better quality 865 .182 Quality is decisive for me in purchasing foods 905 .254 I always aim at the best quality 710 .373 When choosing foods, I try to buy products that do not contain residuals of herbicides and antibiotics .148 930 I am willing to pay somewhat more for food containing natural ingredients .359 .818 For me, wholesome nutrition begins with the purchase of foods of high quality 599 546 Table 7-4 shows further result s of the factor analyses with the tomato data. Once again, three factor scores we re derived for consumer food preferences. The third factor was however unnamed owing to unclear interp retation of the fact or loadings. These results show that sometimes naming of the late nt factors in a factor analysis, even after rotation, can be complicated and at times inappropriate (Hair et al., 2003) This is one reason why there is continued debate in lite rature on the concept and relevance of naming underlying factors. Perhaps the factors underl ying influence on the dependent variable in subsequent double hurdle analyses is all that matters and deriva tion of names is best left alone. In terms of extraction rate, the three factor s were attributable to 68 percent of the variance in the original vari ables and a 0.847 KMO measure of sampling adequacy. It is noteworthy to mention that this and other levels of extract ion described in this study,

PAGE 79

67 though relatively low, are statistically acceptable especially given the high KMO measures of sampling adequacy. Table 7-4. Rotated component matrix(a) for food preference proxy variables-tomato data Factor Proxy Variable 1 (Open to unfamiliar foods) 2 (Choosey) 3 (unnamed) I like foods from different countries .718 -.160 .069 Ethnic food looks too weird to eat -.650 .549 .107 I like to try new ethnic restaurants .809 -.215 .221 At parties, I will try a new food .725 -.100 .488 I am very particular about the foods I will eat -.063 .614 -.493 I am constantly sampling new and different foods .540 -.206 .560 I don't trust new foods -.355 .726 -.090 I will eat almost anything .154 -.263 .853 If I don't know what is in a food, I won't try it -.049 .703 -.273 I am afraid to eat things I have never eaten before -.312 .743 -.171 For the combined apple and tomato data se t, the factor analyses yielded similar results, as expected. 74 percent extraction ra te was achieved for the two factors extracted, with a KMO measure of 0.851 being attained. Table 7-5 summarizes the factor analysis performed for the food quality variables. Once more, the variables I am willing to pay somewhat more for a product of better quality, Quality is decisive for me in purchasing foods and I always aim at the best quality loaded high on the General Quality conscious factor. The other variables loaded high on the Natural Quality conscious factor. Again, the variable For me,

PAGE 80

68 wholesome nutrition begins with the purchase of foods of high quality loaded evenly between the two factors. Table 7-5. Rotated component matrix(a) for food quality proxy vari ables-combined apple and tomato data set Factor Proxy Variable 1(General Quality conscious) 2 (Natural Quality conscious) I usually aim to eat natural food .290 .692 I am willing to pay somewhat more for a product of better quality .812 .266 Quality is decisive for me in purchasing foods .910 .229 I always aim at the best quality .796 .347 When choosing foods, I try to buy products that do not contain residuals of herbicides and antibiotics .204 .861 I am willing to pay somewhat more for food containing natural ingredients .367 .824 For me, wholesome nutrition begins with the purchase of foods of high quality .649 .504 Pertaining to consumer food preferences, th e results presented in Table 7-6 were attained for the combined apple and tomato data set. This time the extraction rate was slightly lower, at 68 percent. The KMO measure of sampling adequacy was 0.863 and a total of three factors were extracted using the scree plot decision rule. Most of the loadings were similar to those in the tomato data set, with the majority clearly loading on a single factor. Only I will ea t almost anything and If I dont know what is in a food, I wont try it did not, possibly indicating the vagueness of these statements.

PAGE 81

69 Table 7-6. Rotated component matrix(a) for food preference proxy variables-tomato data Factor Proxy Variable 1(Open to different foods) 2(Choosey) 3(afraid of unfamiliar foods) I like foods from different countries .689 .003 -.261 Ethnic food looks too weird to eat -.398 .006 .752 I like to try new ethnic restaurants .808 -.161 -.292 At parties, I will try a new food .832 -.286 -.115 I am very particular about the foods I will eat -.140 .723 .272 I am constantly sampling new and different foods .669 -.380 -.229 I don't trust new foods -.263 .305 .703 I will eat almost anything .462 -.766 .040 If I don't know what is in a food, I won't try it -.036 .633 .446 I am afraid to eat things I have never eaten before -.167 .440 .671 On the whole, the factor analyses pres ented here performed well with acceptable levels of significance and extraction rates. Th e only case that warranted discarding of the procedure was that of food safety variable s which had low correlations. Perhaps more questions on food safety should have been asked and framing them in a better way may have improved the survey de sign to yield better results.

PAGE 82

70 CHAPTER 8 ECONOMETRIC MODELING WI TH FACTOR SCORES This chapter presents the models with fact or scores derived from the factor analyses alluded to in Chapter 7. As in the case of Chapter 6, three models are developed (i.e. for apple data, tomato data, and combined appl e and tomato data) and their results are discussed in detail. Apples Model (Model 4) When factor scores for food quality and food preferences were incorporated in place of single variables in the apples pr obit model, the explanatory power increased marginally (significance level of 0.94). However, this model was still insignificant as shown in Table 8-1. Despite this result, the truncated tobit es timation is presented below in Table 8-2 and as is shown, age, gender, income and lo cation demographics were significant at 0.1 significance level. Surprisingl y, the quality variables (QUAL1 and QUAL2) were both insignificant in the truncated model. Neverthele ss, the signs of these quality coefficients were positive as expected, implying that qua lity conscious consumers would pay more for the labeled apples. The signs of the estima ted coefficients for demographics were also as expected save for age, which was nega tive, suggesting that older consumers were unwilling to pay more for labeled apples. With respect to income, consumers with an annual pre-tax income of less than $50,000 were willing to pay more for COOL. The other income group dummy variables were insi gnificant suggesting that they have the same effect as the base ($100,000 and more).

PAGE 83

71 Table 8-1. Apples probit m odel with factor scores (Model 4) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0031 0.0146 0.00 -0.215 0.8301 45.22 GENDER 0.2107 0.4408 0.06 0.478 0.6326 0.90 EDU1 -0.0799 0.3850 -0.02 -0.208 0.8356 0.35 EDU2 0.2305 0.3508 0.06 0.657 0.5112 0.38 LOC1 -0.1216 0.3710 -0.03 -0.328 0.7432 0.41 LOC2 0.5014 0.4963 0.11 1.010 0.3124 0.13 INC1 0.1824 0.3905 0.05 0.467 0.6403 0.26 INC2 -0.0315 0.3903 -0.01 -0.081 0.9356 0.19 INC3 0.2275 0.3890 0.06 0.585 0.5586 0.22 EXPOSE 0.0010 0.2178 0.00 0.004 0.9964 1.90 PC1 0.0820 0.4663 0.02 0.176 0.8604 0.26 PC2 0.6342 0.4935 0.14 1.285 0.1987 0.15 PC3 -0.0463 0.3560 -0.01 -0.130 0.8965 0.37 TRUST 0.1461 0.0962 0.04 1.520 0.1286 3.69 SAFE 0.0091 0.0911 0.00 0.100 0.9206 3.70 PFR1 -0.1714 0.1640 -0.05 -1.045 0.2960 0.00 PFR2 -0.0112 0.1525 0.00 -0.074 0.9412 0.00 PFR3 0.1364 0.1456 0.04 0.937 0.3487 0.00 QUAL1 0.1339 0.1511 0.04 0.886 0.3754 0.00 QUAL2 0.3361 0.1592 0.09 2.111 0.0348 0.00 Restricted log likelihood value, ln L10 = -69.15 R2 (McFadden, 1973) = .0760 Maximum unrestricted log likelihood value, ln L1 = -63.90 R2 (Estrella, 1998)= .0772 Log likelihood 2 (df=19)= 10.50 (p = .9394) % of correct predictions = 80.1 Number of observations = 136 Table 8-2. Apples truncated tobit model with factor scores (Model 4) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0249 0.0128 -0.01 -1.955 0.0506 44.64 GENDER 1.0991 0.5256 0.36 2.091 0.0365 0.91 EDU1 0.4658 0.3691 0.15 1.262 0.2070 0.35 EDU2 0.2450 0.3275 0.08 0.748 0.4543 0.38 LOC1 -0.2101 0.3222 -0.07 -0.652 0.5143 0.40 LOC2 -1.9737 0.6091 -0.64 -3.240 0.0012 0.14 INC1 0.8316 0.3345 0.27 2.486 0.0129 0.27 INC2 0.0412 0.3532 0.01 0.117 0.9072 0.19 INC3 -0.1450 0.3608 -0.05 -0.402 0.6877 0.23 EXPOSE -0.2030 0.1976 -0.07 -1.027 0.3045 1.90 PC1 -0.1521 0.3628 -0.05 -0.419 0.6750 0.26 PC2 -0.1248 0.4131 -0.04 -0.302 0.7625 0.17 PC3 0.3012 0.3179 0.10 0.947 0.3435 0.36 TRUST 0.0565 0.0783 0.02 0.722 0.4704 3.74 SAFE -0.0040 0.0941 0.00 -0.042 0.9662 3.72 PFR1 -0.0071 0.1314 0.00 -0.054 0.9570 0.00 PFR2 0.1620 0.1269 0.05 1.277 0.2015 0.02 PFR3 0.0728 0.1236 0.02 0.589 0.5558 0.05 QUAL1 0.1897 0.1600 0.06 1.185 0.2359 0.04 QUAL2 0.1987 0.1379 0.06 1.441 0.1495 0.08 Sigma 0.7057 0.0904 7.804 0.0000 Number of observations = 136 Log likelihood function = -32.87 Observation after truncation = 108 Threshold values for model: Lower = 0 Upper = +

PAGE 84

72 In terms of location, it was once again reve aled that consumers in Lansing MI are not willing to pay that much for produce la beled U.S.A. Grown unlike consumers in Gainesville FL or Atlanta GA. Tomatoes Model (Model 5) When it came to the tomato model estimation, a slightly different picture was found. Table 8-3 shows that most of the esti mated coefficients for the tomatoes probit model with factor scores had the expected signs even though only quality and location were significant at 0.1 significance level. Table 8-3. Tomatoes probit model with factor scores (Model 5) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0052 0.0131 0.00 -0.396 0.6918 44.21 GENDER -0.0592 0.3266 -0.02 -0.181 0.8561 0.87 EDU1 0.0248 0.3043 0.01 0.082 0.935 0.33 EDU2 0.3589 0.2900 0.13 1.238 0.2159 0.43 LOC1 0.6895 0.3008 0.26 2.292 0.0219 0.38 LOC2 -0.2145 0.2882 -0.08 -0.744 0.4566 0.28 INC1 -0.0578 0.3141 -0.02 -0.184 0.8539 0.27 INC2 -0.2076 0.3093 -0.07 -0.671 0.5022 0.28 INC3 -0.0952 0.3703 -0.03 -0.257 0.7972 0.14 EXPOSE 0.1879 0.1462 0.07 1.285 0.1989 1.94 PC1 -0.2490 0.4260 -0.09 -0.584 0.5589 0.23 PC2 -0.1610 0.3894 -0.06 -0.414 0.6792 0.22 PC3 -0.2101 0.3176 -0.08 -0.662 0.5083 0.37 SAFE 0.0742 0.0765 0.03 0.97 0.3321 3.41 TRUST 0.0589 0.0748 0.02 0.787 0.4314 3.71 PFR1 -0.1192 0.1240 -0.04 -0.961 0.3363 0.00 PFR2 0.0271 0.1175 0.01 0.231 0.8174 0.00 PFR3 0.0819 0.1132 0.03 0.723 0.4695 0.00 QUAL1 0.3140 0.1289 0.12 2.436 0.0149 0.00 QUAL2 0.1378 0.1217 0.05 1.132 0.2576 0.00 Restricted log likelihood value, ln L10 = -104.70 R2 (McFadden, 1973) = .14248 Maximum unrestricted log likelihood value, ln L1 = -89.77989 R2 (Estrella, 1998)= .16800 Log likelihood 2 (df=19)= 29.83458 (p = .0539) % of correct predictions = 74.3 Number of observations = 175 The model predicted correctly for 74.3 percent of the ac tual responses recorded. Marginal effects indicated that being a Gainesville, FL (LOC 1) consumer would increase

PAGE 85

73 the probability of being willing to pay a prem ium for tomatoes labeled U.S.A. Grown by approximately 26 percent as compared to a consumer from Atlanta, GA, ceteris paribus In contrast, being a Lansing, MI (LOC2) consumer did not change the probability of being willing to pay for tomato es labeled U.S.A. Grown when compared to the base (consumers in Atlanta, GA). Also according to model 5, consumers pe rceptions about food quality in general were shown to have a positive impact on th e likelihood of being w illing to pay a premium for COOL in fresh tomatoes. As displayed in Table 8-3, the marginal probability thereof, ceteris paribus is 12 percent. The truncated estimation in model 5 is shown in Table 8-4. Table 8-4. Tomatoes truncated t obit model with factor scores (Model 5) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0011 0.0096 0.00 -0.118 0.9057 44.26 GENDER 0.1995 0.2600 0.09 0.767 0.4429 0.86 EDU1 -0.3170 0.2201 -0.15 -1.440 0.1498 0.32 EDU2 -0.2297 0.2078 -0.11 -1.105 0.2690 0.46 LOC1 0.4261 0.2074 0.20 2.054 0.0400 0.45 LOC2 -0.6325 0.3224 -0.30 -1.962 0.0498 0.22 INC1 0.2769 0.2399 0.13 1.154 0.2484 0.27 INC2 0.3925 0.2384 0.19 1.647 0.0996 0.27 INC3 0.3744 0.2767 0.18 1.353 0.1761 0.14 EXPOSE -0.1740 0.1170 -0.08 -1.487 0.1371 1.94 PC1 -0.2476 0.2976 -0.12 -0.832 0.4054 0.22 PC2 -0.0260 0.2644 -0.01 -0.098 0.9218 0.22 PC3 -0.1813 0.2360 -0.09 -0.768 0.4424 0.36 TRUST 0.0374 0.0562 0.02 0.664 0.5066 3.46 SAFE 0.0987 0.0591 0.05 1.670 0.0950 3.86 PFR1 0.0535 0.1022 0.03 0.524 0.6004 0.02 PFR2 -0.1542 0.0878 -0.07 -1.756 0.0791 -0.01 PFR3 -0.0004 0.0895 0.00 -0.004 0.9968 0.00 QUAL1 0.1021 0.1145 0.05 0.892 0.3723 0.12 QUAL2 0.0916 0.0935 0.04 0.979 0.3274 0.09 Sigma 0.6093 0.0717 8.497 0.0000 Number of observations = 175 Log likelihood function = -47.01 Observation after truncation = 125 Threshold values for model: Lower = 0 Upper = + It was once again noted that location is a significant factor influencing how much consumers will pay for COOL in tomatoes. Co nsumers in Gainesville, FL will pay on

PAGE 86

74 average $0.20 more than Atlanta cons umers while Lansing consumers will pay approximately $0.30 less on average. Also s hown is that household income level affects the premium that consumers will pay for fresh tomatoes labeled U.S.A. Grown. Consumers in the $50,000 to $74,999 (INC2) in come bracket will pay on average $0.19 more than consumers in the $100,000 or more income bracket (the base). Generally, consumers with incomes of less than $100,000 ar e shown to be willi ng to pay a larger premium for COOL in tomatoes. An additional finding of model 5 is that c onsumers views about food safety have a bearing on how much they will pay as a prem ium for U.S.A. Grown labeling in fresh tomatoes. This was strangely not so in th e apple model (i.e. model 4). Consumers who said they think about food safety when purch asing fruits and vegetables are found to be willing to pay more for tomatoes labeled U .S.A. Grown. Perhaps food safety concerns of the surveyed consumers are somewhat greate r in fresh tomatoes than in fresh apples. Model 5 also estimates that consumers food preferences have some bearing on how much consumers will pay for COOL in fresh tomatoes. However, only the second factor score for food preference s, choosey, is significant at = 0.1 and it negatively impacts how much consumers will pay as a premium for tomatoes labeled U.S.A. Grown. The negative impact is quite unexpected because one would think that consumers who are particular about the food they eat will pay more to know the origin of their tomatoes. In this case, the model pred icts the opposite; particular consumers are unwilling to pay more for tomatoes labeled U .S.A. Grown. It is possible that these consumers trust the marketing system to ensure that tomatoes made available are of acceptable quality and standards, thus they ar e not too concerned about country of origin.

PAGE 87

75 This may also be the same for high income level consumers who also proved to be less concerned about country of origin. Combined Apples and Tomatoes Model (Model 6) Table 8-5 shows the probit estimation for the combined apple and tomato data. It shows that both food quality factor scores were significant at = 0.05. Table 8-5. Combined apples and tomato es probit model with factor scores (Model 6) Varible Coefficient Standard Error Marginal Effects Standardized Coefficient p-value Mean of Regressor AGE -0.0077 0.0088 0.00 -0.874 0.382 44.65 GENDER -0.0655 0.2415 -0.02 -0.271 0.7863 0.88 EDU1 0.1513 0.2219 0.05 0.682 0.4953 0.34 EDU2 0.2548 0.2105 0.08 1.211 0.226 0.41 LOC1 0.2909 0.2140 0.09 1.359 0.1741 0.40 LOC2 -0.2283 0.2195 -0.08 -1.04 0.2982 0.21 INC1 -0.0207 0.2280 -0.01 -0.091 0.9276 0.27 INC2 -0.1404 0.2328 -0.05 -0.603 0.5464 0.24 INC3 0.0822 0.2566 0.03 0.32 0.7489 0.17 EXPOSE 0.1826 0.1130 0.06 1.615 0.1062 1.91 PC1 -0.1269 0.2760 -0.04 -0.46 0.6457 0.24 PC2 0.0933 0.2733 0.03 0.341 0.7328 0.19 PC3 -0.0860 0.2209 -0.03 -0.389 0.6971 0.37 SAFE 0.0474 0.0538 0.02 0.881 0.3781 3.71 TRUST 0.1133 0.0545 0.04 2.08 0.0375 3.53 PFR1 -0.1170 0.0890 -0.04 -1.314 0.1889 0.00 PFR2 -0.0192 0.0848 -0.01 -0.226 0.8212 0.00 PFR3 0.0502 0.0849 0.02 0.592 0.554 0.00 QUAL1 0.1762 0.0859 0.06 2.052 0.0402 0.00 QUAL2 0.2264 0.0902 0.07 2.509 0.0121 0.00 Restricted log likelihood value, ln L10 = -179.40 R2 (McFadden, 1973) = .07925 Maximum unrestricted log likelihood value, ln L1 = -165.1843 R2 (Estrella, 1998)= .09086 Log likelihood 2 (df=19)= 28.43597 (p = .0754) % of correct predictions = 74.0 Number of observations = 311 Consumers who were more conscious about f ood quality (be it quality in general or quality associated with natura l foods) were found to be more likely to pay a premium for fresh apples or tomatoes labeled U.S.A. Grow n. The level of trust that consumers have for information they receive from U.S. government agencies (e.g. USDA, FDA, EPA etc.) was the only other significant variable (at = 0.05). Here it was found that consumers who were more trusting of the info rmation they receive from U.S. government

PAGE 88

76 agencies were more likely to pay for COOL. Su rprisingly, all demogra phics turned out to be insignificant in th e participation decision making pro cess, suggesting that it does not matter if one is male or female or if income is high or low. Perhaps the participation decision is simply not a function of demogra phics. Overall, the model was significant (p = 0.075) and had a 74.0 percen t correct prediction rate. The truncated tobit estimation is presented in Table 8-6, where it is reported that gender and location significantly determine how much the consumers are willing to pay once they have decided they are willing to pay for COOL. Marginal effects show the expected result that female consumers are willing to pay 20 cents more per pound than males. In terms of location, the base used was Atlanta, GA and consumers in Lansing, MI were found to be willing to pay substantially less than those in Atla nta (approximately 46 cents per pound less). In contrast, consumer s in Gainesville, FL were willing to pay approximately 4 cents per pound more than thos e in Atlanta, GA. A reason for this could be that MCOOL policy has been prevalent fo r the past 26 years at the state level in Florida. Thus, shoppers in Gainesville FL could be accustomed to MCOOL and therefore are willing to pay for COOL. Conversely, Michigan is geographically far from either Georgia or Florida. Thus it could be the case that Michigan consumer s are less exposed to COOL and are therefore less willing to pay for COOL. Another possible explanation is that Michigan borders Canada, and perhap s Michigan consumers are accustomed to buying and consuming food from Canada. For more details on location-based premium differentials refer to Table 5-3 in Chapter 5. Income level is another demographic that seemed to have an impact on the amount consumers are willing to pay. The greate r-than-$100,000 income group was used as the

PAGE 89

77 base in the model. Findings suggest that co nsumers with an income level of less than $50,000 are the only group with a significantly greater WTP than the base (15 cents per pound more). This finding is similar to that in Loureiro and Umberger (2002) implying that more affluent consumers consider it unimportant to know where their apples or tomatoes come from and thus do not value COOL. Table 8-6. Combined apples and tomatoes truncated tobit model with factor scores (Model 6) Variable Estimated Coefficient Standard Error Marginal Effects Standardized Coefficent p-value Mean of Regressor AGE -0.0156 0.0106 -0.01 -1.478 0.1394 44.46 GENDER 0.5704 0.3406 0.20 1.675 0.0940 0.89 EDU1 -0.0216 0.2406 -0.01 -0.09 0.9285 0.34 EDU2 -0.0498 0.2276 -0.02 -0.219 0.8266 0.42 LOC1 0.1206 0.2082 0.04 0.579 0.5625 0.43 LOC2 -1.3283 0.4310 -0.46 -3.082 0.0021 0.17 INC1 0.4493 0.2492 0.15 1.803 0.0714 0.27 INC2 0.2555 0.2476 0.09 1.032 0.3023 0.23 INC3 0.0596 0.2817 0.02 0.212 0.8324 0.18 EXPOSE -0.2253 0.1327 -0.08 -1.697 0.0897 1.93 PC1 -0.1400 0.2921 -0.05 -0.479 0.6319 0.24 PC2 -0.0887 0.2785 -0.03 -0.319 0.7500 0.20 PC3 0.0583 0.2351 0.02 0.248 0.8041 0.36 SAFE 0.1204 0.0667 0.04 1.804 0.0712 3.80 TRUST 0.0306 0.0570 0.01 0.537 0.5911 3.62 PFR1 -0.0965 0.1051 -0.03 -0.918 0.3587 0.00 PFR2 0.0556 0.0862 0.02 0.645 0.5191 0.01 PFR3 -0.0695 0.0938 -0.02 -0.741 0.4588 0.01 QUAL1 0.1201 0.1188 0.04 1.011 0.3120 0.08 QUAL2 0.1701 0.1042 0.06 1.633 0.1026 0.09 Sigma 0.7937 0.0892 8.901 0.0000 Number of observations = 311 Log likelihood function = -105.90 Observation after truncation = 233 Threshold values for model: Lower = 0 Upper = + The truncated estimation also displayed the food quality factor score that is associated with natural foods as havi ng a positive impact on the price premium consumers are willing to pay for the U.S.A. Grown labeled fresh apples and tomatoes. In other words, consumers who are more na tural-quality-conscious would pay more for COOL. However, the factor score for quality in general, had no significant impact on the amount they were willing to pay.

PAGE 90

78 The food safety variable is the other f actor that was signifi cant at a significance level of 0.1. Consumers th at think about food safety when purchasing fruits and vegetables were found to be willing to pay more for the label U.S.A. Grown. Unexpectedly, this food safety variable had not been significant in th e participation stage, suggesting that it only plays a role on the consumption decision. The consumers selfrated exposure to food safety information in fr esh fruits and vegetabl e is another variable that was significant in model 6 and only with respect to the consumption decision. Consumers that rated themselves as having s een, read or heard less about food safety in fresh fruits and vegetables were willing to pay less for the label U.S.A. Grown. This suggests that consumers may be likely to pa y more for produce labeled U.S.A. Grown if they are more informed about food safety issues in fresh fruits and vegetables.

PAGE 91

79 CHAPTER 9 CONCLUSIONS AND IMPLICATIONS Summary This thesis researches consumers WTP for COOL in fresh apples and tomatoes, specifically the label attribute U.S.A. Grow n. Previous research on WTP for COOL has focused primarily on the beef sector with very little work being done on COOL in the produce sector. Thus, this thesis looks at the situation in the produce sector in order to provide empirical information th at can be used by policy makers and decision makers in the produce sector. Given rising import compe tition in the produce sector and imminent changes in COOL legislation, which w ill see current voluntary COOL becoming mandatory, this information is timely and can contribute to a more informed decision making process. The study determines the nature of consum ers WTP for COOL in fresh apples and tomatoes, ascertaining how much consumers are willing to pay and whether the WTP is product specific or not. Factor s that influence the WTP are researched by applying Craggs double hurdle model to data collected from primary shoppers in Gainesville, FL, Lansing, MI and Atlanta, GA. In total, si x separate double hurdle models are estimated using the same model specification. Three of these estimations incorporate factor scores of food quality and food prefer ence variables as regressors, while the other three do not. In each set of three models, one analyzes fresh apple data, another fresh tomato data and a third combined fresh apple and tomato data.

PAGE 92

80 Findings show that consumers from Michig an, Florida and Georgia are willing to pay a premium for the COOL attribute, U.S.A Grown, in fresh apples and tomatoes. 79 percent of the consumers sampled are found to be willing to pay a premium for fresh apples labeled U.S.A. Grown while 72 percen t are willing to pay a premium in the case of fresh tomatoes labeled U.S.A. Grown. Th e mean WTP for COOL is calculated to be approximately $0.49 and $0.48 per pound of U.S. labeled apples and tomatoes respectively. These findings lead to the failure to reject the first hypothesis of the study, thereby accomplishing the first specific objective outlined in Chapter 1. The thesis also ascertains that the premiu m consumers are willing to pay for fresh apples labeled U.S.A. Grown is statistically equivalent to the pr emium they are willing to pay for fresh tomatoes labeled U.S.A. Grown. This results in the rejection of the second hypothesis of the study and achieves the se cond specific objective. Concerning the factors that influence th e WTP for the COOL attribute, U.S.A. Grown in fresh apples and tomatoes, diffe rent models are developed, which come up with slightly different results, possibly indicating that though th e premiums may be equivalent, the factors affecting them may differ slightly. Nonetheless, most of the estimated coefficients are found to have the expected signs on them and similar findings are made in all models. It is established that cons umer food quality perception is a strong predictor of WTP for COOL in fresh apples and tomatoes. Location is also a significant explanatory variable, with consumers from Lansing, MI being the least willing to pay for the label U.S.A. Grown in both fresh apples and fresh tomatoes. Consumers from Gainesville, FL turn out to be willing to pay the highe st average price premium for COOL in fresh

PAGE 93

81 tomatoes while consumers from Atlanta, GA are willing to pay the highest average price premium for COOL in fresh apples. Consumer food preferences (as defined according to the questions in th e questionnaire in Appendix B ) are found to be insignificant in most models. Whether consumers are choosey or not about the food they eat does not seem to affect their WTP for the label U.S.A. Grown in fresh apples and tomatoes. Similarly, the gender, level of educati on and the presence of children in a household seem not to have much of an effect on consumers WTP for COOL in fresh apples and tomatoes. Also, consumers level of exposure to food sa fety information in fruits and vegetables seem not to have much of an effect either Most psychographics a nd demographics such as gender, age, education, and food safety c oncerns seem to have minor effect depending on the product under consideration. This is particularly so, in the truncated models where the consumption decision (i.e. how much to pa y) is estimated. These findings relate to the third hypothesis of this study and lead to the rejection of some aspects of the hypothesis. Consumer perceptions about food preferences and food safety are not key factors of WTP for COOL in fresh apples and tomatoes, as de monstrated by the lack of significance at 0.1 levels. However, consumer perceptions about food quality are found to be statistically significant, suggesting that th ey are key factors of WTP fo r COOL in fresh apples and tomatoes. Implications There are several implications that arise from the findings made in this thesis. First, by establishing that consumers are willing to pay for the label U .S.A. Grown in both fresh apples and fresh tomatoes, the study imp lies that U.S. consumers want to know the country of origin of the apples and toma toes they consume. This contributes to justification (on the consum ers side) of MCOOL or at the very least COOL on a

PAGE 94

82 voluntary basis, in the apple and tomato mark ets; either mandatory or voluntary COOL legislation would facilitate th e provision of country-of-origin information as desired by consumers at the final point of purchase. Also, the findings suggest that it may be possible for producers and marketers to use the label U.S.A. Grown in order to garner a competitive advantage over import substitutes in the U.S. market. Consumers su rveyed in this study were willing to pay approximately $0.49 and $0.48 more for a pound of U.S. labeled apples and tomatoes respectively, over unlabeled pr oduce. However, the costs asso ciated with incorporating labels have to be calculated before an additional price mark up can be considered. Furthermore, it would be necessary to comp are the net price premium (i.e. the price premium after deducting costs of incorporat ing labeling) for U.S.A. Grown produce with those of other countries. The thesis also establishes premium equivalency in fresh apples and tomatoes labeled U.S.A. Grown, possibl y implying that there is pote ntial for generic promotion of the label U.S.A. Grown to enhance overall demand for U.S. produce over imports. This is, however, not conclusive from the fi ndings of this study, and complementary and extensive research is required to make more than tentative claims about this. By making use of Craggs double hurdle model to estimate the WTP for COOL, this thesis finds that consumers food quality perceptions are critic al factors in both the participation and consumption decision maki ng processes. This implies that before COOL can be used as a tool for enhancing demand for U.S. produce; produce quality will have to be ensured, otherwise, COOL is likel y to induce a negative ma rket response. It is apparent from the research that the label U .S.A. Grown may serve to inform consumers

PAGE 95

83 not only the country-of-origin but also the quality of the produce. This is also an important implication with respect to prospe cts of generic promotion of U.S. labeled produce. If any form of generic promotion is to be chosen by industry players, then ensuring consistent quality will be most criti cal. Poor quality in a ny part of the industry may tarnish the image of the label, consequently causing a decline in the price consumers are willing to pay for produ ce labeled U.S.A. Grown. Location is another factor th at is clearly shown to influence how much consumers will pay for produce labeled U.S.A. Grown. An implication from this is that if a marketing strategy is to be developed for fresh tomatoes and apples labeled U.S.A. Grown, it will have to be sens itive to the location of the ta rgeted consumers. If pricing and promotion strategies are to be successful, they may need to be different in accordance with the nuance of dominant consumer preferen ces at each location. Given that in Florida (where MCOOL has been in place for 26 years) consumers were willing to pay more for COOL, it is implied that U.S. consumers can become loyal to the label U.S.A. Grown if exposed to it for a prolonged time. Perhaps, Floridian consumers may have learned to expect labeling over the years. Other factors that the thesis finds to be of relative importance include the extent to which consumers trust the information they re ceive from U.S. government agencies (e.g. USDA, FDA and EPA) and the consumers fo od safety concerns. While most models find these to be insignificant factors, it is cl ear that they have some bearing on how much consumers will pay for COOL. Concerning th e former, it is found to be a significant determinant of the participation decision in one of the models with factor scores (Model 6). This is likely due to the fact that agen cies are responsible for regulating and enforcing

PAGE 96

84 produce labeling laws. If consumers trust the information they are getting from these agencies, then they are more likely to value COOL because they would believe that there is a trustworthy labeling verifi cation system in place. Since COOL is a credence-attribute label that consumers cannot verify except thr ough a third party (i.e an agency), this would be important. As pertaining to food safety concerns, th ese seem to mainly affect the WTP in tomatoes and not in apples. The truncated estimations of the tomatoes models (Models 2 and 5) show that consumers who take food safety concerns into consideration when making the decision to purchase fruits and ve getables will pay more money for the label U.S.A. Grown in fresh tomatoes. This is somewhat expected, since fresh tomatoes are relatively more perishable and prone to food safety problems. Another possible implication here is that consumers regard U. S. tomatoes to be safer implying that U.S. fresh tomato producers may be able to capital ize on this to garner competitive advantage over import substitutes. With respect to demographics, most ar e found to be insign ificant predictor variables, with gender and income level variable s turning out to be mildly important. It is critical to note that the data analyzed in this study was drawn from primarily high income earners and highly educated in dividuals. Thus, for purposes of marketing the U.S.A. Grown label, information in this thesis could be used to target this particular demographic in the U.S. primary shopper population. All in all, this thesis re ports the findings that consum ers food quality perceptions and consumers location are im portant determinants of WT P for COOL in apples and tomatoes. Most demographics are also somewh at important factors, especially in the

PAGE 97

85 consumption decision on how much to pay, (i.e. once the consumer is willing to pay a premium). Though food safety, trust and prefer ences are psychographics included in the models estimated, these are relatively less important. Areas for Further Research The scope of this thesis, like all others, is and had to be limited. Several aspects relating to the research topic ar e simply not covered, leaving some areas for further study. Though this research finds consumers to be wi lling to pay for fresh apples and tomatoes labeled U.S.A. Grown ther e is little knowledge on how the U.S. label would fare against other country-of-origin labels. Res earching this topic is imperative before substantive conclusions can be drawn about pr omoting the generic U.S.A. Grown label. The ramifications of such a promotional program on import competition in the U.S. produce industry can only be estimated if co mpetitor country labels price premiums are calculated and compared to that for the U.S.A. Grown label. Using the same data that were collected a nd analyzed for this thesis, a forthcoming study addresses some of these issues. Prelim inary findings from the study, suggest that the U.S.A. Grown label fares relatively we ll against several competitor country labels. Table 9-1 presents a synopsis of this and s hows the price premiums that consumers were willing to pay to give up one pound of foreign fr esh apples or fresh tomatoes, (which they had been endowed with) in exchange for id entical produce labeled U.S.A. Grown. As described in Chapter 3, each participating co nsumer had initially been endowed with one pound of unlabeled fresh apples (tomatoes) and was then asked to bid for identical produce labeled U.S.A. Grown. What is not mentioned in Chapter 3 is that after the fourth round of bidding had occurred, consumers were informed the origin of their fresh apples (tomatoes), which they were initially endowed with.

PAGE 98

86 Table 9-1. Comparison of mean bids: U.S. A. Grown versus Other Country labels WTP for U.S. Label Mean Difference Paired Samples Test t-value Number of Observations Before Information After Information Apples U.S.A. Grown versus No label 0.49 136 U.S.A. Grown versus Chile 0.42 0.41 -0.01 -0.240 59 U.S.A. Grown versus China 0.37 0.46 0.09 2.658 39 U.S.A. Grown versus New Zealand 0.71 0.88 0.17 2.043 38 Tomatoes U.S.A. Grown versus No label 0.48 175 U.S.A. Grown versus Canada 0.34 0.38 0.04 1.475 67 U.S.A. Grown versus Mexico 0.58 0.93 0.35 4.432 86 U.S.A. Grown versus the Netherlands 0.56 0.67 0.11 1.941 22 Thus, in Table 9-1, the first two columns titled before information and after information, show the mean WTP for the fr esh produce labeled U.S.A. Grown before the foreign country information was revealed and after the foreign country information was revealed to the consumers, respectively. Since each survey site had been told a different foreign country of origin, the numbe rs of observations va ry though the totals sum up to the sample sizes of 136 for apples and 175 for tomatoes. Further comparative analysis of the WTP by location, and with larg er sample sizes, would be useful in future studies if this is to better inform U. S. producers and marketers of fresh produce. It would also be interesting to find out if U.S. producers would be better off using generic promotion of the label U.S.A. Grown in an effort to enhance market demand for U.S. grown produce. Currently, it is uncle ar if a firm level, sub-sector level or national level approach to promote the U .S.A. Grown label would be appropriate. Establishing this would be of keen interest to several players in the produce market who

PAGE 99

87 may want to enhance demand for U.S. grown fresh tomatoes and apples, given the rising import competition recorded in the last few years. Concerning producers and market ers profitability, it would be useful to find out if costs associated with implementing COOL can be offset by what consumers are willing to pay for it. With the coming of MCOOL, this would be an important question to answer since MCOOL would certainly introduce new co sts. Though this thesis makes findings that add to the justification for MCOOL or at the very least voluntary COOL, it does so by establishing whether or not consumers ar e willing to pay for the label U.S.A. Grown. The cost implications thereof are not addressed.

PAGE 100

88 APPENDIX A EXPERIMENTAL AUCTIONS INSTRUCTIONS Procedures for Experimental WTP Auctions Instructions Thank you for agreeing to participate in toda ys session. As you entered the room, you should have been given $10.00 and a packet. You should also have been assigned an ID number, which is located on the upper right hand corner of the packet. You will use this ID number to identify yourself during this re search session. We use random numbers in order to ensure confidentiality. Before we begin, I want to emphasize that your participation in this session is completely voluntary. If you do not wish to participate in the experiment, please say so at any time. Non-participants will not be penalized in any way. I want to assure you that the information you provide will be kept strictly confidential and used only for the purposes of this research. Please complete the consent form at this time. In todays session, we are ultimately interest ed in your preferences for several different types of foods. First, we will conduct a food evaluation exercise, followed by a consumer survey. Before we begin, I w ould like you all to open your packets and take a minute to fill out the consent form. I will now begin going through a set of instru ctions with you and will read from this script so that I am able to clearly convey the procedures. Importantly, from this point forward, I ask that there be no talking among participants. Are there any questions before we begin? In todays session, we are ultimately inte rested in your preferences for fruits and vegetables. To determine how much these foods may be worth to you, we are going to conduct an auction. To begin, however, we will conduct a candy bar auction so that you can learn how the auction procedures work. First, let me say that I realize the following instructions might be a bit confusing at first. Dont get frustrated. We are conducting the candy ba r auction first so that you will have a chance to learn how things wo rk. I will also provide an example that should help to further clarify any confusion you might have.

PAGE 101

89 Instructions Candy Bar Auction For agreeing to participate in this research session, we are giving each of you one free small Snickers bar. The small Snickers bar is yours to keep, but plea se do not eat it just yet. Although you have been given a free small Snickers bar, in a moment we will give you the opportunity to participate in an auction to obtain a diffe rent size of candy bar if you so desire. Here in the front of the ro om, I also have a large Snickers bar. I will now conduct an auction, where you w ill have the opportunity to exchange your Small Snickers bar for the Large Snickers bar. Most likely, you woul d probably prefer to have the larger Snickers bar rather than the smaller Snickers bar. Therefore, Im interested in how much m oney you would be willing to pay to exchange your smaller Snickers for the larger Snickers. In a moment, you will be asked to indicate the most amount of money you are willing to pay to exchange your Small Snickers bar fo r the Large Snickers bar by writing bids on the enclosed bid sheets. Let me explain how the auction will proceed: Auction Procedures 1) First, each of you has been given a bid sh eet in your packet. On this sheet you will, in a moment, write the most amount of money you are willing to pay to exchange your Small Snickers bar for the Large Snickers bar. This is the amount of money you would be willing to pay to exchange your Small Snickers bar and take the Large Snickers bar instead. Note : Your bids are private information and should not be shared with anyone. 2) After youve finished writing your bids the monitor will go around the room and collect the bid sheets. 3) In the front of the room, bids will be ranked fr om highest to lowest. 4) The four highest bids for the Large Snickers bar will win the auction. The individuals that submitted the four highest bids will pay the 5th highest bid amount to exchange their Small Snickers bar for the Large Snickers. 5) The numbers of the winning bidders and the winning price (the 5th highest bid) will be written on the chalkboard for everyone to see. 6) After posting the price, I will re-conduct the auction for 2 additional rounds. 7) At the completion of the 3rd round, I will randomly draw a number 1 through 3 to determine the binding round. For example, if we randomly draw the number 3, then we will ignore outcomes in all other rounds and only focus on the winning bidders and price in round 3. Importan tly, all rounds have an equally likely chance of being the winning round. 8) Once the binding round has been determined, the winning bidders will come forward and will pay the 5th highest bid amount and exchange their Small

PAGE 102

90 Snickers for the Large Snickers. All ot her participants will pay nothing and will keep their Small Snickers. Important Notes You will only have the opportunity to win an auction for one candy bar. Because we randomly draw a binding round, you cannot win more than one candy bar. You will either leave with one Large Sn ickers if you are a high bidder and some amount of money less than $10, or one Sm all Snickers and $10 if you are a low bidder. You will actually pay money to exchange the Small Snickers for the Large Snickers bar. This procedure is not hypothetical. In this auction, the best strategy is to bid exactly what it is worth to you to exchange your Small Snickers for the Large Snickers. Consider the following: if you bid less than it is worth to you to excha nge your Small Snickers bar for a Large Snickers bar, you may end up not winning the auction even though you could have exchanged your Small Snickers at a price you were actually willing to pay. Conversely, if you bid more than it is worth to you to exchange your Small Snickers for the Large Snickers, you ma y end up having to exchange your Small Snickers at a price higher than you really wa nted to. Thus, your best strategy is to bid exactly what it is worth to you to excha nge your Small Snickers bar for the Large Snickers bar. It is acceptable to bid $0.00 for the Large Snickers in any round. This would mean that you are not willing to give up your Small Snickers for the Large Snickers. Importantly, we are interested in what it is worth to you to exchange your Small Snickers for the Large Snickers. We are not interested in your total value of the Large Snickers, only the amount of the exchange. In other words, what is the premium you would place on your Large Sn ickers versus the Small Snickers?

PAGE 103

91 Auction Example Suppose there were 10 people participating in an auction just like the one you are about to participate in. Suppose that these individu als participated in 3 auction rounds, as you will, and that the 2nd round was randomly selected to be binding. Now, suppose in round 2, participant #1 bid $0.27 to exchange their Small Snickers bar for the Large Snickers, participant #2 bid $0.24, participant #3 bid $0.21, participant #4 bid $0.18, and participant #5 bid $0.15, participant #6 bid $0. 12, participant #7 bid $0.09, participant #8 bid $0.06, participant #9 bid $0.03, and participant #10 bid $0.00 to exchange their Small Snickers bar for the Large Snickers. To furt her illustrate, the bids in round 2 were as follows: Participant Number #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 Bid Amount $0.27 $0.24$0.21$0.18$0.15$0.12$0.09$0.06 $0.03 $0.00 Who would win the auction? Participants #1, #2, #3, and #4 would win the auction because they were the four highest bidders. How much would participants #1, #2, #3, and #4 pay to exchange their Small Snickers bar for the Large Snickers? They would pay them the 5th highest bid amount, which was $0.15. Thus, participants #1, #2, #3, and #4 would come to the front of the room, pay $0. 15 and exchange their Small Snickers bar for a Large Snickers. Participants #5, #6, #7, #8, #9 and #10 would pay nothing and would keep with their free Small Snickers bar. *Note: these dollar amounts were used for illustrative purposes only and should not in any way reflect what the candy bars may be worth to you* Do you have any questions before we begin? Please use the bid sheets marked candy bar auction.

PAGE 104

92 If Apples Auction was conducted first then: Apple Auction Now that you have had the chance to learn how the auction will work, we are interested in your preferences for two types of apples. For agreeing to participate in this research session, we are giving each of you a free pound of apples. These apples are yours to keep. Although you have been given these apples fo r free, we will give you the opportunity to participate in an auction to obtain different a pples if you so desire. Here in the front of the room, we have another apple. These apples have a label on them that identifies that they were grown in the United States. The a pples are the same size and weight as the apples that you have been given. We will now conduct an auction, where you will have the opportunity to exchange your apples that contain NO label for the apples with the label Grown in the United States. In a moment, you will be asked to indicate the most amount of money you are willing to pay to exchange your apples for the apples with the label by writing bids on the enclosed bid sheets. The procedures for this auction are exactly the same as the candy bar auction, with one exception: we request that all participants eat an apple at the end of the todays auctions. To refresh your memory as to how the auction works, I will go through the instructions again. Auction Procedures 1) First, each of you has been given a bid sh eet in your packet. On this sheet you will, in a moment, write the most amount of money you are willing to pay so that you would be willing to exchange your apples for the apples with the label Grown in the United States. This is the most amount of money you will pay to exchange your unlabeled apples for my labeled apples. Note: your bids are private information and should not be shared with anyone. 2) After youve finished writing your bids the monitor will go around the room and collect the bid sheets. 3) In the front of the room, each of your bids will be ranked from highest to lowest. 4) The four highest bids will win the auction. The individuals with the four highest bids will be the 5th highest bid amount for the exchange. 5) We will write the winning bidder nu mbers and the winning price on the chalkboard for everyone to see. 6) After posting the price, we will re-conduct the auction for 6 additional rounds. 7) At the completion of the 7th round, we will randomly draw a number 1 through 7 to determine the binding round. For example, if we randomly draw the number 5, then we will ignore outcomes in all other rounds and only focus on the winning bidders and price in round 5. Importan tly, all rounds have an equally likely chance of being binding.

PAGE 105

93 8) Once the binding round has been determined, the winning bidders will come forward and pay the 5th highest bid amount and exch ange their apples for the labeled apples. All other partic ipants will keep their apples. Important Notes You will only have the opportunity to win an auction for one pound of apples. Because we randomly draw a binding round, you cannot win more than one pound of apples from this auction. The winning bidder will actually pay money to exchange their apples for the labeled apples. This procedure is not hypothetical. We will also expect participants to eat an apple at the conclusion of the auctions. In this auction, the best strategy is to bid exactly what it is worth to you to exchange your apples for the labeled apples. It is acceptable to bid $0.00 in any ro und. This would mean that you are not willing to pay anything to trade your unlab eled apples for the labeled apples. Importantly, we are interested in what it is worth to you to exchange your apples for the labeled apples. Do you have any questions before we begin? Before we begin, we would like to ask you to think about the following statement: This is not a hypothetical setting; you will pa y money to exchange your apples if you win the auction. The reason we use these auctions instead of, say for example, a survey, is that it is closer to real life. However, it is not exactly a real situation; therefore, people do not always bid the same as they would trul y behave in the grocery store. Frequently, people bid a higher amount that they would pay in the grocery store. Let me tell you why I think that we see this, why people behave differently in an auction than they do when the situation is real. I thi nk that when we bid in an auction, we bid our best guess of what the good is really worth on the open market. Plus we are spending our money the money we have been given, but it is money that we didnt necessarily plan on having. But, when we are actually s hopping, and we are actually spending money from our regular budget, we thi nk a different way: if I spend money on this, thats money I dont have to spend on other things we bi d in a way that takes into account the limited amount of money we have In other wo rds, in the grocery store, often, you are going to purchase more than just one item and you are making decisions based on combinations of products and prices that wi ll result in the basket of food products you desire at a certain price, so if you spend more money on th is product, you will have less to spend on other products to stay within your bud get. This is just my opinion, of course, but its what I think may be going on in auctions. So, if I was in your shoes, I would ask myself : if I was grocery shopping, and I had to pay $X more to purchase this pound of apples ove r another pound of appl es, do I really want to spend my money this way? If I really di d, I would bit $X, if I didnt I would bid less than $X.

PAGE 106

94 Tomato Auction Now we will begin another auction, though this one will be slightly different, so please bear with us as we go through the instructions one more time. On the table, you see four different pounds of tomatoes. These tomatoes are labeled as from the United States, Mexico, Holland, a nd Canada. We will give you the opportunity to participate in an auction to obtain one pound of tomatoes if you so desire. In a moment, you will be asked to indicate the most amount of money you are willing to pay to for each pound of tomatoes. The procedures for this auction are similar to the apple auction, however, in this auction, we will randomly select not only one round to be binding, but we will randomly select which tomatoes will be sold. Please pull out your bid sheets labeled Tomato auction. You w ill note that you now have space to write in a price for the each of the four pounds of tomato es. Please note that you are now indicating how much you would be willing to pay for a pound of tomatoes. If you choose to bid 0, that means you are not interested in buying th at pound of tomatoes. If you bid 0 on all 4 pounds of tomatoes, that means you do not wish to leave today with any tomatoes. As this auction is slightly different, le t me go through the instructions again. Auction Procedures 1) First, each of you has been given a bid sh eet in your packet. On this sheet you will, in a moment, write the most amount of money you are willing to pay for each pound of tomatoes, as labeled by country of origin. This is the most amount of money you will pay to purchase these tomatoes. Note: your bids are private information and should not be shared with anyone. 2) After youve finished writing your bids the monitor will go around the room and collect the bid sheets. 3) In the front of the room, each of your bids will be ranked from highest to lowest for each product. 4) The four highest bids will win the auction. The individuals with the four highest bids will be the 5th highest bid amount for the exchange. 5) We will write the winning bidder nu mbers and the winning prices on the chalkboard for everyone to see. There will be four different sets of winning bidder numbers and prices one for each type of tomato. 6) After posting the price, we will re-conduct the auction for 2 additional rounds. 7) At the completion of the 3rd round, we will randomly draw a number 1 through 3 to determine the binding round and we will randomly draw which pound of tomatoes is binding. For example, if we randomly draw the number 2, and tomatoes from Holland, then we will ignor e outcomes in all other rounds and only focus on the winning bidders and price in round 2 for the tomatoes from Holland. Importantly, all rounds and types of tomato es have an equally likely chance of being binding. 8) Once the binding round has been determined, the winning bidders will come forward and pay the 5th highest bid amount and purchase their tomatoes.

PAGE 107

95 Important Notes You will only have the opportunity to win an auction for one pound of tomatoes. Because we randomly draw a binding round, you cannot win more than pound of tomatoes from this auction. The winning bidder will actually pay money to purchase their pound of tomatoes. This procedure is not hypothetical. In this auction, the best strategy is to bid exactly what it is worth to you to purchase a pound of tomatoes. It is acceptable to bid $0.00 in any ro und. This would mean that you are not willing to pay anything for the pound of tomatoes. Do you have any questions before we begin? Again, I ask you to think about this as if you are in the grocery store pu rchasing a pound of tomatoes with your other grocery items. In this case, you are going to repeat this four times and identify the price you would pay for ea ch pound of tomatoes as if it is the only pound of tomatoes you will be purchasing. If there is a pound of tomatoes that you would refuse to purchase, your bid would be $0.00. If Tomato Auction was conducted first then: Tomato Auction Now that you have had the chance to learn how the auction will work, we are interested in your preferences for two types of tomatoes. For agreeing to participate in this research session, we are giving each of you a free pound of tomatoes. These tomatoes are yours to keep. Although you have been given these tomatoes for free, we will give you the opportunity to participate in an auction to obtain different tomatoes if you so desire. Here in the front of the room, we have another tomato. These to matoes have a label on them that identifies that they were grown in the United States. Th e tomatoes are the same size and weight as the tomatoes that you have been given. We will now conduct an auction, where you will have the opportunity to exchange your tomatoes that contain NO label for the tomatoes with the label Grown in the United States. In a moment, you will be asked to indicate the most amount of money you are willing to pay to exchange your tomatoes for the tomatoes with the label by writing bids on the enclosed bid sheets. The procedures fo r this auction are exactly the same as the candy bar auction, with one exception: we request that all participants eat an tomato at the end of the todays auctions To refresh your memory as to how the auction works, I will go through the instructions again.

PAGE 108

96 Auction Procedures 9) First, each of you has been given a bid sh eet in your packet. On this sheet you will, in a moment, write the most amount of money you are willing to pay so that you would be willing to exchange your tomatoes for the tomatoes with the label Grown in the United States. This is the most amount of money you will pay to exchange your unlabeled tomatoes for my labeled tomatoes. Note: your bids are private information and should not be shared with anyone. 10) After youve finished writing your bids the monitor will go around the room and collect the bid sheets. 11) In the front of the room, each of your bids will be ranked from highest to lowest. 12) The four highest bids will win the auction. The individuals with the four highest bids will be the 5th highest bid amount for the exchange. 13) We will write the winning bidder nu mbers and the winning price on the chalkboard for everyone to see. 14) After posting the price, we will re-conduct the auction for 6 additional rounds. 15) At the completion of the 7th round, we will randomly draw a number 1 through 7 to determine the binding round. For example, if we randomly draw the number 5, then we will ignore outcomes in all other rounds and only focus on the winning bidders and price in round 5. Importan tly, all rounds have an equally likely chance of being binding. 16) Once the binding round has been determined, the winning bidders will come forward and pay the 5th highest bid amount and exchange their tomatoes for the labeled tomatoes. All other participants will keep their tomatoes. Important Notes You will only have the opportunity to win an auction for one pound of tomatoes. Because we randomly draw a binding round, you cannot win more than one pound of tomatoes from this auction. The winning bidder will actually pay money to exchange their tomatoes for the labeled tomatoes. This procedure is not hypothetical. We will also expect participants to eat a tomato at the conclusion of the auctions. In this auction, the best strategy is to bid exactly what it is worth to you to exchange your tomatoes for the labeled tomatoes. It is acceptable to bid $0.00 in any ro und. This would mean that you are not willing to pay anything to trade your unlabel ed tomatoes for the labeled tomatoes. Importantly, we are interested in what it is worth to you to exchange your tomatoes for the labeled tomatoes. Do you have any questions before we begin? Before we begin, we would like to ask you to think about the following statement: This is not a hypothetical setting; you will pay money to exchange your tomatoes if you win the auction. The reason we use these auctions instead of, say for example, a survey, is that it is closer to real life. However, it is not exactly a real situation; therefore,

PAGE 109

97 people do not always bid the same as they w ould truly behave in the grocery store. Frequently, people bid a highe r amount that they would pay in the grocery store. Let me tell you why I think that we see this, why people behave differently in an auction than they do when the situation is real. I thi nk that when we bid in an auction, we bid our best guess of what the good is really worth on the open market. Plus we are spending our money the money we have been given, but it is money that we didnt necessarily plan on having. But, when we are actually s hopping, and we are actually spending money from our regular budget, we thi nk a different way: if I spend money on this, thats money I dont have to spend on other things we bi d in a way that takes into account the limited amount of money we have In other wo rds, in the grocery store, often, you are going to purchase more than just one item and you are making decisions based on combinations of products and prices that wi ll result in the basket of food products you desire at a certain price, so if you spend more money on th is product, you will have less to spend on other products to stay within your bud get. This is just my opinion, of course, but its what I think may be going on in auctions. So, if I was in your shoes, I would ask myself : if I was grocery shopping, and I had to pay $X more to purchase this pound of tomatoes over another pound of to matoes, do I really want to spend my money this way? If I really did, I would bit $X, if I didnt I would bid less than $X. Apple Auction Now we will begin another auction, though this one will be slightly different, so please bear with us as we go through the instructions one more time. On the table, you see five different pounds of apples. These apples are labeled as from the United States, Chile, China, Canada, and New Zealand. We will give you the opportunity to participate in an auction to obtain one pound of apples if you so desire. In a moment, you will be asked to indicate the most amount of money you are willing to pay to for each pound of apples. The procedures for th is auction are similar to the tomato auction, however, in this auction, we will randomly select not only one round to be binding, but we will randomly select which a pples will be sold. Please pull out your bid sheets labeled Apple auction. You will note that you now have space to write in a price for the each of the five pounds of apples. Please note that you are now indicating how much you would be willing to pay for a pound of apples. If you choose to bid 0, that means you are not interested in buying that pound of apples. If you bid 0 on all 5 pounds of apples, that means you do not wish to leave today with any apples. As this auction is slightly different, let me go th rough the instructions again. Auction Procedures 9) First, each of you has been given a bid sh eet in your packet. On this sheet you will, in a moment, write the most amount of money you are willing to pay for each pound of apples, as labeled by country of origin. This is the most amount of money you will pay to purchase these a pples. Note: your bids are private information and should not be shared with anyone.

PAGE 110

98 10) After youve finished writing your bids the monitor will go around the room and collect the bid sheets. 11) In the front of the room, each of your bids will be ranked from highest to lowest for each product. 12) The four highest bids will win the auction. The individuals with the four highest bids will be the 5th highest bid amount for the exchange. 13) We will write the winning bidder nu mbers and the winning prices on the chalkboard for everyone to see. There will be five different sets of winning bidder numbers and prices one for each type of apple. 14) After posting the price, we will re-conduct the auction for 2 additional rounds. 15) At the completion of the 3rd round, we will randomly draw a number 1 through 3 to determine the binding round and we will randomly draw which pound of apples is binding. For example, if we randomly draw the number 2, and apples from China, then we will ignore outcomes in all other rounds and only focus on the winning bidders and price in round 2 for the apples from China. Importantly, all rounds and types of apples have an equa lly likely chance of being binding. 16) Once the binding round has been determined, the winning bidders will come forward and pay the 5th highest bid amount and purchase their apples. Important Notes You will only have the opportunity to win an auction for one pound of apples. Because we randomly draw a binding round, you cannot win more than pound of apples from this auction. The winning bidder will actually pay money to purchase their pound of apples. This procedure is not hypothetical. In this auction, the best strategy is to bid exactly what it is worth to you to purchase a pound of apples. It is acceptable to bid $0.00 in any ro und. This would mean that you are not willing to pay anything for the pound of apples. Do you have any questions before we begin? Again, I ask you to think about this as if you are in the grocery store pu rchasing a pound of tomatoes with your other grocery items. In this case, you are going to repeat this four times and identify the price you would pay for each pound of apples as if it is the only pound of apples you will be purchasing. If there is a pound of apples that you would refuse to purchase, your bid would be $0.00

PAGE 111

APPENDIX B WRITTEN QUESTIONNAIRE

PAGE 112

100 SURVEY ON FRUIT AND VEGETABLE CONSUMPTION Introduction: The purpose of this study is to bette r understand consumers thoughts about food consumption. This survey will take approximately 20 minutes to complete. Please note that there are no right or wrong answers to the following questions. We are simply interested in your opinions. Please be assured that all answers w ill be kept strictly confidential and used only for the purposes of this research. Thank you very much for your time and input! This survey is sponsored by the International Agri cultural Trade and Policy Ce nter, Food and Resource Economics Department, University of Florida. IRB#2003-U-936. Section 1: Fresh Fruit and Vegeta ble Production, Quality, and Safety Instructions : Please answer the following questions regarding how much you have heard about food production, quality, and safety. 1. How much have you seen, read, or heard about food quality? A great deal (1) Not much (3) Some (2) Nothing at all (4) 2. How much have you seen, read, or heard about food safety? A great deal (1) Not much (3) Some (2) Nothing at all (4) 3. How much have you seen, read, or heard about food quality in fresh fruits and vegetables? A great deal (1) Not much (3) Some (2) Nothing at all (4) 4. How much have you seen, read, or heard about food safety in fresh fruits and vegetables? A great deal (1) Not much (3) Some (2) Nothing at all (4)

PAGE 113

101 5. How knowledgeable would you say you are about food safety issues? 1 2 3 4 5 6 7 8 9 Not at all Knowledgeable Moderately Knowledgeable Extremely Knowledgeable 6. How knowledgeable would you say you are about labor practices in agricultural production? 1 2 3 4 5 6 7 8 9 Not at all Knowledgeable Moderately Knowledgeable Extremely Knowledgeable 7. How knowledgeable would you say you are about choosing the highest quality fresh fruits and vegetables? 1 2 3 4 5 6 7 8 9 Not at all Knowledgeable Moderately Knowledgeable Extremely Knowledgeable 8. Rinsing fresh fruits and vegetables before eating them is recommended: True (1) False (2) Dont know (3) 9. All fresh fruits and vegetables sold in U.S. grocery stores come from U.S. producers: True (1) False (2) Dont know (3) 10. It is not possible for fresh fruits and vegetables to have bacteria like E. coli that cause food borne illnesses: True (1) False (2) Dont know (3) 11. The method that fruit and vegetable growers use to produce their product can impact the safety of eating those products: True (1) False (2) Dont know (3) 12. U.S. fruit and vegetable growers are more likely to use mechanical harvesting (versus human labor) than foreign producers: True (1) False (2) Dont know (3) 13. Most fresh fruit and vegetables sold in the U.S. have no pesticide residues on them: True (1) False (2) Dont know (3) 14. Integrated Pest Management (IPM) is a production practice that is designed to reduce the frequency and intensity of pesticides used by fruit and vegetable growers. True (1) False (2) Dont know (3)

PAGE 114

102 Section 2: Your Trust of Food Production Instructions : These questions are intended to help us understand the level of trust that you have for food production. In particular, were interested in how y ou determine whether these pr oducts are trustworthy. In answering these questions, consider your feelings in ge neral, rather than your f eelings about a particular type of food. Subsection A : Please indicate the extent to which you agree or disagree with each of the statements you find below by circling the number that most closely describes your personal view, where 1 = Strongly Disagree and 9 = Strongly Agree. Strongly Neither Strongly Disagree Agree nor Disagree Agree Food production in the United States can be trusted. 1 2 3 4 5 6 7 8 9 Food production in Mexico can be trusted. 1 2 3 4 5 6 7 8 9 Food production in Canada can be trusted. 1 2 3 4 5 6 7 8 9 Food production in Europe can be trusted. 1 2 3 4 5 6 7 8 9 Food production in China can be trusted. 1 2 3 4 5 6 7 8 9 Food production in Chile can be trusted. 1 2 3 4 5 6 7 8 9 Instructions : Please indicate the extent to which you trust or distrust information about food production from each of the organizations where 1 = Strongly Distru st and 9 = Strongly Trust. Strongly Strongly Distrust Trust U.S. Government Agencies (e.g., USDA, FDA, EPA, etc.) 1 2 3 4 5 6 7 8 9 Agricultural and Food Businesses (e.g., Monsanto, Del Monte, etc.) 1 2 3 4 5 6 7 8 9 News Reports (e.g., 20/20, Dateline, etc.) 1 2 3 4 5 6 7 8 9 Activist Groups (e.g., Green Peace, P ublic Citizen, etc.) 1 2 3 4 5 6 7 8 9 U.S. Universities 1 2 3 4 5 6 7 8 9 1. Please answer the following questions ab out the auctions you just participated in: In the tomato auction, did you ever bid more than $0.00? Yes (1) No (0) If Yes, please briefly list the reasons why you were willing to pay money to exchange/purchase your tomatoes: 2. In the apple auction, did you ever bid more than $0.00? Yes (1) No (0) If Yes, please briefly list the reasons why you were willing to pay money to exchange/purchase your apples:

PAGE 115

103 Section 3: Food Purchase Decision Methods Instructions : Please indicate the extent to which you agree or disagree with each of the statements you find below by circling the number that most closely describes your personal view, where 1 = strongly disagree and 9 = strongly agree. Strongly Neither Strongly Disagree Agree nor Disagree Agree Fruits and vegetables produced in the United States are more likely to be safe. 1 2 3 4 5 6 7 8 9 I think about food safety when I purchase fresh fruits and vegetables. 1 2 3 4 5 6 7 8 9 I think about food safety when I purchase meats. 1 2 3 4 5 6 7 8 9 The U.S. agricultural food industry provides the safest, most affordable food supply in the world. 1 2 3 4 5 6 7 8 9 If the U.S. could buy all its food from other countries cheaper than it can be produced and sold here, we should. 1 2 3 4 5 6 7 8 9 I believe there are currently too many chemical pesticide residues on fresh fruits and vegetables. 1 2 3 4 5 6 7 8 9 To protect the environment, we must change the way we produce our nations food. 1 2 3 4 5 6 7 8 9 If used as directed, fertilizers, pesticides, and ot her agricultural chemicals are not a threat to the environment. 1 2 3 4 5 6 7 8 9 Fruits and vegetables produced in the United States are more likely to be fresher. 1 2 3 4 5 6 7 8 9 Fruits and vegetables produced in other countries are more likely to be of higher quality. 1 2 3 4 5 6 7 8 9 I have made purchase decisions on consumer products (e.g. shoes or clothes) based on the labor practices of the producer. 1 2 3 4 5 6 7 8 9 I have made purchase decisions on food products based on the labor practices of the producer. 1 2 3 4 5 6 7 8 9 I believe companies should follow fair labor practices. 1 2 3 4 5 6 7 8 9 Smaller agricultural operations are more likely to use better management practices. 1 2 3 4 5 6 7 8 9

PAGE 116

104 I am willing to eat fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9 I am willing to serve my family fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9 School lunch programs should be able to purchase fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9 When I grocery shop, I am willing to purchase fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9 At restaurants, I am willing to eat fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9 Prisons should be able to purchase fresh fruits and vegetables from any country. 1 2 3 4 5 6 7 8 9

PAGE 117

105 Section 4: Your Views on Food and the Environment Instructions : We are interested in your personal opinions on a variety of topics including food, health, the natural environment and technology, in general Subsection A : Environment Please indicate the extent to which you agree or disagree with each of the statements you find below by circling the number that most closely describes your personal view. Please answer the questions with regard to your view on the natural environment, in general. Strongly Neither Strongly Disagree Agree nor Disagree Agree We are approaching the limit of the number of people the earth can support. 1 2 3 4 5 6 7 8 9 The balance of nature is very delicate and easily upset by human activities. 1 2 3 4 5 6 7 8 9 Humans have no right to modify the natural environment to suit their needs. 1 2 3 4 5 6 7 8 9 When humans interfere with nature, it often produces disastrous consequences. 1 2 3 4 5 6 7 8 9 Humans must live in harmony with nature in order to survive. 1 2 3 4 5 6 7 8 9 Humans need not adapt to the natural environment because they can remake it to suit their needs. 1 2 3 4 5 6 7 8 9 The earth is like a spaceship with only limited room and resources. 1 2 3 4 5 6 7 8 9 Plants and animals exist primarily to be used by humans. 1 2 3 4 5 6 7 8 9 Modifying the environment for human use seldom causes serious problems. 1 2 3 4 5 6 7 8 9 Mankind was created to rule over the rest of nature. 1 2 3 4 5 6 7 8 9 Mankind is severely abusing the environment. 1 2 3 4 5 6 7 8 9 There are limits to growth beyond which our industrialized society cannot expand. 1 2 3 4 5 6 7 8 9

PAGE 118

106 Subsection B : Food Preferences Please indicate the extent to which you agree or disagree with each of the statements you find below by circling the number that most closely describes your personal view. Please answer the questions with regard to your view on food preferences, in general. Strongly Neither Strongly Disagree Agree nor Disagree Agree I like foods from different countries. 1 2 3 4 5 6 7 8 9 Ethnic food looks too weird to eat. 1 2 3 4 5 6 7 8 9 I like to try new ethnic restaurants. 1 2 3 4 5 6 7 8 9 At parties, I will try a new food. 1 2 3 4 5 6 7 8 9 I am very particular about the foods I will eat. 1 2 3 4 5 6 7 8 9 I am constantly sampling new and different foods. 1 2 3 4 5 6 7 8 9 I dont trust new foods. 1 2 3 4 5 6 7 8 9 I will eat almost anything. 1 2 3 4 5 6 7 8 9 If I dont know what is in a food, I wont try it. 1 2 3 4 5 6 7 8 9 I am afraid to eat things I have never eaten before. 1 2 3 4 5 6 7 8 9

PAGE 119

107 Subsection C : Food Quality Please indicate the extent to which you agree or disagree with each of the statements you find below by circling the number that most closely describes your personal view. Please answer the questions with regard to y our view on food quality, in general. Strongly Neither Strongly Disagree Agree nor Disagree Agree I usually aim to eat natural food. 1 2 3 4 5 6 7 8 9 I am willing to pay somewhat more for a product of better quality. 1 2 3 4 5 6 7 8 9 Quality is decisive for me in purchasing foods. 1 2 3 4 5 6 7 8 9 I always aim at the best quality. 1 2 3 4 5 6 7 8 9 When choosing foods, I try to buy products that do not contain residuals of herbicides and antibiotics. 1 2 3 4 5 6 7 8 9 I am willing to pay somewhat more for food containing natural ingredients. 1 2 3 4 5 6 7 8 9 For me, wholesome nutrition begins with the purchase of foods of high quality. 1 2 3 4 5 6 7 8 9

PAGE 120

108 Section 5: Questions About You Instructions: In this part of the survey, we would like some background information about you, as it is a critical part of our analysis. This is an anonymous survey and your name is in no way linked to the responses. In addition, all of th is information will be treated as confidential. Results of the survey will only be used in aggregate form and only for research purposes. 1) Gender: Male (0) Female (1) 2) Age: _____ 3) What is the highest level of education you have completed? No formal education (1) University undergraduate degree (5) Less than High School diploma (2) University postgraduate degree (6) High School diploma (3) Other:_____________________ (7) Some College (4) 4) Current Employment: Full-time (1) (30 hours per week or more) Currently unemployed (4) Part-time (2) (less than 30 hours per week) Student (5) Unpaid family worker (3) (i.e., homemaker) Retired (6) 5) Please indicate your approximate household income before taxes: less than $15,000 (0) $35,000 to $49,999 (5) $100,000 to $124,999 (10) $15,000 to $24,999 (1) $50,000 to $74,999(6) $125,000 to $149,999 (11) $25,000 to $34,999(2) $75,000 to $99,999(7) more than $150,000 (12) 6) Number of children unde r age of 16 living at home: ___________? 7) Number of people currently living in your household (including yourself): _________? 8) Are you the primary shopper for groceries in your household? (The primary shopper is the person responsible for at least 50% of food purchased for the household.) Yes (1) No (0) 9) Please indicate your race as classified by the U.S. Census (select all that apply): White (1) Black or African U.S. (2) Asian or Pacific Islander (3) U.S. Indian (4) Other_________ (5) 10) Are you of Hispanic descent? Yes (1) No (0) 11) Please list your Religion: __________________________ Not applicable 12) How involved are you with your Religion? Very involved (1) Not very involved (3) Somewhat involved (2) Not involved (4) 13) Are you employed in some capacity related to the food industry? Yes (1) No (0) 14) Does anyone in your immediate fam ily farm or ranch for a living? Yes (1) No (0)

PAGE 121

109 APPENDIX C BID SHEETS Bid Sheet Apple auction Participant Number ______________ Round Number _____1 ________ The most amount of money I am willing to pay to exchange my apples for . the apples labeled Grown in the United States is . $__________ Bid Sheet Apple auction Participant Number ______________ Round Number _____2 ________ The most amount of money I am willing to pay to exchange my apples for . the apples labeled Grown in the United States is . $__________ Bid Sheet Apple auction Participant Number ______________ Round Number _____3 ________ The most amount of money I am willing to pay to exchange my apples for . the apples labeled Grown in the United States is . $__________ Bid Sheet Apple auction Participant Number ______________ Round Number _____4 ________ The most amount of money I am willing to pay to exchange my apples for . the apples labeled Grown in the United States is . $__________

PAGE 122

110 Bid Sheet Tomato auction Participant Number ______________ Round Number _____1 ________ The most amount of money I am willing to pay to exchange my tomatoes for . the tomatoes labeled Grown in the United States is . $__________ Bid Sheet Tomato auction Participant Number ______________ Round Number _____2 ________ The most amount of money I am willing to pay to exchange my tomatoes for . the tomatoes labeled Grown in the United States is . $__________ Bid Sheet Tomato auction Participant Number ______________ Round Number _____3 ________ The most amount of money I am willing to pay to exchange my tomatoes for . the tomatoes labeled Grown in the United States is . $__________ Bid Sheet Tomato auction Participant Number ______________ Round Number _____4 ________ The most amount of money I am willing to pay to exchange my tomatoes for . the tomatoes labeled Grown in the United States is . $__________

PAGE 123

111 APPENDIX D EQUIVALENCY TESTING AND SAS CODE Table D-1. SAS equiva lency testing results Obs MX MY NOBS X NOBS Y ID NN CSSX CSSY TT 1 0.48787 0.48497 136 175 1 311 45.4918 53.0168 0.044873 ALPHA M N EPS1 EPS2 IT C1 C2 ERR1 POW0 0.05 136 175 0.5 1 24 -2.72150825 7.04388936 1.339E-11 .99656648 Where: Obs = number of observed values tested for (the differential between the means) MX is the mean WTP for apples MY is the mean WTP for tomatoes NOBSX = M = number of observations in apples data NOBSY = N = number of observations in tomatoes data ID = identification number used in the data file CSSX = cross sums of squares for WTP values in apples data CSSY = cross sums of squares for WTP values in tomatoes data TT = resulting t statistic ALPHA = significance level EPS1 = lower interval scale EPS2 = upper interval scale IT = number of iterations completed by SAS program C1 = lower limit of critical region C2 = upper limit of critical region ERR1 = error in the estimation POWO = statistical power of the t statistic based test SAS Code: Testing Statistical Hypothesis of Equivalence by Stefan Welleck p. 105 macro tt2st.sas, mentioned in pg 104 was downloaded from http://zima04.zi-mannheim.de/wktsheq/ Program name: EQUIVS.SAS; OPTIONS PS=72 MISSING='.' NODATE NONUMBER; TITLE' '; %MACRO TT2ST(ALPHA,TOL,ITMAX,M,N,EPS1,EPS2); DATA ; ALPHA= &ALPHA; TOL= &TOL; ITMAX= &ITMAX; M= &M; N= &N; EPS1= &EPS1; EPS2= &EPS2; NY= M+N-2; DEL1= -EPS1*SQRT(M*N/(M+N)); DEL2= EPS2*SQRT(M*N/(M+N)); ERR1= -ALPHA; C1= (DEL1+DEL2)/2;

PAGE 124

112 DO WHILE (ERR1 <0); C1= C1-.05; AREAC1_2= PROBT(C1,NY,DEL2) ; IF AREAC1_2=. THEN GOTO S8; H= ALPHA + AREAC1_2 ; C2= TINV(H,NY,DEL2); ERR1= PROBT(C2,NY,DEL1)-PROBT(C1,NY,DEL1) -ALPHA; END; C1L= C1 ; C1R= C1 +.05; IT= 0; DO WHILE (ABS(ERR1) >=TOL AND IT <=ITMAX); IT= IT+1; C1= (C1L+C1R)/2; H= ALPHA + PROBT(C1,NY,DEL2) ; C2= TINV(H,NY,DEL2); ERR1= PROBT(C2,NY,DEL1)-PROBT(C1,NY,DEL1) -ALPHA; IF ERR1 <=0 THEN GO TO S1; ELSE GO TO S2; S1: C1R= C1; GO TO S3; S2: C1L= C1; S3: END; GOTO S9; S8: C2=TINV(ALPHA,NY,DEL2); C1=TINV(1-ALPHA,NY,DEL1); AREAC1_1=PROBT(C1,NY,DEL1); AREAC2_1= PROBT(C2,NY,DEL1); IF AREAC2_1=. THEN AREAC2_1=1; ERR1= AREAC2_1-AREAC1_1-ALPHA; AREAC1_2=PROBT(C1,NY,DEL2); AREAC2_2= PROBT(C2,NY,DEL2); IF AREAC1_2=. THEN AREAC1_2=0; ERR2= AREAC2_2-AREAC1_2-ALPHA; S9: POW0= PROBT(C2,NY)-PROBT(C1,NY); OUTPUT; RUN; PROC PRINT NOOBS; VAR ALPHA M N EPS1 EPS2 IT C1 C2 ERR1 ERR2 POW0; FORMAT C1 C2 11.8 POW0 9.8 ERR1 ERR2 E10.; RUN; %MEND TT2ST; PROC IMPORT DATAFILE="F:\APPTOM.XLS" OUT=APTO REPLACE; Variables are PMTA PMTT; DATA APTO; SET APTO; ID=1; X=PMTA; Y=PMTT; PROC MEANS MEAN N DATA=APTO; VAR X Y; OUTPUT OUT=MXY MEAN = MX MY N = NOBSX NOBSY; DATA MXY; SET MXY; KEEP MX MY NOBSX NOBSY ID NN; ID=1; NN = NOBSX + NOBSY; DATA PMXY; MERGE APTO MXY; BY ID; CXSQ = (X-MX)**2; CYSQ = (Y-MY)**2; PROC MEANS SUM DATA=PMXY; VAR CXSQ CYSQ; OUTPUT OUT=CC SUM= CSSX CSSY; DATA CC; SET CC; KEEP CSSX CSSY; DATA TT; MERGE MXY CC; TT = sqrt((NOBSX*NOBSY*(NN-2))/NN)*(MX-MY)/sqrt(CSSX + CSSY);

PAGE 125

113 PROC PRINT; %TT2ST(.05,1.E-10,50,136,175,0.50,1.00) RUN; QUIT;

PAGE 126

114 APPENDIX E DESCRIPTION OF VARIABLES Table E-1. A detailed description of va riables used in the econometric models Variable Description WTP The dichotomous consumers willingness to pa y (i.e. participation dependent variable), expressed as a probability PMT The quantitative consumers willingness to pay (i.e. consumption dependent variable) expressed as the dollar amount AGE Numerical age of the respondent GENDER A dichotomous Gender variable: Whether th e respondent was female (1) or male (0) (Male dropped) EDU1 Highest level of education completed : Some College or less EDU2 Highest level of education completed : University undergraduate degree EDU3 Highest level of education completed : Un iversity postgraduate degree (dummy dropped) LOC1 Location: Gainesville, Florida LOC2 Location: Lansing, Michigan LOC3 Location: Atlanta, Georgia (dummy dropped) INC1 Pre-tax Household Income: Less than $50,000 INC2 Pre-tax Household Income: $50,000 to $74,999 INC3 Pre-tax Household Income: $75,000 to $99,999 INC4 Pre-tax Household Income: $100,000 and above (dummy dropped) EXPOSE Self rating on level of exposure to informa tion about food safety in fruits and vegetables: Likert scale of 1-4 treated as a numerical variab le. 1 is a great deal and 4 is nothing at all PC1 Presence of Children under 16 years in the household: 0 present PC2 Presence of Children under 16 year s in the household: 1 child present PC3 Presence of Children under 16 years in the household: 2 children present PC4 Presence of Children under 16 years in the household: 3 or more present SAFE Likert scale rating on statement I think about food safety when purchasing fruit and vegetables-6 scale treated as a numerical vari able. 1 is strongly disagree and 6 is strongly agree TRUST Likert scale rating on level of trust that consumer has on information about food production from U.S. government agencies, (e.g. USDA, FDA, EPA) 1-6 scale tr eated as a numerical variable. 1 is strongly distrust and 6 is strongly trust PFR1 First numerical Food Preferen ce factor score: The more posi tive and higher it is the more preference for various foods and more open to different foods. (Open to unfamiliar foods) PFR2 Second numerical Food Preference factor score: The more positive and higher it is the less preference for unfamiliar foods and risk (Choosey about foods) PFR3 Third numerical Food Preference factor score: The more positive and higher it is the less preference for unfamiliar foods and risk (afraid of unfamiliar foods) QUAL1 First numerical Food Quality factor score: The more positive and higher it is the more conscious about food quality in general QUAL2 Second numerical Food Quality factor scor e: The more positive and higher it is the more conscious about food quality associated with natural foods Sigma Disturbance standard deviation (Included as an independent variable in the truncated model)

PAGE 127

115 APPENDIX F FACTOR ANALYSIS RESULTS Table F-1. Correlation matrix for food safety proxy variables I think about food safety when I purchase fresh fruits and vegetables Fruits and vegetables produced in the U.S. are more likely to be safe I think about food safety when I purchase meats The U.S. agricultural food industry provides the safest, most affordable food supply in the world I believe there are currently too many chemical pesticide residues on fresh fruits and vegetables I think about food safety when I purchase fresh fruits and vegetables 1.000 .116 .502 .110 .339 Fruits and vegetables produced in the U.S. are more likely to be safe .116 1.000 .044 .448 -.042 I think about food safety when I purchase meats .502 .044 1.000 .232 .156 The U.S. agricultural food industry provides the safest, most affordable food supply in the world .110 .448 .232 1.000 -.034 Correlation I believe there are currently too many chemical pesticide residues on fresh fruits and vegetables .339 -.042 .156 -.034 1.000 I think about food safety when I purchase fresh fruits and vegetables .091 .000 .103 .000 Fruits and vegetables produced in the U.S. are more likely to be safe .091 .309 .000 .317 I think about food safety when I purchase meats .000 .309 .004 .036 The U.S. agricultural food industry provides the safest, most affordable food supply in the world .103 .000 .004 .348 Sig. (1-tailed) I believe there are currently too many chemical pesticide residues on fresh fruits and vegetables .000 .317 .036 .348

PAGE 128

116 Table F-2. KMO and Bartlett's te st for food safety proxy variables Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .519 Approx. ChiSquare 94.151 df 10 Bartlett's Test of Sphericity Sig. .000 Table F-3. Total variance explaine d in food safety factor scores Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 1.807 36.139 36.139 1.807 36. 139 36.139 1.683 33.657 33.657 2 1.377 27.541 63.680 1.377 27. 541 63.680 1.501 30.023 63.680 3 .828 16.568 80.249 4 .583 11.668 91.917 5 .404 8.083 100.000 Figure F-1. Food quality factor s scree plot-apple data Table F-4. Total variance explained in food quality factor scores-apple data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.487 64.095 64.095 4.487 64. 095 64.095 2.946 42.082 42.082 2 .914 13.063 77.158 .914 13. 063 77.158 2.455 35.076 77.158 3 .491 7.014 84.172 4 .464 6.630 90.802 5 .306 4.367 95.169 6 .197 2.818 97.988 7 .141 2.012 100.000 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue

PAGE 129

117 Figure F-2. Food preferences factors scree plot-apple data Table F-5. Total variance explained in f ood preferences factor scores-apple data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.839 48.394 48.394 4.839 48. 394 48.394 2.949 29.492 29.492 2 1.196 11.955 60.349 1.196 11. 955 60.349 2.025 20.248 49.740 3 .861 8.615 68.964 .861 8. 615 68.964 1.922 19.224 68.964 4 .703 7.026 75.990 5 .634 6.341 82.331 6 .491 4.910 87.242 7 .450 4.504 91.746 8 .383 3.830 95.576 9 .244 2.438 98.013 10 .199 1.987 100.000 Figure F-3. Food quality factor s scree plot-tomato data 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue 10 9 8 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue

PAGE 130

118 Table F-6. Total variance explained in food quality factor scores-tomato data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.182 59.739 59.739 4.182 59. 739 59.739 2.774 39.624 39.624 2 .898 12.833 72.571 .898 12. 833 72.571 2.306 32.948 72.571 3 .649 9.271 81.843 4 .536 7.653 89.496 5 .380 5.429 94.925 6 .200 2.855 97.780 7 .155 2.220 100.000 Figure F-4. Food preferences f actors scree plot-tomato data Table F-7. Total variance explained in f ood preferences factor scores-tomato data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.747 47.469 47.469 4.747 47. 469 47.469 2.663 26.635 26.635 2 1.132 11.318 58.788 1.132 11. 318 58.788 2.443 24.431 51.066 3 .928 9.279 68.067 .928 9. 279 68.067 1.700 17.001 68.067 4 .695 6.949 75.016 5 .639 6.387 81.403 6 .541 5.410 86.814 7 .433 4.333 91.146 8 .389 3.890 95.036 9 .292 2.915 97.951 10 .205 2.049 100.000 10 9 8 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue

PAGE 131

119 Figure F-5. Food quality factors scree pl ot-combined apple and tomato data Table F-8. Total variance explained in f ood quality factor scores-combined apple and tomato data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.319 61.697 61.697 4.319 61. 697 61.697 2.803 40.044 40.044 2 .882 12.594 74.292 .882 12. 594 74.292 2.397 34.247 74.292 3 .584 8.337 82.629 4 .414 5.909 88.537 5 .382 5.452 93.989 6 .253 3.617 97.606 7 .168 2.394 100.000 Figure F-6. Food preferences factors scr ee plot-combined apple and tomato data 10 9 8 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue 7 6 5 4 3 2 1 Factor Number 5 4 3 2 1 0 Eigenvalue

PAGE 132

120 Table F-9. Total variance explained in food preferences factor scor es-combined apple and tomato data Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Factor Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.771 47.709 47.709 4.771 47. 709 47.709 2.756 27.560 27.560 2 1.146 11.463 59.172 1.146 11. 463 59.172 2.049 20.488 48.048 3 .891 8.911 68.083 .891 8. 911 68.083 2.003 20.035 68.083 4 .696 6.957 75.040 5 .606 6.057 81.097 6 .511 5.109 86.206 7 .459 4.592 90.798 8 .425 4.253 95.050 9 .274 2.737 97.788 10 .221 2.212 100.000

PAGE 133

121 APPENDIX G CHI-SQUARE SPECIFICATION TEST Table G-1. Chi-square values used in the specification test of each model Simple Models Lprobitrestricted Lprobit Ltruncated Ltobit 2 df Apples (Model 1) -69.14917 -65.83477 -35.4999 -126.3534 -50.0375 16 Tomatoes (Model 2) -104.6972 -93.05584 -49.69198 -152.5791 -19.6626 16 Combined (Model 3) -175.1603 -163.1073 -102.7799 -291.9253 -52.0762 16 Factor scores Models Lprobitrestricted Lprobit Ltruncated Ltobit 2 df Apples (Model 4) -69.14917 -63.89789 -32.86963 -124.4763 -55.4176 19 Tomatoes (Model 5) -104.6972 -89.77989 -47.00739 -149.1206 -24.6666 19 Combined (Model 6) -175.1603 -161.3024 -101.6101 -288.8708 -51.9166 19

PAGE 134

122 APPENDIX H PROGRESSION OF MEAN BIDS BY LOCATION Progression of mean bidsGainesville, FL0.00 0.20 0.40 0.60 0.80 1.00Round1Round2Round3Round4Round of bidding$/Lb Tomatoes (n=67) Apples (m=56) Figure H-1. Progression of mean bids in Gainesville, FL Progression of mean bidsLansing, MI0.00 0.05 0.10 0.15 0.20 0.25 0.30Round1Round2Round3Round4Round of bidding$/Lb Tomatoes (n = 49) Apples(m=17) Figure H-2. Progression of m ean bids in Lansing, MI

PAGE 135

123 Progression of mean bidsAtlanta, GA0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70Round1Round2Round3Round4Round of bidding$/Lb Tomatoes (n=59) Apples (m=63) Figure H-3. Progression of m ean bids in Atlanta, GA

PAGE 136

124 LIST OF REFERENCES Aadland, D. and A.J. Caplan, Cheap ta lk revisited: New Evidence from CVM Selected Paper presented at the American Agricultural Economics Association, Annual Meeting, July 27-30 (2003a), Montreal, Canada . Willingness to pay for Curbside Recyc ling with Detection and Mitigation of Hypothetical Bias American Journal of Agricultural Economics 85 (2003b): 492502. Adamowicz, W., R. Boxall, M. Williams and J. Louviere Stated Preference Approaches for Measuring Passive Use Values: Choice Experiments and Contingent Valuation American Journal of Agricultural Economics 80 (1998):64-75 Akgungor, S., B. Miran and C. Abay, Consumer Willingness to Pay for Reduced Pesticide Residues in Tomatoes: The Turkish Case Selected Paper presented at the American Agricultural Economics Asso ciation, Annual Meeting, August 8-11, 1999, Nashville, Tennessee Alfnes, F. and K. Rickertsen, European Cons umers WTP for U.S. Beef in Experimental Auction Markets American Journal of Agricultural Economics 85 (2003): 397406. Andersen, E. and K. Phillipsen, The Evoluti on of Credence Goods in Customer Markets: Exchanging Pigs in Pokes Paper presented at the DRUID Winter Conference, January 8-10, 1998, Middlefart, Denmark Blend, J.R. and E.O. Van Ravenswaay Consumer Demand for Eco-labeled Apples: Results from Econometric Estimation American Journal of Agricultural Economics 81 (1999): 1072-1077 Burton, M., D. Rigby, T. Young and S. Ja mes Consumer Attitudes to Genetically Modified Organisms in Food in the UK European Economic Review of Agricultural Economics 28(2001):479-498 Cragg, J.G. Some Statistical Models for Li mited Dependent Variables with Application to the Demand for Durable Goods Econometrica 39 (1971):829-844 Cox, J.C., V.L. Smith and J.M. Walker, Expected Revenue in Discriminative and Uniform Price Sealed-Bid Auctions In Research and Experimental Economics Vol. 3, ed., V.L. Smith, pp 183-232. Greenwich: JAI Press (1985)

PAGE 137

125 Darby, M.R., and E. Karni, Free Competition and the Optimal Amount of Fraud Journal of Law and Economics, 16 (1973): 67-88 Dong, D., and Gould, B., A Double-Hurd le Model of Food Demand with Endogenous Unit Values Selected Paper presented at the American Agricultural Economics Association, Annual Meeting, August 8-11, 1999, Nashville, Tennessee Eastwood, D.B, J.R. Brooker and R.H. Orr, Consumer Preferences for Local Versus Out-of-state Grown Selected Fresh Produ ce: The Case of Knoxville, Tennessee Southern Journal of Agricultural Economics 19(2) (1987): 183-194 Florida Statutes of 1979, 17th edition, Chapter 504, 504.011504.014, Produce Labeling Act of 1979 vol. 2, p1353, The State of Florida, Tallahassee Fox, J.A., J.F. Shogren, D.J. Hayes and J.B. Kleibenstein, Experimental Auctions to Measure Willingness to Pay for Food Safety In Valuing Food Safety and Nutrition ed., J.A. Caswell, p117 Westview Press, Boulder CO (1995) Experimental Methods in Consumer Preference Studies Journal of Food Distribution Research 27 (2) (1996): 1-7 Gao, X.M., Wailes, E.J. and G.L. Cramer, D ouble-Hurdle Model w ith Bivariate Normal Errors: An Application to U.S. Rice Demand Journal of Agricultural and Applied Economics. 27 (2) (1995): 363-376 Golan, E., F. Kuchler and L. Mitchell, (2000), Economics of Food Labeling Economic Research Services, U.S. Department of Agriculture Agricultural Economic Report No. 793 Hansen, F., Consumer Choice Behavior: A Cognitive Theory. The Free Press, New York (1972) Hair, J.F. Jr., R.E. Anderson, R.L. Tatham and W.C. Black, Multivariate Data Analysis Fifth Edition. Pearson Education, Delhi (2003) Harrison, G.W. and E.E. Rutstrom, (2005) Experimental Evidence on the Existence of Hypothetical Bias in Value Elic itation Methods. Forthcoming in Handbook of Results in Experimental Economics ed. C. Plott. Elsevier Science, New York Hayes, D.F., J.F. Shogren, S.Y. Shin and J.B. Kliebenstein, Valuing food safety in experimental auction markets. American Journal of Agricultural Economics 77 (1995):40-53 Hurley, S.P., D.J. Miller and J.B. Kliebens tein, Estimating Willingness-To-Pay Using a Polychotomous Choice Function: An A pplication to Pork Products with Environmental Attributes Selected Paper presented at the American Agricultural Economics Association, Annual Meeting August 3, 2004, Denver, CO

PAGE 138

126 Kreps, D. M. (1988), Notes on the Theory of Choice Westview Press, Boulder, Colorado. Lancaster, K. A New Approach to Consumer Theory Journal of Political Economy 74 (2) (1966): 132-157 List, J.A., Do Explicit Warnings Eliminat e the Hypothetical Bias in Elicitation Procedures? Evidence from Field Auctions of Sportscards American Economic Review. 91(5) (2001): 1498-1507 List. J.A., and G.A. Gallet. What Experiment al Protocol Influence Disparities between Actual and Hypothetical Stated Values? Environmental an d Resource Economics 20(2001):241-254 List. J.A., and J.F. Shogren. Calibration of the Differences between Actual and Hypothetical Valuations in a Field Experiment Journal of Economic Behavior and Organization 37(1998):193-205 Loureiro, M.L, J. J. McCluskey and R. C. Mittelhammer, Assessing Consumer Preferences for Organic and Ec o-labeled and Regular Apples Journal of Agricultural and Resource Economics 26(2) (2001):404-416 Loureiro, M.L. and W.J. Umberger, Es timating Consumer Willingness-to-Pay for Country-of-Origin Labels for Beef Products Selected Paper presented at the American. Agricultural Economics Association, Annual Meeting, July 28-31, 2002, Long Beach, CA . A Choice Experiment Model for Beef Attributes: What Consumer Preferences Tell Us Selected Paper Presented at the Americ an Agricultural Economics Association Annual Meeting, August 1-4, 2004, Denver CO Lusk, J. L., J. Roosen, and J.A. Fox, Dema nd for Beef from Cattle Administered Growth Hormones or Fed Genetically Modified Corn: A Comparison of Consumers in France, Germany, the United Kingdom, and the United States American Journal of Agricultural Economics 85(1) (2003): 16-29 Lusk, J. L., C. Alexander, and M. Rousu, Designing Experimental Auctions for Marketing Research: Effect of Value, Di stributions, and Mechanisms of Incentives for Truthful Bidding Selected Paper presented at the American Agricultural Economics Association, Annual Meeting, August 3, 2004, Denver Colorado McFadden, D. Conditional Logit Analysis of Qualitative Choice Behavior In Frontiers in Econometrics ed. P. Zarembka pp 105-142. Academic Press, New York (1973) Maine Statute of 1989, Title 7, Chapter 101, Subchapter 6, Section 530, General Provisions, Heading PL 1989, c.527 @1(new), p29. http://janus.state.me.us/legi s/statutes/7/title7sec530.html Retrieved January 18, 2005

PAGE 139

127 Maine Statute of 1999, Title 7, Chapter 101, Subchapter 6, Section 530, General Provisions, Heading PL 1999, c.405 @2 (amd), p29. http://janus.state.me.us/legi s/statutes/7/title7sec530.html Retrieved January 18, 2005 Maynard L. J., J. G. Hartell, A. L. Meyer, and J. Hao (2003), An Experimental Approach to Valuing New Differentiated Products Contributed paper presented at the International Association of Ag ricultural Economis ts, Annual Meeting, August 16-22, 2003, Durban, South Africa McCluskey J. J. and M. L. Loureiro, Consumer Preferences and WTP for Food Labeling: A Discussion of Empirical Studies Journal of Food Distribution Research 34 (3) (2003): 95-102 Menkhaus, D.J. (2001), Experimental Auctio ns: New Theoretical Developments and Empirical FindingsDiscussion Selected Paper presented at the Western Agricultural Economics Association, Annual Meeting, July 8-11, 2001, Logan, Utah Nelson, P. Information and Consumer Behavior Journal of Political Economy, 78 (1970): 311-329 Roosen, J., J.A. Fox, D.A. Hennessy and A. Schreiber, Consumers Valuation of Insecticide Use Restrictions: An Application to Apples Journal of Agricultural and Resource Economics, 23 (2)(1998): 367-384 Roosen, J., J.L. Lusk and J.A. Fox, Consumer Demand for and Attitudes towards Alternative Beef Labeling Strategies in France, Germany and the UK Agribusiness: An In ternational Journal 19(1) (2003): 77-90 Rousu, M., W. Huffman, J.F. Shogren and A. Tegene, Are U.S. Consumers Tolerant of G.M. Foods? Working Paper No. 02014, De partment of Economics, Iowa State University, October 2002 Samnaliev, M., T. Stevens and T. More, A Comparison of Cheap Talk and Alternative Certainty Calibration Techni ques in Contingent Valua tion Working Paper No. 2003-11, Department of Resource Economi cs, University of Massachusetts Amherst Schupp, A. and J. Gillespie, Consumer Attit udes towards Potential Country of Origin Labeling of Fresh and Frozen Beef, Journal of Food Distribution Research 32 (3) (2001): 34-44 Shogren, J.F., J. List and D. Hayes, Pre ference Learning in Consecutive Experimental Auctions American Journal of Agricultural Economics 82 (4) (2000):1016-1021 Shogren, J.F., S.Y. Shi, D.J. Hayes and J. B. Klienbenstein, Resolving differences in WTP and Willingness to Accept American Economic Review, 84 (2002): 255-270

PAGE 140

128 Solomon, M.R., Consumer Behavior: Buying, Having and Being. Pearson Prentice Hall, Upper Saddle River, New Jersey (2004) Thompson, B., Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications American Psychological Associ ation, Washington, DC (2004) Thompson, G.D., Retail Demand for Gr eenhouse Tomatoes: Differentiated Fresh Produce A selected paper presented at th e American Agricultural Economics Association Annual Meeting, July 27-30, 2003, Montreal, Canada Tobin, J. Estimation of Relationships for Limited Dependent Variables Econometrica 26 (1958):24-36 Umberger, W.J., D.M. Feuz, C.R. Calkins a nd K. Killinger. U.S. Consumers Preference and Willingness-to-Pay for Domestic Corn-fed versus International Grass-Fed Beef Measured through an E xperimental Auction Agribusiness: An International Journal 18 (4) (2002): 491-504 Umberger, W.J., D.M. Feuz, C.R. Calkins and B.M. Stitz, Country of Origin Labeling of Beef Products: U.S. Consumers Perceptions Journal of Food Distribution Research 34(3) (2003): 103-116 U.S. Census Bureau. United States Census, 2000 Internet address: http://www.census.gov/prod/2001pubs/c2kbr01-9.pdf Retrieved March 12, 2005 United States Statutes at Large, Agricultural Marketing Act, 79th Congress, 2nd Session vol. 60, Part 1, Public Laws Reorganization Plans, (1947), p1089 VanSickle, J.J., Country of Origin La beling-A COOL Update, International Agricultural Trade and Policy Center. Policy Brief Series, PBTC 03-17 December 2003 Verbeke, W. and R.W. Ward, Importance of EU Label Requirements: An Application of Ordered Probit Models to Belgium Beef Labels A Selected Paper presented at the U.S. Agricultural Economics Association Annual Meeting, July 27-30, 2003, Montreal, Canada Wellek, S., Testing Statistical Hypotheses of Equivalence Chapman and Hall/CRC, Boca Raton (2003)

PAGE 141

129 BIOGRAPHICAL SKETCH Born in the city of Bulawayo, Zimbabwe Athur graduated high school at Christian Brothers College, Bulawayo, in 1998 before attending the University of Zimbabwe where he earned a Bachelor of Science degr ee with Honors in agri cultural economics. In August 2003 Athur began the Food and Resource Economics Master of Science program at the University of Florida and specia lized in agricultural food marketing.


Permanent Link: http://ufdc.ufl.edu/UFE0011841/00001

Material Information

Title: Estimating Consumers' Willingness-to-Pay for Country-of-Origin Labels in Fresh Apples and Tomatoes: A Double-Hurdle Probit Analysis of U.S. Data Using Factor Scores
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011841:00001

Permanent Link: http://ufdc.ufl.edu/UFE0011841/00001

Material Information

Title: Estimating Consumers' Willingness-to-Pay for Country-of-Origin Labels in Fresh Apples and Tomatoes: A Double-Hurdle Probit Analysis of U.S. Data Using Factor Scores
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011841:00001


This item has the following downloads:


Full Text












ESTIMATING CONSUMERS' WILLINGNESS-TO-PAY FOR
COUNTRY-OF-ORIGIN LABELS IN FRESH APPLES AND TOMATOES: A
DOUBLE-HURDLE PROFIT ANALYSIS OF U.S. DATA USING FACTOR SCORES















By

ATHUR MABISO


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Athur Mabiso



































To my parents















ACKNOWLEDGMENTS

Firstly, I thank Almighty God for the opportunity and blessing of education. I

consider the successful completion of this work a gift from God and I am truly grateful.

To my family and friends who provided unquantifiable support and encouragement I

express my gratitude.

I sincerely thank the International Agricultural Trade and Policy Center, the Florida

Agricultural Experiment Station and the Food and Resource Economics Department for

providing the necessary funding for this research and my master's studies at the

University of Florida. This thesis would not have been without their financial support.

My utmost thanks to my thesis committee, Drs. James Stems, Lisa House, Allen

Wysocki, and John VanSickle, who on numerous occasions offered their time, solid

advice, guidance, and instruction in the production of this work. Special

acknowledgement goes to my committee chair Dr. James Sterns whose guidance and

confidence to let me try different approaches made the process of writing this thesis an

adventurous learning experience. Also, special thanks go to Dr. Lisa House for being

available to share her knowledge and insights on data analysis. For assistance with certain

aspects of statistics and equivalency testing, I thank Mr. Carlos Jauregui.

Finally, to my fellow graduate students, the FRE-GSO, BGSO, ASU, OCS,

MANRRS, Florida Rotaract Club and the rest of the alphabet soup; I am thankful. They

made my graduate studies experience at Florida forever memorable.
















TABLE OF CONTENTS

Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ............................... ............. .............. viii

LIST OF FIGURES ............................... ... ...... ... ................. .x

ABSTRACT .............. .................. .......... .............. xi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

B background .................. ....................................... .... ........ ................1
Problem atic Situation................................................... 3
Problem Statem ent .................. ......................................... ................ .4
Testable H ypotheses ........................................ ....... ... ........ .. ........ .. ..
R research O bjectives.......... ................................................................ ........ .... .5
Specific Objectives .............. ................. ................................... .5
O outline of T hesis ............................................................................................ ........ 5

2 LITERATURE REVIEW ........................................................................7

The History of Country-of-Origin Labeling.....................................................7
P reviou s Stu dies ................................. ............................................. 8
Food Labeling in Fresh Apples and Tomatoes..........................................8
Country-of-Origin Labeling Studies......................................... ............... 10
E xperim mental A u action s ....................................................................... ..................13

3 RESEARCH METHODS AND DATA ........................................ ............... 18

Stu dy D design an d D ata .................................................................... .......... .. .. .. 18
A auction B id D ata C collection ........................................................... .....................23
Candy B ar A auction ............................................. .... .... ............... ... 23
A pple and T om ato A uctions.................................................................... ..... 25
W written Q uestionnaire........ ................................................................ .... .... ..... .. 26

4 THEORETICAL FRAMEWORK AND ANALYTIC TOOLS.............................28




v









The Psychology and Theory of WTP Decisions...................................................28
T heoretical F ram ew ork .............................................. ........................... ........... 1
Analytic Tools ............................. ............... 33
The D ouble H urdle M odel............... ... ............................. ......................... ...33
Factor Analysis (Principle Component Analysis) ............................................35

5 AUCTION BID ANALYSIS AND RESULTS..................... ...... ............... 40

6 EMPIRICAL SPECIFICATION AND RESULTS .......................................... 48

E m pirical Sp ecification .................................................................... .. ............... 4 8
D description of V ariables......................................................... .............. 49
Organization of M odel Estimations.................. ............................................. 54
M odels w without Factor scores.......................................................... ............... 54
A pple M odel (M odel 1)............................................................ .....................54
Tomatoes Model (Model 2)............................................................. 58
Combined Apples and Tomatoes Model (Model 3)................. ............ .....60

7 FACTOR ANALYSIS RESULTS ........................................ ........................ 63

8 ECONOMETRIC MODELING WITH FACTOR SCORES................................70

A pples M odel (M odel 4) ................................................................. .....................70
Tom atoes M odel (M odel 5) ............................................................ ............... 72
Combined Apples and Tomatoes Model (Model 6) ....................................... 75

9 CONCLUSIONS AND IMPLICATIONS ...................................... ............... 79

S u m m ary ...................................... .................................................. 7 9
Im p location s .......................................................................... 8 1
A reas for Further Research........................................................... ............... 85

APPENDIX

A EXPERIMENTAL AUCTIONS INSTRUCTIONS........................ .....................88

B W R ITTEN Q U ESTION N A IRE ................................................................................99

C B ID S H E E T S ................................................................................. .................... 10 9

D EQUIVALENCY TESTING AND SAS CODE.................. ...................................111

E DESCRIPTION OF VARIABLES................................................................ 114

F FACTOR ANALYSIS RESULTS ................................................................ 115

G CHI-SQUARE SPECIFICATION TEST......................................................121

H PROGRESSION OF MEAN BIDS BY LOCATION....... ...... ....... ...........122









L IST O F R E FE R E N C E S ......................................................................... ................... 124

BIOGRAPHICAL SKETCH ............................................................. ..................129
















LIST OF TABLES


Table page

3-1. Data distribution across location ...................... ............. ... ............ 19

3-2. Demographic profile of respondents.................... .......... .. .. ................ 21

5-1. Average WTP for apples (n = 136) and tomatoes (n = 175)...............................42

5-2. Average WTP for apples (n = 108) and tomatoes (n = 126): Sampling only those
consumers who were W TP more than $0.00................................. ............... 43

5-3. M ean W TP across location ............................................. ............................. 44

6-1. Apples probit model without factor scores (Model 1).......................................56

6-2. Apples truncated tobit model without factor scores (Model 1) .............................57

6-3. Tomatoes probit model without factor scores (Model 2) ......................................58

6-4. Tomatoes truncated tobit model without factor scores (Model 2) .........................59

6-5. Combined apples and tomatoes probit model without factor scores (Model 3).......60

6-6. Combined apples and tomatoes truncated tobit model without factor scores (Model
3) .............. ................... ..................................... ......... ....... 6 1

7-1. Rotated component matrix(a) for food quality proxy variables-apple data .............64

7-2. Rotated component matrix(a) for food preference proxy variables-apple data........65

7-3. Rotated component matrix(a) for food quality proxy variables-tomato data..........66

7-4. Rotated component matrix(a) for food preference proxy variables-tomato data.....67

7-5. Rotated component matrix(a) for food quality proxy variables-combined apple and
tom ato data set ................................................... ................. 68

7-6. Rotated component matrix(a) for food preference proxy variables-tomato data.....69

8-1. Apples probit model with factor scores (Model 4).................... ...............71









8-2. Apples truncated tobit model with factor scores (Model 4) ...................................71

8-3. Tomatoes probit model with factor scores (Model 5) ...........................................72

8-4. Tomatoes truncated tobit model with factor scores (Model 5)..............................73

8-5. Combined apples and tomatoes probit model with factor scores (Model 6)............75

8-6. Combined apples and tomatoes truncated tobit model with factor scores (Model 6)77

9-1. Comparison of mean bids: U.S.A. Grown versus Other Country labels..................86

D -1. SA S equivalency testing results ....................................................... ....... ........ .111

E-1. A detailed description of variables used in the econometric models .....................114

F-l. Correlation matrix for food safety proxy variables.............................115

F-2. KMO and Bartlett's test for food safety proxy variables ............. ... .................116

F-3. Total variance explained in food safety factor scores.......................... .........116

F-4. Total variance explained in food quality factor scores-apple data ......................116

F-5. Total variance explained in food preferences factor scores-apple data................117

F-6. Total variance explained in food quality factor scores-tomato data ................118

F-7. Total variance explained in food preferences factor scores-tomato data...............118

F-8. Total variance explained in food quality factor scores-combined apple and tomato
d ata .............................................................................................. 1 1 9

F-9. Total variance explained in food preferences factor scores-combined apple and
to m ato d ata ................................................................... ............... 12 0

G-1. Chi-square values used in the specification test of each model ...........................121
















LIST OF FIGURES


Figure p

3-1. Sample distribution across locations................. .................... .................19

3-2. Comparison of gender between survey data and U.S. Census data ........................20

5-1. Line graph showing the progression of combined bids in both tomato and apple
au ctio n s ...................... .. .. ......... .. .. ................................................ 4 1

F-1. Food quality factors scree plot-apple data................................. .......... 116

F-2. Food preferences factors scree plot-apple data ................................................117

F-3. Food quality factors scree plot-tomato data .............................. ...............117

F-4. Food preferences factors scree plot-tomato data...................................................118

F-5. Food quality factors scree plot-combined apple and tomato data .......................119

F-6. Food preferences factors scree plot-combined apple and tomato data.................19

H-1. Progression of mean bids in Gainesville, FL .................................................... ... 122

H-2. Progression of mean bids in Lansing, M I .................................................... .. .......... 122

H-3. Progression of mean bids in Atlanta, GA ................ ............................... 123















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

ESTIMATING CONSUMERS' WILLINGNESS-TO-PAY FOR
COUNTRY-OF-ORIGIN LABELS IN FRESH APPLES AND TOMATOES: A
DOUBLE-HURDLE PROFIT ANALYSIS OF U.S. DATA USING FACTOR SCORES

By

Athur Mabiso
August 2005

Chair: James. A. Sterns
Major Department: Food and Resource Economics

In September 2006, Mandatory Country-of-Origin Labeling (MCOOL) policy is set

to take effect in accordance with the Omnibus Appropriations bill passed by Congress in

January 2003. U.S. producers and marketers of fresh apples and tomatoes will be subject

to this law while contending with increasing import competition for the domestic market.

This thesis explores U.S. consumers' willingness to pay (WTP) for Country-of-Origin

Labeling (COOL) in fresh apples and tomatoes, particularly the label "U.S.A. Grown," in

order to inform policymakers and industry participants of some of the implications of

COOL in the U.S. market. By analyzing how much U.S. consumers will pay for the label

"U.S.A. Grown" in fresh apples and tomatoes, and establishing the major determinants of

the WTP, the thesis provides useful empirical evidence to give good reason for MCOOL

or at the very least voluntary COOL, if costs of labeling are less than the price premiums.

U.S. consumers are found to be willing to pay approximately $0.49 more for a pound of

apples labeled "U.S.A. Grown" and $0.48 more for a pound of tomatoes labeled "U.S.A.









Grown" in comparison to unlabeled yet identical fresh apples and tomatoes, respectively.

Of the consumers surveyed, 79% are willing to pay a premium for fresh apples

labeled "U.S.A. Grown" while 72% will pay a premium for fresh tomatoes labeled

"U.S.A. Grown."

Applying the double hurdle probit model, we developed six variants of the same

model specification. Results show that consumer food quality perceptions and location

are statistically significant explanatory variables for determining the amount consumers

are willing to pay for the U.S. labeled produce. Some demographics and psychographics

also showed a significant relationship with WTP, though with less magnitude. Data were

collected using a written questionnaire and a Vickrey auction (fifth-bid sealed-price).

The written questionnaire collected data on consumer demographics and perceptions of

food safety, food quality, and food preferences among other variables. The Vickrey auction

determined the actual amount consumers are willing to pay to exchange identical unlabeled

fresh apples and tomatoes for those labeled "U.S.A. Grown." The information in this

thesis can be used by U.S. producers and marketers of fresh apples and fresh tomatoes to

develop formidable marketing strategies in their efforts to boost demand for U.S. produce

in the face of rising import competition. It is also informative to policymakers who are in

the process of making laws on COOL and traceability of agricultural food products.














CHAPTER 1
INTRODUCTION

Background

As consumer demand for agricultural food-products becomes more complex and

dynamic, food labeling is taking an increasingly important role in the food marketing

system (McCluskey and Loureiro, 2003). Consumers are constantly obtaining different

kinds of information about food-product attributes via food labels and their purchasing

decisions are influenced by these. Theoretically, consumers demand food-product

attributes (e.g. food quality or taste) not the food-product per se and the food-product is

considered to be merely a bundle of these individual attributes that give rise to utility.

Thus purchasing decisions made by consumers are based on specific food-product

attributes embodied in a food-product (Lancaster, 1966).

This is important if one is to recognize the significance of food labeling. Food

labels present information about specific food-product attributes, which potentially can

affect consumer willingness to pay (WTP) and in turn the effective demand for a food-

product. Some studies have found consumers to be willing to pay premiums for eco-

labels, organic labels, or country of origin labels (COOL) (Loureiro et al., 2001; Burton

et al., 2001; Umberger et al., 2002 respectively). Alternatively, these can be viewed as

premiums for desired food-product attributes, which the labels make claim to. Thus, in

the case of eco-labels the desired attribute would be "Environmentally-friendly,"

"organically produced" in the case of organic labels, "non-GM food" in non-GM labels,









and "U.S.A. Grown," in the case of Country of Origin Labeling (COOL). This alternative

view is more consistent with Lancaster.

Though the studies have shown consumers to be willing to pay premiums for

labels, which indicate specific food-product attributes, it is not necessarily the case that

all labels will command a price premium from consumers. Different consumers respond

differently to different labels; some willing to pay and others not. Even more so, those

who are willing to pay will pay different amounts. Producers and marketers alike are

aware of this complex mesh of possibilities, and know that labels making claim to a key

product attribute can ultimately determine if the product will be purchased and/or how

much is purchased and/or at what price.

In the U.S. fresh produce industry there is topical interest in these complexities of

food labeling, particularly COOL. This is largely due to COOL legislation in the 2002

Farm Security and Rural Investment Act. Subtitle D of this composite act specifies that

currently market actors can voluntarily label their products with COOL so as to inform

the shopper at the final point of purchase. Guidelines for voluntary COOL, which were

issued by the USDA in October 2002, apply to fresh meats (beef, pork, lamb and fish) as

well as fresh peanuts, fruits and vegetables the so-called covered products (VanSickle,

2003). These products were selected to be "covered" by the policy because proponents

asserted that these are food products most prone to food safety and health problems.

Though currently voluntary, the law on COOL is set to change to mandatory

country of origin labeling (MCOOL); unless new proposed bills, such as the Meat

Promotion Bill of 2005 are passed. The change to MCOOL was initially set to be

effective on the 30th of September 2004 but this was postponed by two years with the









passing of the Omnibus Appropriations bill in January 2003. The imminent change to

MCOOL brings to the forefront several issues of debate, surrounding the justification of

MCOOL policy. One of these major issues concerns consumers' WTP for COOL. There

is little empirical information on how much consumers are willing to pay for COOL in

different products. Information on this would be important in pointing policy and/or

decision makers in a particular direction in as far as implications of MCOOL are

concerned (Menkhaus, 2001).

This study seeks to fill part of this information gap by analyzing if U.S. primary

shoppers are willing to pay a premium for fresh apples and tomatoes labeled "U.S.A.

Grown." In addition, the study examines whether premiums for "U.S.A. Grown" labels in

fresh apples and tomatoes are product specific or not, then ascertains what the prominent

factors affecting the WTP for the "U.S.A. Grown" labels are. Thus, the study gives

several insights about COOL in the produce sector.

Problematic Situation

In 2006, MCOOL in fresh fruits and vegetables is expected to take effect, in

accordance with the revised Farm Bill of 2002. Predicted consumer response to MCOOL,

however, is lacking. U.S. producers of apples and tomatoes are uncertain if MCOOL

policy, which producers and packers expect to be very costly to implement, will have a

negative impact on consumer purchases and industry profits. They do not know what to

expect in terms of how much consumers will pay for COOL, if indeed they are willing to

pay. In addition it is unclear what the key factors affecting the consumers' WTP for

COOL are. Food safety and food quality concerns are some of the factors hypothesized to

be major drivers of WTP for COOL. However, debate remains on the issue due to lack of

empirical information.









Access to such information would be useful as U.S. producers strive to market

more competitive produce relative to import substitutes. Several players within the

marketing channel are also interested in obtaining information on whether the WTP for

U.S. labels is product specific or not. This is particularly so, given rising import

competition and speculation on the prospects of differentiating domestic produce on the

basis of "U.S.A. Grown" labels. Moreover, the potential of segmenting the market on the

basis of COOL is a reason why market players are interested in knowing more about the

consumers' WTP for COOL.

Problem Statement

It is not known if U.S. primary shoppers who purchase and consume fresh

tomatoes and apples are willing to pay a premium for produce labeled "U.S.A. Grown"

over identical produce whose country of origin is not specified. Furthermore, it is

uncertain which key factors determine their WTP for fresh apples and tomatoes labeled

"U.S.A. Grown" This leaves agribusinesses and policymakers to decision-making based

on asymmetric information about COOL policy in the fresh apple and tomato industries.

Testable Hypotheses

The primary hypotheses that this research will test are as follows:

1. If fresh apples and tomatoes are labeled "U.S.A. Grown," then U.S. primary
shoppers will be willing to pay more money for them as compared to what they are
willing to pay for identical unlabeled produce whose country of origin is unknown.

2. If consumers are willing to pay a premium for fresh apples and fresh tomatoes
labeled "U.S.A. Grown" then the premiums will be product specific and unequal.

3. Consumer perceptions about food quality, food preferences and food safety are key
factors of WTP for COOL in fresh apples and tomatoes.









Research Objectives

The overall objective of this research is to determine if a differential premium

exists between U.S. primary shoppers' WTP for fresh apples labeled "U.S.A. Grown"

and fresh apples without a label of origin. The same is sought for in the case of fresh

tomatoes. Underlying determinants of the differential premiums (if they exist) will also

be ascertained in both cases.

Specific Objectives

1. To determine if U.S. primary shoppers will be willing to pay a premium for fresh
apples and fresh tomatoes labeled "U.S.A. Grown" by calculating the mean
premiums (PMTi) recorded from each product sample and testing if they are
significantly different from zero

2. To ascertain if WTP premiums for COOL are product specific and unequal. This
will be done by performing the z-test for independent samples followed by a
premium equivalency hypothesis test for two unrelated samples.

3. To assess if perceptions about food quality, food preferences and food safety are
key factors of WTP for COOL in fresh apples and tomatoes

The last objective will be achieved by performing a double hurdle probit analysis.

In order to test the hypothesis that perceptions about food quality and food preferences

are key factors of WTP for COOL in fresh apples and tomatoes, principle component

factor analysis will be done and resulting factor scores will be used as explanatory

variables in the double hurdle probit analysis. The significance of each factor score

variable will be assessed, and so will the respective marginal effects.

Outline of Thesis

This chapter introduced the thesis, by giving an overview of the problematic

situation to be addressed and the objectives central to the study. In Chapter 2, I begin

with a discussion of the history of COOL in the U.S. followed by a review of the

literature on food labeling and COOL. Relevant previous studies are referenced and a









section on experimental auctions is also provided. In the third chapter, a description of

the data set and how it was collected is presented. The chapter includes a summary of the

auction bid procedures and the survey questionnaire used in the study. Chapter 4 opens

with a section on the psychology and theory of WTP decisions, before presenting the

theoretical framework and the analytic tools used in this thesis.

Chapter 5 is a presentation of the auction-bids price analysis and results, which

mainly entail univariate statistical analyses, while Chapter 6 is a presentation of the

double hurdle probit specification and estimations without factor scores. Chapter 7

follows, showing the results from the factor analysis performed in the process of deriving

factor scores to be incorporated in subsequent double hurdle estimations. These

estimations are presented in Chapter 8 before Chapter 9 concludes the thesis with a

summarized discussion of findings and their implications. Areas for further research are

also made reference to in the final chapter of the thesis.














CHAPTER 2
LITERATURE REVIEW

The History of Country-of-Origin Labeling

The rudiments of COOL of fresh produce in the United States can be traced back to

the Agricultural Marketing Act of 1946 (United States Statutes, 1946). In this act, the

consumer's right to knowledge or information about the agricultural produce s/he

consumes is fostered. Emphasis is, however, on information that pertains to livestock and

in most part, the act deals with various aspects of food distribution and consumer welfare

(e.g. food safety, informational labeling and general packaging standards), not COOL in

fresh produce marketing. Nonetheless, these are legal issues that underpin COOL of fresh

produce and thus unsurprisingly form the foundations of today's all encompassing COOL

policy and the proposed 2006 MCOOL policy.

Until 1998, not many amendments pertaining to COOL are seen in the 1946 Act. In

July of 1998 amendments to the act were appended to the 1999 fiscal year appropriations

bill by the Senate, only to be removed by the House-Senate conferees prior to its passage

(Schupp and Gillespie, 2001). It is in 2002 that one can actually see serious amendments

to the act that legislatively introduce COOL at the federal level. This set the stage for a

debate on COOL, pitting voluntary COOL versus MCOOL in the on-going deliberation

about the appropriate policy direction to take on the matter.

At the state level, the history of COOL presents a somewhat homogeneous profile.

In most states no law on COOL existed. It is in the private sectors of the industry that

COOL occurred and only on a voluntary basis. However, in Florida a statute was passed









in 1979 making it compulsory for retailers to label fruit and vegetables' country of origin

(Florida Statute, 1979). Similarly, in the state of Maine, a COOL law was passed in 1989

making it a requirement for any vegetable or fruit produced outside the United States to

have COOL if the origin is deemed to have lower pesticide residue standards and

regulations compared to those of the United States (Maine Statute, 1989). This was later

modified to MCOOL of all imported produce, irrespective of standards and regulations in

1999 (Maine Statute, 1999). Overall, there is limited prevalence of legislative acts on

COOL at the state level. However, in the two instances that it exists, the main premise for

its existence is food safety concerns and minimum acceptable quality standards. This

same premise stands today as one of the fundamental arguments for MCOOL. The other

is consumer freedom of informed choice. All these come under the overarching principle

of the consumer's right to information and alleviation of informational asymmetry in the

food markets (Golan et al., 2000).

Previous Studies

Food Labeling in Fresh Apples and Tomatoes

In the last decade several studies have been carried out on food labeling in apples

and very few in the case of tomatoes. Many of these have looked at consumer behavior

and consumer perceptions in the context of WTP for food labeling. For example, Blend

and Van Ravenswaay (1999) studied eco-labeling in apples and how eco-labels influence

consumer behavior by affecting their perceptions about food-product attributes. As their

study confirmed, changing the labeling on the apples and inadvertently the eco-

information provided to the consumers at the point of purchase, changed consumer

buying behavior. Consumers were willing to try eco-labeled apples as well as pay a

premium for them if made available in the marketplace. Specifically, the study showed









that at the time, more than a third of surveyed consumers were willing to pay a $0.40

premium for the eco-labeled apples.

In a more recent paper, Loureiro et al. (2001) studied WTP for organic labels and

eco-labels in apples. Using contingent valuation methods (CVM), they estimated a

maximum-likelihood multinomial logit model which included food safety, perceived

produce quality and environmental concerns together with demographic variables as

independent variables. Results are that food safety and produce quality concerns were

significant determinants of WTP for the labels. The same was ascertained for

environmental concerns and the presence of children under the age of 18 in the

household, which was found to increase the probability of an U.S. consumer to purchase

organic labeled apples. Regarding regular unlabeled apples, they established that all

significant independent variables (i.e. food safety concerns, quality perceptions,

environmental preferences, presence of children in the household, income level) had a

negative impact on the likelihood to purchase with the exception of household size,

which had a positive effect on the WTP. Generally, U.S. consumers were found to be

willing to pay for organic labeled and eco-labeled apples, with those that are eco-labeled

ranking as a second choice while regular unlabeled apples ranked last.

In the case of tomatoes, not many studies dealing with WTP for labeling have been

conducted. However, Akgungor et al. (1999) analyzed WTP for "reduced pesticide use"

labeling in tomatoes, though in the context of metropolitan areas in Turkey. Using CVM,

the study found that consumers were willing to pay a 2 percent premium for the labels. In

addition, demographics such as income, gender and household size were significant

determinants positively increasing the probability of consumers' WTP for labels.









However, age was not significant. Though this study clearly suggested a causal

relationship between WTP for labels in tomatoes and demographics, this was specific to

Turkey. The scenario in the United States may be different.

In the United States, a relatively older study by Eastwood et al. (1987) explored

some of these aspects in the produce industry. By surveying households in Knox County,

TN, they found that just over one half of the consumers from Knoxville, TN were likely

to be concerned about the origin of their tomatoes. In addition, housewives and

professionally occupied respondents were most likely to be willing to pay a premium for

locally produced tomatoes over out-of-state tomatoes. A probit estimation was used to

ascertain these results. In the case of apples, broccoli and cabbage, it did not seem to

matter where the produce came from. This suggests that different produce yield different

consumer responses when it comes to WTP for origin labels.

Thompson (2003) looked at retail demand for differentiated fresh tomatoes and

particularly greenhouse tomatoes, which consumers identify through labels. Using

descriptive statistics he established that the average market price of greenhouse labeled

tomatoes was higher than that for field-grown tomatoes in the cities of Albany, NY,

Chicago, IL, Dallas, TX, Los Angeles, CA, and Miami, FL. Only in Atlanta, GA was the

average market price for greenhouse tomatoes lower ($1.56 compared to $1.99). In

addition, he presented empirical evidence showing that greenhouse and on-the-vine

tomatoes experienced market-share growth at the expense of field-grown tomatoes,

despite a reduction in market prices of field-grown tomatoes.

Country-of-Origin Labeling Studies

Schupp and Gillespie (2001) are part of the first group of researchers to have turned

the focus onto COOL in agricultural products and specifically beef. By sampling









households in Louisiana in order to establish if consumers support MCOOL of beef in

grocery stores and restaurants, they found that on average 90.3 percent of surveyed

consumers were willing to support MCOOL. Also, they looked at factors influencing the

support for MCOOL by estimating a probit model and found the food safety concern

variable to be significant and positively affecting the probability of a consumer to support

MCOOL. Consumer preference for locally produced beef also turned out to be a

statistically significant independent variable positively affecting the likelihood to support

MCOOL.

In another paper, Loureiro and Umberger (2002) researched COOL when they

explored consumer WTP for a MCOOL program in general as well as WTP for steak and

beef hamburgers labeled "U.S. Certified beef'. By survey-sampling consumers in

different grocery store locations in Colorado, they established that consumers were

willing to pay for a MCOOL program. They then found that consumers who had

completed a high level of education and had a high income level were less likely to pay a

premium for COOL or "U.S. Certified" labels in beef. This disproved their initial

hypothesis that a more educated and wealthier consumer would pay attention to COOL

and be more likely to pay a premium for it. Moreover they showed that female consumers

are most likely to pay a premium for COOL and are more supportive of MCOOL. All

these results pertained to the U.S. certified beef hamburger.

In the case of U.S. certified steak, slightly different results were obtained with the

presence of children in the household being the only significant demographic factor of

the likelihood of WTP. Overall, the findings suggested that WTP is not only based on

variables such as food safety concern, food quality and demographics but also on the









product itself (i.e. different products would have different factors influencing the

consumers' WTP under consideration.)

Loureiro and Umberger (2004) used experimental auctions to solicit information on

U.S. consumers' WTP. Again they looked at beef and the reason being that most of the

debate pertaining to MCOOL has centered on beef owing to its link to BSE food safety

concerns which have, in the recent past, plagued the industry in isolated circumstances.

According to these authors, COOL in beef is a less important factor of consumers' WTP

as compared to food safety inspection labels, product quality labels (tenderness) or

traceability of the beef. This presents mixed results on COOL and suggests a weaker link

between COOL and food safety concerns than previously thought to be.

Similar studies in Europe, though using stated choice data, have revealed similar

responses. In Belgium Verbeke and Ward (2003) conducted a survey seeking to

determine those informational labeling cues on beef that really attract consumers' interest

(attention and importance ratings) as part of consumer risk aversion associated with food

safety. By analyzing a data set of 278 observations and using ordered probit models the

authors found that COOL is of moderate interest to consumers, while traceability is of

low interest. Instead, food quality concerns rank very high and advertising/publicity of

labeling policy positively affects consumers' interest in labels pertaining to quality and

origin.

Roosen et al. (2003), however, found that European consumers are more concerned

about COOL. English, French and German consumers were surveyed and analyses found

that French and German consumers are more concerned about the origins of their beef

than about product attributes (e.g. quality). British consumers were shown to have a high









regard for COOL but not to the same extent. Interestingly enough, high WTP premiums

were recorded for "hormone-free" labels on beef, this being the underlying unobservable

variable postulated to be the cause of the high regard for COOL in Europe.

While all these studies suggest a relationship between consumers' WTP for COOL

and food safety concerns, food quality and demographics, it is evident that other variables

may have a bearing on the nature of the relationship, e.g., the types of products under

consideration or the consumers' geographic location. With respect to fruits and

vegetables the situation is unclear, since the majority of previous research has focused on

the beef sector. Though some work has been done with apples, none is specific to the

question on COOL and little has to do with tomatoes as the product of interest. This

thesis delves into these specific issues to address COOL in fresh apples and tomatoes.

Experimental Auctions

Experimental techniques are not widely used in economic research because it is

difficult to incorporate experimental controls when dealing with people as the subject of

study. Constraining the human subject to a control is exacting if not impossible, without

breaching the necessary unbiasedness and validity in the research. Nevertheless, of late

the use of experimental techniques to research consumer WTP and product value has

gained much popularity and has had satisfactory degrees of success (Maynard et al.,

2003, Alfnes and Rikertsen 2003, Shogren et al., 2002, Rousu et al., 2002, Fox et al.,

1996, and Hayes et al., 1995).

In the case of WTP for COOL studies, a few researchers have used experimental

auctions to collect data. Others have continued to base their findings solely on Stated

Choice (SC) surveys and CVM, and yet questions have been raised about their validity.

List and Gallet (2001), List (2001) and List and Shogren (1998), for example, point out









that respondents of SC and CVM surveys tend to exaggerate their WTP. Harrison and

Rutstrom (2005) further augment this view when they summarize 39 CVM based studies

and find that hypothetical bias is significant in 31 of the 39 studies.

This has led some researchers to move away from CVM and towards a variant of

CVM, where calibration is performed (e.g. by using statistical bias functions, uncertainty

scale adjustment or "cheap talk"). Others have tended towards combining these

calibration methods with the use of experimental auctions instead of CVM. These

methods are believed to do away with most hypothetical bias associated with SC and

CVM surveys. Based on this, this study used a combination of experimental auctions,

cheap talk and a written questionnaire in an effort to minimize hypothetical bias in the

data.

However, neither calibration of CVM nor experimental auctions are without

problems. For instance, research has found that sometimes hypothetical bias remains

even after calibration of CVM. For example, Aadland and Caplan (2003a) and Samnaliev

et al. (2003) show that using "cheap talk" does not always eliminate hypothetical bias and

neither does using an uncertainty scale or a "not sure/don't know" option in the

questionnaire.

With respect to experimental auctions the sealed-bid, second-price (Vickrey)

auction has received wide recognition in terms of dealing with hypothetical bias. Named

after William Vickrey who pioneered auction research in 1961, the auction is desirable

for its truth revealing traits in the consumer's bid. Since the auction is set up in such a

way that winning bidders pay the second highest bid price for the product being

auctioned, the bidder is encouraged to bid truthfully as his/her weakly dominant strategy.









According to Lusk et al., (2004, p.1) this is the "incentive compatibility" of the Vickrey

second bid auction, meaning that it inherently provides individuals with incentives for

truthful bidding while simultaneously punishing untruthful bidding. If a consumer

underbids, s/he runs the risk of losing the auction, whereby s/he could have won at

his/her true WTP value. As for the consumer that overbids, s/he runs the risk of winning

but at a much higher price than his/her true bid. Variants of the second bid auction have

also been popular because they have similar incentive compatibility features (e.g. the

Vickrey fifth- priced sealed bid variant used in this study).

Comparing the Vickrey auction to another incentive compatible method, the Becker

DeGroot Marschak (BDM) mechanism, Lusk et al. (2004), show that the Vickrey auction

is more punitive to untruthful bidding in the case of high value bidders. However, when

low value bidders are involved, the BDM is found to be more punitive. Lusk et al. thus

recommend the application of the Vickrey auction when research interest is in high value

individuals.

Besides incentive compatibility other advantages of using the Vickrey auction

second bid price or variants thereof are noted basically that respondents actually engage

in a transaction and have the opportunity to see the product of interest. This adds to the

validity of research results as it closely mirrors the shopping experience within which

purchasing decisions are made in the real world.

However, there are a few challenges associated with the Vickrey auction and other

experimental auction methods. Menkhaus lists these as disadvantages in his commentary

of 2001. First he mentions that the bidding process in auctions "do not naturally mimic

how consumers reveal preferences in grocery stores" (Menkhaus, 2001, p. 2). Instead of









contending with administered pricing, which is the case in grocery stores, consumers

have to bid their own prices. In addition, there are no competing substitutes in the

experiments, which is the case in grocery stores. Thus the WTP decision-making process

is arguably different.

Despite these assertions, the Vickrey auction has proved to be formidable

especially when "cheap talk" is incorporated in order to frame a grocery store mindset for

the consumers participating in the auction. In such instances, it has shown effectiveness

in revealing WTP. For example, in a Vickrey auction-based study, Hurley et al. (2004)

apply this method when they study 329 consumers in Ames, IA; Manhattan, KS; Raleigh,

NC; Burlington, VT; Iowa Falls, IA and Corvallis, OR and effectively establish WTP in

pork. Researching the WTP for environmental attributes embedded in ten different pork

loin chop packages, they incorporate cheap talk as well as provide some form of

substitutes. They use several related environmental attributes in their experiment, thus

providing consumers with varying levels of environmental friendliness to contend with.

Admittedly there is essentially one product (pork chops) in the experiment and this does

not cover a wide enough spectrum of substitutes as would be the case in grocery stores.

Nevertheless, statistically valid findings show that 62 percent of the consumers are WTP

a premium for the most environmentally friendly pork in the experiment with more than

40 percent WTP a dollar or more. These findings are evaluated and found to be

hypothetical bias-free.

In another Vickrey auction based study, Roosen et al. (1998) measured the WTP

for the removal of insecticide use in apple production. They analyzed a sample of 54

primary shoppers in a Midwestern university town, using nonparametric statistical









analyses and the double hurdle model. Findings revealed that the consumers are willing

to pay $0.34 for apples not treated with Azinphos-methy (APM) pesticide and $0.43 for a

similar bag of apples not treated with any kind of neuroactive insecticides (NAI). The

double hurdle analysis reveals that concern about pesticide use in food production is a

variable with a positive effect on the probability to bid greater than $0.00 for the apples

with reduced pesticide use. The same is found for household income. Concern about food

prices is found to be insignificant in influencing the probability to purchase the reduced

pesticide use apples. Roosen et al. also establish that the presence of children below the

age of five in a household has mixed effects. If a household has a child below the age of

five, then the probability of being willing to pay more than $0.00 is reduced. However,

for those households that are willing to pay, they will pay a higher premium if they have

children below the age of five.

Overall the findings add to the growing list of studies that use the Vickery auction

(second bid price or variants thereof) to successfully elicit consumer WTP for different

product attributes without hypothetical bias. This study follows suit and uses the fifth-

price sealed bid variant given that there was a large number of participants per group that

was surveyed in each auction site. This ensures incentive compatibility even with the

relatively large groups (as will be explained in Chapter 3).














CHAPTER 3
RESEARCH METHODS AND DATA

This chapter documents how the data set was collected and gives a description of

the data.

Study Design and Data

To estimate WTP for COOL in apples and tomatoes a sample of U.S. consumers

was recruited for each product. Data were collected in December 2003 and January 2004

using a Vickrey experimental auction (fifth-priced sealed-bid) followed by a written

questionnaire. The Vickrey auction solicited data on WTP premiums then the written

questionnaire solicited data on numerous variables including demographics, food safety

concerns, food quality concerns and food preferences. Data were collected in Gainesville,

FL, Atlanta, GA and Lansing, MI. Respondents were randomly recruited through local

civic organizations, ranging from faith based organizations to Parents Teachers

Associations (PTA) at schools. The survey was conducted in each respective

organization's facilities and compensation for use of these facilities was made. In total

335 primary shoppers were sampled and the chart below shows the breakdown of these

across locations.










Sample Distribution Across Locations

n i


Lansing, MI Gainesville, FL Atlanta, GA

Figure 3-1. Sample distribution across locations

Of the 335 observations sampled, 311 were useable for analysis; 175 from the

tomato auctions and 136 from the apple auctions. The 24 observations deleted were

unusable due to missing data. The table below shows how the data distributed between

the two products vis-a-vis location.

Table 3-1. Data distribution across location
Product Location Observations Observations Percent Cumulative
Deleted Retained Percent

TOMATOES GNV, FL 3 67 38.3 38.3

LAN, MI 2 49 28.0 66.3

ATL, GA 9 59 33.7 100.0

Total (n) 14 175 100.0

APPLES GNV, FL 1 56 41.2 41.2

LAN, MI 4 17 12.5 53.7

ATL, GA 5 63 46.3 100

Total (n) 10 136 100.0









In general, fewer respondents were recruited in Lansing, MI, making the data

unevenly distributed across locations, particularly in the case of the apple data. The

useable data were such that the respondents' ages ranged from 25 to 65 years and

included primary shoppers only. Here, a primary shopper was defined as an individual

responsible for at least 50 percent of food purchases in the household.

Comparing the sample data to the U.S. population census revealed several

disparities. For instance, a larger proportion of the sample was female, as shown in

Figure 3-2 below. This was nonetheless expected considering that most primary shoppers

are female and that this was the target population. Research protocol had specifically

asked for primary shoppers only.


Comparison of Gender Between Survey data and
U.S. Census Data


100.0%

80.0%

60.0%
m Male
40.0% m Female

20.0%

0.0%
Apples data Tomatoes data U.S. Census

Figure 3-2. Comparison of gender between survey data and U.S. Census data

For the apple data, up to 90 percent of the respondents were female and 86

percent in the case of tomatoes data. This is notably different from 50.9 percent females

in the U.S. census data. In addition, a few other demographics did not mirror those of the









national census data. These observations call for caution in the interpretation of results

from this study, as extrapolative generalizations based on the study may be erroneous.

The table below shows more details on the comparability of the survey data to the

national census data.

Table 3-2. Demographic profile of respondents
Category U.S. Census Sample Average (%)
Average (%)
Apples Tomatoes

Age

25-34 27 11.0 9.1

35-44 31 39.0 48.9

45-54 26 36.8 34.6

55-65 16 13.2 7.4

Race

White 75 84.6 90.9

Black or African U.S. 12 7.4 4.6

Asian 4 2.2 1.7

Other 9 5.8 2.8

Ethnicity

Hispanic 12 6.6 2.3

Income

<$15,000 15.2 3.0 1.7

$15,000 to $24,999 13.2 5.1 4.5

$25,000 to $34,999 12.3 7.4 8.0

$35,000 to $49,999 15.1 11.0 13.1

$50,000 to $74,999 18.3 19.1 27.8









Table 3-2. Continued.
Category U.S. Census Sample Average (%)
Average (%)
Apples Tomatoes

$75,000 to $99,000 11.0 22.0 13.6

$100,000 or above 14.1 32.4 31.3

Education

Bachelor's Degree or 24 64.7 67.0
Higher

Some College 27 28.6 23.9

High School Diploma (or 29 5.9 8.5
equivalent)

Less than High School 20 0.8 0.6


Major disparities are found in the ethnicity, the highest level of education attained

and the pretax household income variables. In proportional terms, fewer Hispanics

participated in the survey (6.6 percent and 2.3 percent in the apples and tomatoes data

respectively, as compared to 12 percent in the 2000 census). The greatest disparity

however, showed up in the highest level of education attained where the majority of the

sample had attained a much higher level of education than the general U.S. population.

More than 60 percent of the respondents in both the apples and tomatoes data had

attained at least a Bachelors degree. Moreover, the survey captured household primary

shoppers that were more affluent relative to the census data. Nearly a third of the

respondents had a pretax household annual income greater than $100,000 as compared to

only 14 percent for the census data (U.S. Census Bureau, 2000). Negligible disparities are

also noted in terms of age and race with more respondents falling in the 35 to 54 age

ranges and more Caucasians being sampled. Even though the data used in the study were









somewhat deviant from the U.S. population they were useable because the deviation was

initially expected, since the target population of the study was U.S. primary shoppers and

not the general U.S. population. The demographic profile of U.S. primary shoppers does

differ slightly from the census data profile. Also, some similarities between the data and

the U.S. census data do exist.

Auction Bid Data Collection

The following section describes the manner in which the auction bid data were

collected. Respondents participated in two auctions which were fashioned as random

fifth-price sealed-bid (Vickrey) auctions. The first auction was a preliminary run used

merely to familiarize the respondents to the auction bidding process. Here the

respondents bid for a large Snickers candy bar.

Candy Bar Auction

Initially, respondents were assigned to a seat with a corresponding I.D. numbered

envelope containing unfilled bid sheets, questionnaire, consent form and instructions.

Each respondent was asked not to communicate with others participating in the survey,

thus all of the consumers' bids and questionnaire responses were made independently.

The respondents were then given $10 in cash for their voluntary participation in the

survey. This became entirely theirs and available for use if they so chose, in the event that

they won an auction. They were also endowed with a small Snickers candy bar and asked

not to consume it. Each respondent voluntarily signed the consent form, which was then

collected by one of the survey monitors. Following this, instructions on the whole auction

procedure were read aloud to the respondents by the lead monitor, including a "cheap

talk" script before the bidding process began (see Appendix A for the detailed

instructions which include the "cheap talk" script). The purpose of reading aloud the









"cheap talk" script to the respondents was to further defray any potential of hypothetical

bias. According to List (2001) incorporating a "cheap talk" script in an experimental

auction can effectively improve the quality of data in terms of the accuracy of WTP

measures elicited. Respondents were asked to raise their hand if they had any questions

after instructions were read out to them. Furthermore, they were told they could raise

their hand to ask questions for clarification at any point in time.

In the first auction, respondents bid the amount they were willing to exchange their

small Snickers bar for a larger Snickers bar. All auction bids in the survey elicited WTP

premiums for exchange. A total of five bidding rounds were conducted for the Snickers

candy bar auction, giving the respondents ample bidding practice and familiarity with the

process. At the end of the auction a binding round was randomly selected and the winners

of that round (i.e. the four highest bidders) had to pay the 5th highest price of the

randomly selected round in order to exchange their small Snickers candy bar for a larger

Snickers candy bar. Since the size of groups participating in the auctions was large

(approximately 25 consumers in each group) the 5th price auction was appropriately

chosen instead of the commonly used 2nd price. This better engaged low-end bidders who

could have been disengaged by a 2nd price, if the high-end bidders' bid too high for the

low-end bidders. Disengaging low end bidders would have resulted in a less

representative estimate of the mean WTP. Using the 5th price ensured incentive

compatibility, which encouraged truthful bidding, and in turn better revealed the

consumers' demand. In addition, more than one top-bidder could be a winner (i.e. the top

four bidders could each win). This too, ensured incentive compatibility and maintained

the truth-revealing trait of the Vickrey auction by eliminating the chance of artificial









scarcity, which usually arises if only one bidder can win. Similarly, the random selection

of the winning bid and winning round ensured truthful bidding by eliminating wealth

effects that could arise if a bidder wins an earlier round. After the winning bids were

established and exchanges done with the winners, the respondents were told they were

free to consume their candy.

Apple and Tomato Auctions

Since there were two products of main interest (i.e. fresh apples and tomatoes),

respondents were either exposed to the apple-auction procedures or tomato-auction

procedures after completing the candy bar auction. Exactly which one the respondents

were exposed to depended on the survey site. For each location (e.g. Atlanta), there had

been several survey sites (i.e. civic organizations) where the auctions were conducted. At

each site random selection was done to establish whether the apple auction or the tomato

auction procedures would be administered. Only in a couple of instances where the sites

surveyed had a large number of voluntary respondents was there need to randomly

divided respondents into two separate groups. In such case each randomly created group

was ushered into a different room right at the onset and thus later exposed to either the

apple or the tomato auction procedures after their candy auction was over. The different

auction procedures for the second auction had been so designed in order to create two

samples (one for each product; see Appendix A for details).

The manner in which the apple (tomato) auction was conducted was similar to the

candy bar auction. Respondents were first endowed with a pound of unlabeled gala

apples (off-the-vine tomatoes). Again, as in the candy bar auction, they were not allowed

to immediately consume their apples (tomatoes) as they would have the opportunity to

exchange them for identical apples (tomatoes) with the label "U.S.A. Grown" (i.e. if their









bid won the auction). Considerable efforts were made to ensure that the unlabeled apples

(tomatoes) bestowed to the respondents had identical visible attributes with those labeled

"U.S.A. Grown." Visible attributes here refers to the size, color, blemishes and variety of

the product. Thus the only visible difference was labeling. Prior to bidding, respondents

were given the opportunity to visually inspect the apples (tomatoes) labeled "U.S.A.

Grown" allowing them to compare them with their own unlabeled ones. Thereafter, they

were asked to make their private sealed bid. In contrast to five rounds of bidding in the

candy bar auction, four rounds of bidding were conducted in the apples (tomatoes)

auction.

After each round, survey monitors would collect the sealed bids, sort and rank them

then post the top five bids in rank order on a board in the front of the room, thus making

the bids visible to all participants. This was done to give bid price information feedback

to the bidders. At the end of all four rounds, one round was randomly selected as

binding, thus establishing the price of exchange and the winning bidders. Once the whole

auction was over and exchanges with winners were done, respondents were told they

were free to consume their apples (tomatoes).

Written Questionnaire

After participating in either the apple or tomato auction, the respondents were

asked to complete a questionnaire which solicited information about their buying

behavior and stated preferences for fresh produce and labeling. The questionnaire took an

average of fifteen minutes to complete, with most questions being based on a nine point

Likert-scale. These questions were designed to elicit the levels of attribute-related

preferences for each respondent. The detailed questionnaire is shown in Appendix B. Not

all questions in the questionnaire were used in this study. Only those pertaining to the









variables that were included in the double hurdle model (later elaborated in the analysis

section of this thesis) were used. A 100 percent response rate for the questionnaire was

achieved since all volunteer participants gave responses and none declined to participate

after coming to the survey site. However, not all of the respondents answered every

question thus giving a completion rate of just over 80 percent. Most of the respondents

that did not complete the questionnaire did not respond to the question on their pre-tax

household income level. Generally consumers may be reluctant to divulge their incomes

as this is a somewhat personal question. Nevertheless, most questions were answered and

the auctions data collection went smoothly in all survey sites to provide the necessary

data analyzed in this study.














CHAPTER 4
THEORETICAL FRAMEWORK AND ANALYTIC TOOLS

The Psychology and Theory of WTP Decisions

Consumer cognitive theory in the context of neoclassical consumer theory forms

the theoretical framework for this study. Consumer cognitive theory proposes that

consumer behavior in food markets is invariably affected by a multitude of factors, which

can be viewed as either internal or external to the consumer (Hansen, 1972). Those

factors considered to be external are stimuli derived from the product itself, as well as

information gained by the consumer about the product and various other environmental

sources that may influence the decision to consume. Theoretically, these are all

measurable by the researcher and are exogenous to the WTP modeling framework.

Practically, not all factors are measurable, at least not directly and some factors may be

endogenous.

Similarly, not all internal factors are directly measurable. The internal factors

include demographics, tastes and preferences (both congenital and learned), perceptions

about the product and intrinsic randomness owing to asymmetric processing of

information by the consumer. Several authors (e.g. Kreps, 1988; Solomon, 2004)

consider the randomness factor to be non-existent as per the random utility model.

According to this school of thought, the individual consumer is assumed to be rational

and capable of perfect discrimination. Thus, the consumers' behavior is inherently

deterministic and any randomness factor in the WTP modeling framework can only result

from the researcher's failure to accurately capture or measure the explanatory variables









influencing the consumer's decision making phenomenon. This view has wide

acceptance, especially with econometricians and is adopted in this study.

Also of theoretical interest is the distinction between the decision to purchase and

the decision on how much to consume (commonly referred to as the participation and

consumption decisions, respectively). Cragg (1971) makes the case that these are two

separate decisions made by the consumer when buying a product. The first is whether to

buy or not to buy the product and the second is how much to buy if indeed the consumer

chooses to buy. Both decisions may or may not be determined by the same explanatory

variables. This forms the basis for using double-hurdle econometric models to explain

consumers' WTP and consumer demand (e.g. Maynard et al., 2003; Dong and Gould,

1999; Gao et al., 1995).

Another important aspect of theory, which lends itself to the work of Lancaster

(1966), is that pertaining to the trichotomy of product attributes, namely search,

experience and credence attributes. As Lancaster articulated, consumers demand

attributes that provide utility to them subject to economic constraints. They do not

demand the products in and of themselves. Following from this theoretic assertion,

Nelson (1970) and Darby and Karni (1973) developed the trichotomy as a means of

classifying product attributes, which are considered to be the focal points of demand.

Criterion used to classify attributes was the pre-cost and post-cost associated with quality

detection for each attribute (i.e. the economic cost incurred by consumers when

identifying the level of quality of the product attribute and the economic cost incurred by

consumers after detecting the quality.)









Search attributes were defined as those with a low pre-cost of detection and zero

post-cost such that consumers would be willing to shop around (hence the name

"search") in order to find the best quality. A key distinction of the search attribute was

that it was observable and measurable by the consumer prior to purchase, e.g. product

color. Thus, no economic cost can be incurred after detection. Andersen and Philipsen

(1998) elaborate on this when they note that the trichotomy is dependent on the level of

uncertainty that consumers will face when provided information about the attribute. In

the case of search attributes, there is zero uncertainty as the attribute is observable and

measurable by the consumer.

In contrast, experience attributes cannot be deciphered by the consumer prior to

purchase and consumption (e.g. taste). For this reason, there is a high level of uncertainty

(and/or pre-cost); however, once the participation and consumption decisions are made,

the level of quality of the attribute is detected. The consumer can then choose if they

want to engage in repeated buys or not. If the quality was low, it is likely that repeated

buys will not occur and the high economic cost of continued consumption of a low

quality product is averted. Nevertheless, a post-cost and uncertainty still exists as there is

a possibility that additional units of the same product may have higher quality which

could yield greater marginal utility.

In the case of credence, both the pre-cost and post-cost are high because at any

stage, the consumer is unable to measure or observe the level of quality. An example of a

credence attribute is country of origin. As Andersen and Philipsen assert, the consumer

must rely on a third party for detection purposes, e.g. a product certifying body. Credence

is closely linked to beliefs and trust. For instance, the consumer can only believe and trust









that a tomato s/he bought was produced in the U.S. There is no way that s/he can test this

and in most cases a third party such as the USDA can act as a detection arbiter. In such a

case, the consumer's level of trust in the USDA then has an effect on the participation

and consumption decisions.

Understanding the trichotomy described above is important to this study because

the focus of this study is on a credence attribute (COOL). Model specification draws

from this understanding as factors that affect consumer WTP for these attributes will

differ due to the different nature of each type of attribute.

Theoretical Framework

Drawing from the aforementioned theories, this study assumes that the individual

consumer can attain utility from a specific product attribute, in this case COOL in apples

or tomatoes ("U.S.A. Grown"). This utility is a function of: (i) consumer characteristics

that influence consumer choice and (ii) the cost that the consumer will pay in order to

obtain the attribute. Thus,

U =U(7, -c,)2 0, (4-1)

where r, is a combination of consumer characteristics and c, the cost that the consumer

will pay to obtain the attribute.

Utility gained from the attribute is zero when the consumer is not willing to pay

anything to obtain the attribute otherwise it is greater than zero. The case of disutility is

disregarded (i.e. U < 0 ) because a rational consumer who does not make mistakes is

assumed, whereby the buying decision ultimately must yield positive utility. In addition,

a consistent consumer is assumed, such that it is not possible for him/her to be willing

and unwilling to pay for the attribute at the same time (i.e. the consumer is a perfect









judge of his/her utility function and remains in his/her chosen state of WTP.) This

assumption allows for mathematical consistency, without which the theoretical

framework cannot hold.

The utility function is unobservable and cannot be measured by the researcher;

however, a proxy measure of utility can be estimated by the WTP. Similarly, it is

assumed that not all consumer characteristics are directly observable and quantifiable,

e.g. consumers' perceptions about food quality or consumers' feelings about food

preferences. These are, instead, latent constructs whose phenomena are observed via

other directly quantifiable proxy variables. Thus, the utility function is deconstructed in

similar fashion to Adamowicz et al (1998), with the only difference being that this study

proposes a directly observable deterministic part A,, an indirectly observable

deterministic partp, and a stochastic error term E,. The error term is assumed to be

independent and identically distributed with a mean of zero and a constant variance.

U(.-, c, )= V(A, ,p, c, )+ E, (4-2)

It is postulated that the variance of the indirectly observable p, can be better

estimated by way of a factor analysis of the directly observable and quantifiable

proxies 0, rather than by using an individual observable proxy variable. Mathematically,

p, = VO, (4-3)

where V/ is a vector of factor loadings

Thus the WTP decision can then be framed in likelihood terms as

Pr{PMT > 0}= Pr{V(A,, p,) > -E,} (4-4)









even though p, is unobservable directly. The above forms the basic theoretical

framework for the double hurdle model estimation with factor scores, which is proposed

in this study.

Analytic Tools

In analyzing the consumer WTP for "U.S.A. Grown" labeling, various analytic

tools are used in this study. In order to ascertain the mean WTP, summary statistics are

computed and parametric hypothesis testing is performed to assess significance.

Equivalency testing is also performed to assess price premium equivalency between the

mean price premium consumers are willing to pay for "U.S.A. Grown" labeling in fresh

apples and that in fresh tomatoes. As earlier alluded to, the double hurdle probit model is

applied to estimate the effects of various factors on consumer WTP. In the process, factor

analysis is performed on a set of variables that, in general, concern food safety, food

quality and food preference subject areas, to derive factor scores that are incorporated as

independent variables in the model estimations. The following sections present these

analytic tools.

The Double Hurdle Model

The double hurdle model is a product of work by Cragg (1971) after the realization

that the Tobit model originally developed by Tobin (1958) was inadequate for a complete

analysis of censored or truncated data. Cragg noted that the Tobit model was restrictive

as it used the same explanatory variables to estimate the dichotomous choice variable and

the quantitative extent of choice variable (i.e. the participation and consumption variables

respectively). In contrast the double hurdle model that he proposed segregated the two,

making it possible for each one to have different explanatory variables and error terms.









By assuming a random utility function that explains the latent dependent utility

variable expressed firstly as a likelihood function of the WTP and then as the actual WTP

amount, this double hurdle model is adapted here to analyze WTP for COOL.

Thus,

WTP, = ZIS + u, (4-5)

for the participation equation denoting the dichotomous willing to pay or not willing to

pay part of the framework.

Then,

PMT* = Xy + E, (4-6)

for the quantitative consumption part of the framework.

In (4-5) the variable WTP* is the consumer willingness to pay assuming 0 if not

and 1 if willing to pay. This dependent variable represents the underlying utility

associated with the participation decision; essentially whether or not the consumer

derives utility from the attribute. In (4-6) PMT* is the actual premium that consumers are

willing to pay for the apples or tomatoes with COOL, if in (4-5) WTP* was equal to 1.

This represents the magnitude of the latent utility associated with the COOL attribute.

Z, and X, are predictor vectors while / and y are parameter vectors to be

estimated for the respective predictor vectors. Zj and Xj can potentially be identical and

include reduced variables in the form of factor scores derived from factor analysis. If Z1

and X, are equal and / and y are also equal then the tobit model results instead of a









truncated tobit. u, and E, are random error terms, normally independently distributed.

u, NID(0,1) and E, ~ NID(0, c2).

Theoretically the underlying utility which is non-measurable can also be expressed

as U,* = f(X Z,") where U'* is the individual consumer's utility. Equations (4-5) and (4-

6) are estimated separately with the (4-5) being estimated first because its results are used

in the estimation of the second (i.e. in estimation of the censoring rule). A probit model

can be estimated for the first equation using the maximum likelihood function:

Pr(PMT, = O I Z X) = O(-X / y / o) + O(X,*y / c)(-Z, ) (4-7)



Then the second equation can be estimated using,

(PMTexp{-(PMT *X2)2 /U212}(Z*,8)
f(PMT, Z,X ,PMT > 0)= ex 2 /2 (4-8)


Where D signifies the standard normal cumulative density function.

Factor Analysis (Principle Component Analysis)

Factor analysis is a data reduction tool whose main objective is to define an

underlying/latent structure of data and reveal interrelations between correlated variables

(Hair et al., 2003). It can be used as an intermediate analytic tool to generate a reduced

form of data for further analysis. Essentially there are two ways that results of a factor

analysis can be used in subsequent data analyses. The first is by analyzing the factor

matrix to select a surrogate variable, whereby that variable with the highest factor loading

is chosen to represent the whole set of variables in the analysis. This method is less

commonly used as it involves omission of those variables found to have lower loadings

but potentially pertinent to the subsequent analyses; hence it is not used in this study.









The other scenario involves generation of an entirely new yet parsimonious set of

variables created from summated scales or factor scores. In the case of summated scales,

a single composite measure is created when several original variables with high loadings

are combined to come up with either the total or (more commonly used) the average

score of the variables for use as a replacement variable. In contrast, when using factor

scores a composite measure incorporating the variance of all variables under analysis is

computed for every case/respondent. This composite measure portrays the extent to

which each case's responses are correlated to the factors derived in the factor analysis.

For each factor generated, a factor score will be generated. Basically, factor scores differ

from summated scales in that they are generated for all variables not just those with high

factor loadings. Furthermore, factor scores can be orthogonally generated thus doing

away with multicollinearity problems that may arise in a subsequent regression analysis

or variants of a regression analysis.

There are two classes of factor analysis, namely exploratory factor analysis and

confirmatory factor analysis (Thompson, 2004). In the case of exploratory factor analysis,

the researcher simply performs the analysis to unveil structure among variables and take

the results as generated. The researcher does not start off with any preconceptions about

the number of underlying factors or their nature. In addition the researcher can use the

results in subsequent analyses. Conversely, confirmatory factor analysis begins with a

preconceived notion by the researcher about the structure of the data, with this being

founded on theory or previous research. This a priori notion sets limits to the analysis in

terms of component estimation or the number of factors to be extracted. Moreover the

analysis adopts the arrangement of a structural equation model (SEM), incorporating









more complex systems of correlations, which are based on prior information (Hair, et al.

2003). Given limited research on the latent constructs of most variables in market

research and the tautological complexities of SEMs, the majority of factor analyses tend

to follow the exploratory perspective. This is the same perspective taken by this study.

Prior to performing a factor analysis, the researcher must perform a correlation

analysis of the variables. Usually, this entails inspection of the Pearson correlation matrix

together with the respective significance levels of these correlations. If correlation

coefficients are greater than +0.6 and significant at 0.01, then this is regarded

meritorious. However, given the large number of variables that are often included in

factor analyses, correlations of this magnitude are rare; hence lower levels of correlations

are considered to be acceptable if the Kaizer-Meyer-Olkin (KMO) measure of sampling

adequacy and the Bartlett's test of sphericity are satisfactory.

According to Hair et al. (2003), the KMO measure of sampling adequacy assesses

the factorability of the variables by quantifying the level of inter-correlations among the

variables. It tests if each partial correlation is small and if it is significant, to give a

summarized index of the individual correlation coefficients. The KMO measure ranges

from 0 to 1; with 0 signifying absolute lack of correlation (i.e. inappropriate for factor

analysis) and 1 signifying perfect correlation. The Bartlett test of sphericity is somewhat

similar, providing the statistical probability that the correlation matrix, as a whole, is

correlated to the individual variables. Essentially, the Bartlett test assesses if the

correlation matrix is an identity matrix, which if so would make factor analysis

inappropriate. These measures are used to determine if factor analysis is appropriate in

this study.









Factor scores are derived and the scree plot decision rule is employed to determine

how many factors to extract, because the variables used are less than 20. In addition, as a

cautionary measure, the acceptable extraction rate is set at approximately 70 percent or

greater. Usually, the latent root criterion would be used to determine how many factors to

extract, with each factor having a latent root/eigenvalue of greater than 1. The rationale

behind this is that each variable contributes a value of 1 to the total eigenvalue, thus each

factor would have to contribute more variance than a single variable if it is to be

considered significant. However, this is most applicable to factor analyses involving 20 to

50 variables (Hair et al., 2003). If less than 20 variables are involved, then this decision

rule is likely to extract too few factors; hence it is not used in this study.

In addition, Varimax rotations are used in this study to facilitate naming of the

derived factors. Fundamentally, a factor rotation is an adjustment of the factor axes,

which results in the transformation of the derived factors through the redistribution of the

variance extracted. The total variance extracted is maintained and does not change, and

only the position of each factor in the rotated plane is shifted; thereby making it easier to

interpret associations and name the factors. Different forms of factor rotations can be

performed, but the Varimax is widely used because it is a 90 degree rotation that leads to

orthogonal factors. This eliminates chances of multicollinearity without complicating the

interpretation of each factor. Loadings tend towards 0 or 1, with 0 signifying a lack of

association and 1 signifying high degrees of association. The signs on the loadings only

serve directional purpose, to indicate the nature of the relationship and not the magnitude.

Various other rotations can be performed in factor analysis but these are irrelevant to this






39


study and beyond the scope of this thesis. Only the Varimax rotation is of significance

here.














CHAPTER 5
AUCTION BID ANALYSIS AND RESULTS

The manner in which the collected data were analyzed can be broken into two

categories: i) univariate-parametric statistical analyses of the WTP bids and ii) the

econometric analyses of WTP bids. The latter entailed a factor analysis followed by a

double hurdle probit model analysis. This chapter presents the former (i.e. the univariate

parametric statistical analysis of the auction bids).

Both the apple and tomato auction bids were initially evaluated separately. As

previously alluded to, four rounds of bidding had occurred in each auction. The amount

that each participant bid was the premium s/he was willing to pay in order to exchange

one pound of unlabeled fresh apples (tomatoes) that s/he was initially endowed with, for

one pound of fresh apples (tomatoes) labeled "U.S.A. Grown." The bidding progression

for all rounds in all locations are combined and outlined in the line graph in Figure 5-1.

Also, the analysis of bidding progressions by location was done and results of this are

shown in Appendix H.

In the fresh apples' line graph, in Figure 5-1, the average bid for the first round was

considerably lower than that of all four rounds combined. Otherwise the subsequent bids

converged at the approximate average of $0.48. With respect to tomatoes, the last bid is

the one that was relatively deviant. The first three rounds of bidding had a mean premium

of approximately $0.44 while the last average bid jumped to $0.53. This invariably raised

the mean bid, making the average for all four bids to be $0.46. A look at Appendix H

reveals that the unprecedented deviation may have been due to a spike recorded in the









fourth bid in the Gainesville, FL data. Nevertheless, increases from the third to the fourth

bid were also recorded in the other locations.


Progression of Mean bids-All Locations

0.55

0.50


0.45 --+ Tomatoes (n=175)
$/Lb
0.40 -- Apples (m=136)

0.35

0.30
Round Round2 Round3 Round4
Round of bidding

Figure 5-1. Line graph showing the progression of combined bids in both tomato and
apple auctions

The mean of the four bids as well as the mean of the last two bids (i.e. PMTi, which

is the variable used as the quantitative WTP variable in subsequent econometric analyses)

are shown in Table 5-1. The mean of the last two bids (PMTi) is used as the dependent

variable in the econometric analysis instead of the mean of all four bids because the last

two bids are considered to be better estimates of the WTP. This is because in the first few

rounds of bidding, consumers learn the nature of their WTP for the product attribute at

auction as value formation takes place. Thus, bids often change substantially from one

round to the next before stabilizing at the true WTP value after the first few rounds have

passed. According to Cox et al. (1985) and Fox et al. in Caswell (1995) this is why the

Vickrey auction with multiple rounds is better suited to obtain true WTP because it

accommodates for value formation. Shogren et al. (2000) add that the first rounds of









bidding in a Vickrey auction may reveal less accurate WTP measures because the value

formation process may take place during these rounds, especially if a new unfamiliar

product is at auction. They refer to this as the lab novelty effect.

Table 5-1. Average WTP for apples (n = 136) and tomatoes (n = 175)
Mean Standard Deviation

All four bids 3ra and 4th round All four bids 3ra and 4th
bids (PMTi) round bids

Apples $0.48 $0.49 $0.55 $0.58

Tomatoes $0.46 $0.48 $0.53 $0.55


As shown in Table 5-1, the mean WTP for COOL in apples and tomatoes is

approximately $0.49 and $0.48 respectively. The standard deviations in both cases are

high, 0.58 and 0.55 respectively. This high level of dispersion shows that different

consumers have distinctly different levels of WTP (i.e., it can be viewed as an indicator

of consumer surplus). Univariate hypothesis testing of the mean WTP (i.e. PMTi) proves

the bids to be significantly greater than zero. Thus:

Ho: PMTi = 0

Ha: PMTi > 0 (One-tailed Test)

0.05 is chosen as the significance level (a = 0.05)

For fresh apples:

t(i- 0)- 0.49-0 9.85
t= = =9.85
se 0.58/_


Therefore, I reject the null hypothesis that the WTP for COOL in fresh apples is

equal to 0 since the t-value yields a probability value < 0.001

Similarly for fresh tomatoes:










t = 0) -0.48 0 =11.58, which also yields a probability value < 0.001.
se 0.5 5/_1


Therefore the null hypothesis that the WTP for COOL in fresh tomatoes equals 0 is

rejected.

On calculating the means for only those consumers who were willing to pay more

than $0.00 for either fresh apples or fresh tomatoes labeled "U.S.A. Grown," the expected

increase was registered as shown in Table 5-2. This calculation was done to give insight

on the existing differentials between the sub-sample of those willing to pay and the whole

sample.

Table 5-2. Average WTP for apples (n = 108) and tomatoes (n = 126): Sampling only
those consumers who were WTP more than $0.00
Mean Standard Deviation

All four bids 3ra and 4th All four bids 3" and 4th
round bids round bids

Apples $0.60 $0.61 $0.56 $0.59

Tomatoes $0.64 $0.68 $0.53 $0.54


Overall, 79 percent of the consumers were willing to pay more than $0.00 for fresh

apples labeled "U.S.A. Grown" while 72 percent were willing to pay a premium in the

case of fresh tomatoes labeled "U.S.A. Grown." Detailed analysis by location revealed

that in general, consumers in Lansing were willing to pay substantially less for the label

"U.S.A. Grown" irrespective of the product under consideration. As shown in table 5-3, it

was only the consumers in Gainesville who seemed to have a high WTP for COOL in

fresh tomatoes, while those in Atlanta had a high WTP for COOL in fresh apples.

Using the findings on the price premium means (PMTi) for fresh apples and

tomatoes, a comparison was made to test if these means were statistically different in









order to begin testing of the second hypothesis in this study. As noted earlier, the

hypothesis is:

2. If consumers are willing to pay a premium for fresh apples and fresh tomatoes
labeled "U.S.A. Grown" then the premiums will be product specific and unequal.


Table 5-3. Mean WTP across location
Gainesville, FL Lansing, MI Atlanta, GA

($/Lb) ($/Lb) ($/Lb)

Apples 0.41 0.18 0.64

Tomatoes 0.78 0.20 0.39



In addition, it was imperative do so because subsequent econometric analyses

entailed combining the two data sets with the premiums as the dependent variable.

Combining the two would only be statistically plausible if the two data sets' PMTi were

independent of the product treatment that the respondents received. In other words, the

premium that consumers were willing to pay could not be a function of whether the

product they bid for was an apple or a tomato. If it was, then the PMTi would be regarded

as two different variables making it implausible to combine them. Thus the z-test for

independent samples was performed:

Ho: PMTapples = PMTtomatoes

Ha: PMTapples PMTtomatoes

Choosing a significance level of 0.05 (a = 0.05)

Criterion: Reject Ho if z < -1.96 or z > 1.96, where









PMT PMT,
PIapples P tomatoes
2 2
apples tomatoes
+
apples tomatoes

0.49 0.48
= 0.154
Calculation: 0.337 0.303 154

136 176

Since z = 0.154 < 1.96 we fail to reject the null hypothesis that the two means are

not different. This analysis of the data suggests premium equivalency in fresh apples and

tomatoes labeled "Grown in the U.S. Nevertheless, due to the inherent logical asymmetry

between the null and alternative hypotheses, we cannot statistically confirm equivalence.

To confirm equivalence, the following two-sample t-test for equivalence is performed:

First, I assume that the distributions of PMTi observations follow the basic

parametric model i.e. they are normal with a common variance and potentially unequal

expected values. Mathematically,

X, N(,a 2)Vi = 1,...,m

Y N(0, a )Vj = l,...,n with 77 e IR, a2 e IR+

where X is the premium bid for apples and Y is that for tomatoes; m is the sample size of

the apple data and n is the sample size of the tomato data, while is the mean PMTi for

apples and 7 that for tomatoes. C2 is the common variance of the PMTi which is a

positive real number.

This is also assumed for both previous hypothesis tests and all other analyses in this

study. I follow Wellek (2003) and define equivalence of the apples "treatment" to the

tomatoes "treatment" by the condition that the difference between the standardized









premium means falls in a sufficiently narrow interval (si, a 2) in the neighborhood of

zero. Thus I then formulate the hypothesis testing problem as

Ho:: <-1 or > !-E2


versus Ha: 1 <(- < s (s1, S2>0) where is the mean PMTi for apples

and q that for tomatoes; a is the standard error of the difference in PMTi.

A critical region, which is synonymous to the confidence interval in the traditional

t-test for independent samples, is then selected based on the researcher's discretion.

Discretion here would depend on how narrow the desired equivalence interval is to be.

Here, I choose an alpha level of 0.05 and calculate the region using the SAS program (see

Appendix D) because this alpha level corresponds to the narrowest interval, compared to

other commonly used critical regions for this kind of hypothesis test. The test statistic is

calculated as


T= m+n-2 -Y




where Xi is the WTP for apples and Yi is that for the tomatoes; m is the sample size of the

apples data (136) and n that for tomatoes (175). Thus we would reject the null hypothesis

of non-equivalence of the premium for apples to the one for tomatoes if the calculated T-

statistic falls within the critical region.

The resulting T-value from the data is 0.045. This signifies premium equivalence

since it falls within the critical region of -2.72 and 7.04. The null hypothesis of non-

equivalence of the premium for apples to the one for tomatoes is rejected. See Appendix

D for detailed results from the SAS program. It is concluded that the mean premium that






47


consumers are willing to pay for apples labeled "U.S.A. Grown" over unlabeled apples is

equivalent to the mean premium they are willing to pay for tomatoes labeled "U.S.A.

Grown" over unlabeled tomatoes. The detailed implication of this finding is discussed in

Chapter 9.















CHAPTER 6
EMPIRICAL SPECIFICATION AND RESULTS

Empirical Specification

To address the last hypothesis of this study, the factors influencing consumers'

WTP for the label "U.S.A. Grown" were analyzed using the double hurdle probit model

specification in (6-1). Additional consideration of demographic variables was also

included in the analysis.

4 6 9
WTP, = fAge + f2Gender + ,Edu, + /0,Loc, + Inc, + lExpose +
=3 1=5 :=7 (6-1)
13 18 20
Y /PC, + P4Trust + ,15Safe + ,Y Pfr, + ,Qual, + u,
=11 z=16 1=19

where WTPi is the dichotomous willingness to pay (i.e. participation dependent variable),

expressed as a probability

For the second hurdle PMTi replaces WTPi, where PMTi is the quantitative

willingness to pay (i.e. consumption dependent variable) and e, takes the place of u, as

the error term.

* Age = Age of respondent

* Gender = Gender of the respondent

* Edu = Highest level of education completed by respondent

* Loc= Location (Atlanta, Gainesville, or Lansing)

* Inc = Income group

* Expose = Self rating on exposure to food safety information in fresh fruit and
vegetables









* PC = Presence of children under age of 16 in the household

* Trust = Extent of respondent's trust in information about food production obtained
from U.S. Government Agencies, (e.g. USDA, FDA, EPA, etc.)

* Safe = Perceptions about food safety

* Pfr = Food preferences factor scores

* Qual = Food quality factor scores

Description of Variables

A detailed description of the variables is provided in Appendix E. Age was

measured in terms of the number of years that the respondent had, while gender was

measured as a dichotomous variable, denoting the respondent's sex. In all of the models

estimated in this study, the male sex was dropped and used as the base1 variable. Females

were expected to be willing to pay more for COOL since they are considered to be more

concerned about details or food safety, which would normally induce a desire to know

the product's origin. Similarly, older consumers were expected to be more willing to pay

for COOL for the same reasons.

The highest level of education completed by the respondent is a variable measuring

how well-educated a respondent is, based on the formal education system. Given the

ordinal nature of the education variable, 3 dummy variables (EDU1, EDU2 and EDU3)

had to be created, while taking into account the distribution of the education variable in

the samples analyzed; (most respondents were relatively well-educated thus necessitating

the creation of fewer dummy variables). The "University postgraduate degree" (EDU3)

dummy variable was dropped and became the base in all model estimations.



1 In all of the models estimated in this thesis, the same base was used; this is the case for all dummy
variables i.e. GENDER, EDU, LOC, INC and PC.









Three dummy variables were also created for the respondents' location/state of

residence given that location is a nominal variable (LOC1, LOC2 and LOC3). In this

case, Atlanta, GA (LOC 3) was used as a base by dropping it out of the estimations. Prior

to model estimation, it was anticipated that consumers from Gainesville, FL would be

more willing to pay for COOL because mandatory COOL has been prevalent in Florida

for the past 26 years, at the state level and it would be expected that Floridian consumers

are accustomed to COOL. In addition, it was expected that consumers from Lansing, MI

would be willing to pay for COOL particularly in the case of apples, since Michigan

apples may compete with apples from other countries. The premise here was that

Michigan consumers would want to support local producers and since COOL would

enable identification of product origin, consumers would then be willing to pay for

COOL in apples.

Income is another ordinal variable, which was included in the model specification.

In this case, 4 dummy variables (INC1... INC4) were created based on the income

distribution of the samples analyzed; most respondents had relatively high incomes. The

highest income bracket (INC4) "$100,000 and above" was dropped, thus making it the

base. Prior to estimation of the model, consumers with higher incomes were thought to be

more willing to pay for COOL, based on the notion that labeled produce would be

regarded as a type of luxury good when compared to unlabeled produce. In addition, high

income consumers were expected to have a higher marginal propensity to spend.

Economic theory shows the high income consumers usually have a higher marginal

propensity to spend, thus it would be more likely to find high income consumers who are

willing to pay for COOL.









The presence of children under the age of 16 years in the respondent's household

was another demographic included in the model. Four dummy variables were created and

the highest "3 or more children" (PC4) was dropped to make it the base variable.

Anticipated results were that consumers with more children would be more concerned

about food safety and the origin of the food they feed to their children. Thus, they would

be more willing to pay a premium for COOL.

Inclusion of all these demographic variables was grounded on economic theory.

Consumer demand theory maintains that consumers make expenditure choices with the

objective of maximizing utility subject to economic constraints. These constraints arise

from scarcity, which inevitably leads to the economic problem of choice, while the

objective utility function stems from endogenous preferences/desires, which may differ

from one individual to the other. Hence, different consumers may obtain different levels

of utility from the same measure of the same product attribute. In other words, the

endogenous preferences determine the unique level of consumer-surplus that each

consumer may attain from a product attribute, if purchased at a given market price;

consumer surplus arises from the difference between WTP and the market's equilibrium

price.

As outlined in the theoretical framework in Chapter 4, demographics influence the

nature of an individual's endogenous preferences, which determine the utility s/he can

obtain from a particular product attribute. This is why, for example, older consumers will

normally demand different product attributes in comparison to younger consumers (e.g.

more health-oriented product attributes) and females will normally demand different

attributes compared to males.









Therefore, demographics (such as age, gender, income, level of education, location

and number of children in the household) were appropriately included as explanatory

variables in the model specification. In addition, the choice of these explanatory variables

followed several previous studies, which have hypothesized them to be major drivers of

WTP for COOL (e.g. Schupp and Gillespie, 2001; Loureiro and Umberger, 2002; Lusk et

al., 2003).

Based on consumer demand theory and consumer cognitive theory, psychographics

were included as explanatory variables in the specified model. Also, previous studies

have included similar psychographics in their analyses of WTP for COOL (e.g. Loureiro

and Umberger, 2002; Umberger et al., 2003). EXPOSE is a psychographic variable that

was included in the model specification, and it is an ordinal Likert-scale rating, indicating

the level of exposure to information on food safety in fruits and vegetables that

respondents had previously received. This variable was included because theory suggests

that the level of information that a consumer has been exposed to, affects his/her WTP

decisions. Perceptions about trust, food safety, food preferences and food quality were

also included as explanatory psychographic variables on the same basis. As alluded to in

Chapter 4, psychographic variables such as these may have a significant bearing on

consumers' WTP.

Trust was measured by a Likert scale rating, indicating how trusting consumers

were of information they receive from U.S. agencies (e.g. USDA, FDA, EPA, etc.) about

food production. Given that COOL is a credence attribute whose verification and

enforcement is in the hands of such agencies, it was necessary to include this variable in









the model specification in order to assess the impact it may have on WTP for a credence

attribute such as COOL.

Consumer food safety concerns in fresh fruits and vegetables were also included,

because COOL can be associated with food safety. The origin of a product will usually

influence whether the product is safe for consumption, because the product's contact with

its environment of origin can contaminate the product with harmful bacteria or viruses.

Based on this, consumers that think about the food safety, when making purchasing

decisions, were expected to be willing to pay more to know the product's country of

origin.

Consumer preferences were also included, and factor scores were generated from

10 psychographic questions related to food preferences, thus creating 3 factors to be

included in the model specification. The 3 quantitative factors on consumers food

preferences (PFR1, PFR2, PFR3) were named "Open to unfamiliar foods," "Choosey"

and "Afraid of unfamiliar foods," respectively (see Chapter 7 for details). The

expectation was that conservative consumers who prefer to consume local produce or are

less adventurous in the kinds of foods they will eat, would be willing to pay a premium

for produce labeled "U.S.A. Grown."

In the case of consumers' perceptions about food quality, 2 factors (QUAL1 and

QUAL2) were generated from 7 psychographic questions related to food quality

perceptions and these were named "General quality conscious" and "Natural quality

conscious," respectively. It was anticipated that consumers who were more quality

conscious would be willing to pay more for the produce labeled "U.S.A. Grown," and the

premise being that U.S. produce would be regarded as better quality produce. Consumers









inclined to naturally produced foods were expected to be willing to pay even more for

COOL, as these consumers are assumed to be relatively more particular about the origin

of the produce they consume.

Organization of Model Estimations

Prior to the estimation of the model in (6-1), a variant thereof was specified and

estimated where the factor scores on food preferences and food quality (Pfr and Qual)

were not incorporated. Instead, single variables on food preferences (DFF) and food

quality (PAYQ) replaced them. Both of these were based on consumer Likert-scale

ratings (nine-point scale) indicating the consumers' level of agreement with the

statements: "I like foods from different countries" and "I am willing to pay somewhat

more for a product of better quality" (refer to questionnaire in Appendix B).

In addition, the model was first estimated separately for apples and then separately

for the tomato data; hence the presentation of the results is organized into the apple and

then tomato sections. A third section is then presented, where the estimation was

performed for the combined apple and tomato data set. Combining the two data sets was

done since the analysis of the bids by consumers in Chapter 5 indicated that the WTP was

not product specific. This made it statistically plausible to combine the two data sets and

analyze them as one data set. The results presented in this chapter are only for models

without factor scores.

Models without Factor Scores

Apple Model (Model 1)

The apple model without factor scores is presented in Tables 6-1 and 6-2 below.

Table 6-1 shows the probit part of the estimation, where an approximate chi-square value

of 6.6 with 16 degrees of freedom yielded a significance level of 0.98. Thus, this model is









insignificant. Though none of the explanatory variables are statistically significant, the

probit model correctly predicted 78.6 percent of the consumers' responses.

Also, despite the insignificance of the probit estimation, the truncated estimation

has been presented in Table 6-2. In this part of the estimation, as well as all other

estimations of the second stages of the double hurdle models in this study, the chi-

squared specification test was performed to evaluate if the truncated tobit estimation was

a better fit than the tobit. The formula used is, Z = 2(ln Lprobit + In Lncaor In -nLob,)

and the truncated tobit proved to be a superior fit (See Appendix G for details).

Age, gender, income and location were all significant at 90 percent confidence

level while the rest of the explanatory variables were insignificant. In terms of age the

marginal effects show that for every year that a consumer is younger, they would be

willing to pay an extra $0.01 for the "U.S.A. Grown" labeling. Females were willing to

pay approximately $0.35 more than the males and consumers in Lansing MI were willing

to pay considerably less than the consumers in Atlanta GA or Gainesville, FL. Regarding

the income levels, only consumers who earned below $50,000 per year were willing to

pay more for the labels, otherwise all other income groups were not willing to pay more

for COOL.














Table 6-1. Apples probit model without factor scores (Model 1)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.016 0.014 0.00 -1.1180 0.2637 45.22
ENDER 0.218 0.439 0.06 0.4980 0.6187 0.90
EDU1 0.030 0.367 0.01 0.0820 0.9347 0.35
EDU2 0.128 0.343 0.03 0.3730 0.7088 0.38
LOC1 -0.109 0.348 -0.03 -0.3130 0.7544 0.41
LOC2 0.484 0.482 0.11 1.0050 0.3149 0.13
INC1 0.196 0.377 0.05 0.5200 0.6031 0.26
INC2 -0.020 0.376 -0.01 -0.0530 0.9574 0.19
INC3 0.167 0.369 0.04 0.4540 0.6500 0.22
EXPOSE -0.049 0.215 -0.01 -0.2300 0.8183 1.90
PC1 0.414 0.424 0.10 0.9760 0.3290 0.26
PC2 0.505 0.462 0.12 1.0920 0.2749 0.15
PC3 0.054 0.339 0.01 0.1600 0.8732 0.37
TRUST 0.073 0.094 0.02 0.7700 0.4411 3.69
SAFE 0.027 0.089 0.01 0.2980 0.7659 3.70
DFF -0.017 0.064 0.00 -0.2680 0.7884 6.90
PAYQ 0.117 0.094 0.03 1.2510 0.2110 7.29
Restricted log likelihood value, In Llo = -69.15 R2 (McFadden, 1973) = .0479
Maximum unrestricted log likelihood value, In L1 = -65.83 R2 (Estrella, 1998)= .0487
Log likelihood 2(df-16)= 6.629 (p = 0.9800) % of correct predictions = 78.6
Number of observations = 136














Table 6-2. Apples truncated tobit model without factor scores (Model 1)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor

AGE -0.0317 0.0133 -0.01 -2.3890 0.0169 44.64
GENDER 1.1218 0.6012 0.35 1.8660 0.0621 0.91
EDU1 0.3950 0.3829 0.12 1.0320 0.3023 0.35
EDU2 0.1335 0.3369 0.04 0.3960 0.6919 0.38
LOC1 -0.2111 0.3447 -0.07 -0.6120 0.5403 0.40
LOC2 -1.9665 0.6501 -0.62 -3.0250 0.0025 0.14
INC1 0.6580 0.3314 0.21 1.9860 0.0470 0.27
INC2 -0.1328 0.3572 -0.04 -0.3720 0.7101 0.19
INC3 -0.2104 0.3681 -0.07 -0.5710 0.5677 0.23
EXPOSE -0.2729 0.2062 -0.09 -1.3240 0.1856 1.90
PC1 0.0042 0.3716 0.00 0.0110 0.9909 0.26
PC2 -0.1079 0.4299 -0.03 -0.2510 0.8019 0.17
PC3 0.3531 0.3325 0.11 1.0620 0.2882 0.36
TRUST 0.0403 0.0869 0.01 0.4640 0.6427 3.74
SAFE 0.1053 0.0912 0.03 1.1550 0.2481 3.72
DFF -0.0271 0.0646 -0.01 -0.4190 0.6752 6.88
PAYQ 0.0461 0.0904 0.01 0.5100 0.6101 7.36
Sigma 0.7348 0.0979 7.5060 0.0000
Number of observations = 136 Log likelihood function= -35.50
Observation after truncation= 108 Threshold values for model: Lower = 0 Upper = +x










Tomatoes Model (Model 2)

In the case of the tomatoes data, the probit estimation without factor scores had a p-

value of 0.11. Only consumers' perceptions about food quality and location turned out to

be significant independent variables. Consumers, who indicated in the questionnaire that

they were willing to pay somewhat more for a product of better quality, actually had a

higher probability of being willing to pay a premium for tomatoes labeled "U.S.A.

Grown," according to auctions data analysis. Also, consumers who were located in

Lansing MI were least likely to be willing to pay a premium for the tomatoes labeled

"U.S.A. Grown" in contrast to consumers in Gainesville, FL who were most likely to pay

a premium. Table 6-3 below presents the probit model estimation.

Table 6-3. Tomatoes probit model without factor scores (Model 2)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0098 0.0147 0.00 -0.671 0.5024 44.21
GENDER -0.1092 0.3260 -0.03 -0.335 0.7376 0.87
EDU1 0.0788 0.2903 0.03 0.272 0.7859 0.33
EDU2 0.4440 0.2824 0.14 1.572 0.1159 0.43
LOC1 0.5479 0.2866 0.17 1.912 0.0559 0.38
LOC2 -0.2992 0.2775 -0.10 -1.078 0.2809 0.28
INC1 -0.1154 0.3048 -0.04 -0.379 0.7049 0.27
INC2 -0.2644 0.2994 -0.09 -0.883 0.3771 0.28
INC3 -0.0809 0.3675 -0.03 -0.22 0.8257 0.14
EXPOSE 0.0856 0.1420 0.03 0.603 0.5467 1.91
PC1 -0.3590 0.4147 -0.12 -0.866 0.3867 0.23
PC2 -0.2022 0.3841 -0.07 -0.526 0.5985 0.22
PC3 -0.2482 0.3175 -0.08 -0.782 0.4343 0.37
TRUST 0.0803 0.0697 0.03 1.152 0.2495 3.71
SAFE 0.0282 0.0698 0.01 0.404 0.6859 3.41
DFF -0.0747 0.0559 -0.02 -1.336 0.1815 6.99
PAYQ 0.1534 0.0847 0.05 1.812 0.0700 7.32
Restricted log likelihood value, In Lo = -104.70 R2 (McFadden, 1973) = 0.1111
Maximum unrestricted log likelihood value, In L1 = -93.06 R2 (Estrella, 1998)= 0.1315
Log likelihood 2(df-16)= 23.28 (p = 0.1064) % of correct predictions = 74.2
Number of observations = 175










As seen above, the model had an R2 (McFadden, 1973) of 0.11 which is considered

feeble for cross sectional data such as that used in this analysis. Nevertheless, the model

had a 74% correct prediction rate.

In the truncated part of the model, the consumers' food quality perceptions

variable was insignificant while some demographic variables turned out to be significant.

Table 6-4 below has more details on this. Those consumers whose highest level of

education completed was some college or less (EDU1) were willing to pay $0.20 less

than the base (Postgraduate degree). Also, consumers in Lansing MI were once again

shown to be willing to pay less for the labeling (this time $0.27 less). Those consumers,

who had indicated that they have seen, read or heard less about food safety in fresh fruits

and vegetables were willing to pay $0.09 less for COOL. Consumers who indicated a

high level of trust for U.S. government agencies such as the USDA, FDA, EPA etc.

showed a greater willingness to pay for "Grown in the U.S" fresh tomatoes.

Table 6-4. Tomatoes truncated tobit model without factor scores (Model 2)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0005 0.0111 0.00 -0.041 0.9674 44.26
GENDER 0.2288 0.2918 0.10 0.784 0.4330 0.86
EDU1 -0.4375 0.2336 -0.20 -1.873 0.0611 0.32
EDU2 -0.2519 0.2147 -0.11 -1.173 0.2407 0.46
LOC1 0.5136 0.2220 0.23 2.314 0.0207 0.45
LOC2 -0.6104 0.3421 -0.27 -1.784 0.0744 0.22
INC1 0.2859 0.2492 0.13 1.147 0.2513 0.27
INC2 0.3649 0.2514 0.16 1.451 0.1467 0.27
INC3 0.2968 0.2902 0.13 1.023 0.3064 0.14
EXPOSE -0.2085 0.1266 -0.09 -1.647 0.0996 1.94
PC1 -0.3528 0.3244 -0.16 -1.087 0.2769 0.22
PC2 -0.1270 0.2805 -0.06 -0.453 0.6507 0.22
PC3 -0.2249 0.2537 -0.10 -0.887 0.3753 0.36
TRUST 0.1407 0.0646 0.06 2.176 0.0296 3.86
SAFE 0.0277 0.0558 0.01 0.495 0.6203 3.46
DFF -0.0209 0.0560 -0.01 -0.373 0.7094 6.99
PAYQ 0.0124 0.0768 0.01 0.162 0.8717 7.49
Sigma 0.6382 0.0791 8.064 0.0000
Number of observations = 175 Log likelihood function= -49.70
Observation after truncation= 125 Threshold values for model: Lower = 0 Upper = +o











Combined Apples and Tomatoes Model (Model 3)

When the data were combined into one data set and the same model estimated,

similar results were obtained. The probit part of the estimation was found to be

significant with a log likelihood chi-square value of 24.11 at 16 degrees of freedom,

which corresponds to a p-value of 0.087. Food quality concerns and age turned out to be

the only variables that were significant, at a 0.1 alpha level. Marginal probabilities show

that for every year older that the consumer is the probability of being willing to pay a

premium for COOL is decreased by 0.6%. This is holding all dummy variables at the

base; the base being a female with a university postgraduate degree education, household

income of greater than $150,000, with three or more children in the household and

located in Atlanta GA.

Table 6-5. Combined apples and tomatoes probit model without factor scores (Model 3)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0171 0.0098 -0.01 -1.739 0.082 44.65
GENDER -0.0666 0.2476 -0.02 -0.269 0.7878 0.88
EDU1 0.1142 0.2195 0.04 0.52 0.603 0.34
EDU2 0.2848 0.2122 0.10 1.342 0.1796 0.41
LOC1 0.3288 0.2130 0.11 1.544 0.1227 0.40
LOC2 -0.2251 0.2184 -0.08 -1.03 0.3028 0.21
INC1 0.0270 0.2292 0.01 0.118 0.9061 0.27
INC2 -0.1705 0.2269 -0.06 -0.752 0.4522 0.24
INC3 0.0211 0.2520 0.01 0.084 0.9334 0.17
EXPOSE 0.0631 0.1149 0.02 0.549 0.583 1.89
PC1 -0.0303 0.2814 -0.01 -0.108 0.9144 0.24
PC2 0.0165 0.2735 0.01 0.06 0.952 0.19
PC3 -0.1219 0.2236 -0.04 -0.545 0.5856 0.37
TRUST 0.0656 0.0532 0.02 1.232 0.2179 3.53
SAFE 0.0514 0.0530 0.02 0.97 0.3318 3.71
DFF -0.0354 0.0404 -0.01 -0.875 0.3815 6.95
PAYQ 0.1488 0.0592 0.05 2.512 0.012 7.31
Restricted log likelihood value, InLlo = -175.16 R2 (McFadden, 1973) = .06881
Maximum unrestricted log likelihood value, In L, -163.1073 R2 (Estrella, 1998)= .07717
Log likelihood 2(df=16)= 24.10601 (p = .8721459E-01) % of correct predictions = 75.2
Number of observations = 311










The truncated tobit part of the analysis is presented in the Table 6-6. According to

this model, it is noted that location, level of exposure to information about food safety in

fruits and vegetables as well food safety concerns are significant variables at a 0.1

significance level. Consumers in Lansing MI were once more shown to be less willing to

pay for COOL with marginal effects at the means showing that they were willing to pay

$0.49 less than what the consumers in Atlanta GA were willing to pay. Those consumers

who rated themselves as having seen or heard or read less about food safety in fresh fruits

and vegetables were also found to be willing to pay marginally less for fresh tomatoes

labeled "U.S.A. Grown." In contrast, those who said they thought about food safety when

purchasing fresh fruits and vegetables were found to be willing to pay more for produce

labeled "U.S.A. Grown." This may imply that these consumers consider U.S. produce

safer hence they were willing to pay more for produce with COOL.

Table 6-6. Combined apples and tomatoes truncated tobit model without factor scores
(Model 3
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0192 0.0123 -0.01 -1.558 0.1193 44.43
GENDER 0.6725 0.4091 0.20 1.644 0.1002 0.88
EDU1 -0.0747 0.2624 -0.02 -0.285 0.7759 0.33
EDU2 -0.0975 0.2540 -0.03 -0.384 0.7010 0.42
LOC1 0.1938 0.2331 0.06 0.832 0.4057 0.42
LOC2 -1.6638 0.5437 -0.49 -3.060 0.0022 0.18
INC1 0.4257 0.2763 0.12 1.541 0.1234 0.27
INC2 0.2584 0.2800 0.08 0.923 0.3561 0.23
INC3 -0.0163 0.3150 0.00 -0.052 0.9586 0.18
EXPOSE -0.2727 0.1542 -0.08 -1.769 0.0769 1.92
PC1 -0.0703 0.3274 -0.02 -0.215 0.8300 0.24
PC2 -0.0024 0.3145 0.00 -0.008 0.9939 0.20
PC3 0.1562 0.2675 0.04 0.584 0.5593 0.36
TRUST 0.0265 0.0646 0.01 0.410 0.6821 3.59
SAFE 0.1814 0.0779 0.05 2.330 0.0198 3.79
DFF -0.0640 0.0585 -0.02 -1.093 0.2742 6.94
PAYQ 0.0280 0.0802 0.01 0.349 0.7271 7.43
Sigma 0.8516 0.1063 8.014 0.0000
Number of observations = 311 Log likelihood function = -102.78
Observation after truncation = 233 Threshold values for model: Lower = 0 Upper = +co






62


Overall, the models without factor scores all suggested an association between

WTP for produce labeled "Grown in the U.S. and some of the demographic variables as

well as the quality and food safety related variables. In all models without factor scores

food preferences and presence of children in a household were insignificant implying that

these have no effect on the WTP for COOL in fresh apples or tomatoes. Most of the

models without factor scores were significant, having a relatively better predictive power

than the naive alternative.














CHAPTER 7
FACTOR ANALYSIS RESULTS

The models with factor scores are presented in Chapter 8, and include factor scores

for both food quality and food preference variables derived from principle component

factor analyses. Initially, they were also supposed to include food safety factor scoress.

However, results from the factor analysis of food-safety-concern variables showed low

correlations plus extremely low Kaiser-Meyer-Olkin (KMO) measures of sampling

adequacy of less than 0.6 (see Appendix F). This necessitated the use of a single question

from the questionnaire as a food safety variable in the final model specifications. For

those variables for which factor analysis is used to derive factor scores, results are

presented in this chapter. Table 7-1 summarizes results of the first factor analysis. The

table shows that when food-quality-related variables were factor analyzed for the apple

data, two factors were extracted. This number of factors extracted was established based

on the scree plot analysis decision rule, and in the process a 77 percent extraction rate

was achieved as well as a 0.843 KMO measure of sampling adequacy. The factor

loadings distributed evenly between the two factors named "general quality conscious"

and "natural quality conscious". As is true in all factor analyses, naming of factors was

solely based on the researcher's intuition. Thus, the names should be viewed in that

context. Factor scores were saved using the regression method for later use in the double-

hurdle models presented in Chapter 8.










Table 7-1. Rotated component matrix(a) for food quality proxy variables-apple data
Factor

Proxy Variable 1 (General Quality conscious) 2 (Natural Quality conscious)

I usually aim to eat natural food .227 .813

I am willing to pay somewhat more for a .764 .333
product of better quality

Quality is decisive for me in purchasing .919 .181
foods

I always aim at the best quality .860 .343

When choosing foods, I try to buy .271 .799
products that do not contain residuals of
herbicides and antibiotics

I am willing to pay somewhat more for .396 .825
food containing natural ingredients


For me, wholesome nutrition begins with .704 .462
the purchase of foods of high quality


Table 7-2 shows results of the second factor analysis. In this case, food preference

variables were reduced to yield three factors and these being extracted also based on a

scree plot decision rule. This time, a 69 percent extraction rate was achieved with a high

KMO measure of sampling adequacy of 0.86. Scree plots and extraction percentages of

all factor analyses in this study are shown in Appendix F. Again, names were intuitively

assigned to the factors derived. Notably, the variable, "I will eat almost anything," had

relatively high loadings on two factors at the same time, i.e. "Open to unfamiliar foods"

and "Choosey". This made naming of the factors a little more challenging.

In the tomato data, two factors were also derived for the quality variables with 73

percent of the variance in the original variables being captured by the factors. Table 7-3

shows how the variables' factor loadings distributed.













Table 7-2. Rotated component matrix(a) for food preference proxy variables-apple data
Factor


1 (Open to unfamiliar
foods)


I like foods from different countries

Ethnic food looks too weird to eat

I like to try new ethnic restaurants

At parties, I will try a new food

I am very particular about the foods I
will eat

I am constantly sampling new and
different foods

I don't trust new foods

I will eat almost anything

If I don't know what is in a food, I
won't try it

I am afraid to eat things I have never
eaten before


.712

-.279

.832

.852

-.190


.685


-.285

.585

-.080


-.105


2 (Choosey) 3 (Afraid of
unfamiliar foods)

-.045 -.161

.040 .778

-.165 -.288

-.224 -.155

.765 .135


-.280


.283

-.654

.730


.490


-.382


.700

-.023

.337


.645


The variables "I usually aim to eat natural food" and "For me wholesome nutrition

begins with the purchase of foods of high quality" had more or less the same levels of

correlation with each factor, hence these variables contributed significantly to both

factors' variances. The interpretation of this could be that the nutrition and natural

product attributes are perceived by consumers as indicators of quality in general and

specific quality associated with natural foods. Perhaps no real distinction exists for these

variables in relation to these quality factors.


Proxy Variable










Table 7-3. Rotated component matrix(a) for food quality proxy variables-tomato data
Factor


Proxy Variable


I usually aim to eat natural food

I am willing to pay somewhat more for a product of better
quality

Quality is decisive for me in purchasing foods

I always aim at the best quality

When choosing foods, I try to buy products that do not
contain residuals of herbicides and antibiotics


I am willing to pay somewhat more for food containing
natural ingredients

For me, wholesome nutrition begins with the purchase of
foods of high quality


1 (General Quality 2 (Natural Quality


1 (General Quality
conscious)

.439

.865


.905

.710

.148



.359


.599


2 (Natural Quality
conscious)

.488

.182


.254

.373

.930



.818


.546


Table 7-4 shows further results of the factor analyses with the tomato data. Once

again, three factor scores were derived for consumer food preferences. The third factor

was however unnamed owing to unclear interpretation of the factor loadings. These

results show that sometimes naming of the latent factors in a factor analysis, even after

rotation, can be complicated and at times inappropriate (Hair et al., 2003). This is one

reason why there is continued debate in literature on the concept and relevance of naming

underlying factors. Perhaps the factors' underlying influence on the dependent variable in

subsequent double hurdle analyses is all that matters and derivation of names is best left

alone.

In terms of extraction rate, the three factors were attributable to 68 percent of the

variance in the original variables and a 0.847 KMO measure of sampling adequacy. It is

noteworthy to mention that this and other levels of extraction described in this study,










though relatively low, are statistically acceptable especially given the high KMO

measures of sampling adequacy.

Table 7-4. Rotated component matrix(a) for food preference proxy variables-tomato data
Factor


Proxy Variable 1 (Open to unfamiliar 2 (Choosey) 3
foods) (unnamed)

I like foods from different countries .718 -.160 .069

Ethnic food looks too weird to eat -.650 .549 .107

I like to try new ethnic restaurants .809 -.215 .221

At parties, I will try a new food .725 -.100 .488

I am very particular about the foods I will eat -.063 .614 -.493

I am constantly sampling new and different foods .540 -.206 .560

I don't trust new foods -.355 .726 -.090

I will eat almost anything .154 -.263 .853

If I don't know what is in a food, I won't try it -.049 .703 -.273

I am afraid to eat things I have never eaten before -.312 .743 -.171


For the combined apple and tomato data set, the factor analyses yielded similar

results, as expected. 74 percent extraction rate was achieved for the two factors extracted,

with a KMO measure of 0.851 being attained. Table 7-5 summarizes the factor analysis

performed for the food quality variables.

Once more, the variables "I am willing to pay somewhat more for a product of

better quality," "Quality is decisive for me in purchasing foods" and "I always aim at the

best quality" loaded high on the "General Quality conscious" factor. The other variables

loaded high on the "Natural Quality conscious" factor. Again, the variable "For me,










wholesome nutrition begins with the purchase of foods of high quality" loaded evenly

between the two factors.

Table 7-5. Rotated component matrix(a) for food quality proxy variables-combined apple
and tomato data set
Factor


Proxy Variable

I usually aim to eat natural food

I am willing to pay somewhat more for a product of better
quality

Quality is decisive for me in purchasing foods

I always aim at the best quality

When choosing foods, I try to buy products that do not
contain residuals of herbicides and antibiotics


I am willing to pay somewhat more for food containing
natural ingredients


For me, wholesome nutrition begins with the purchase of
foods of high quality


1 (General Quality
conscious)

.290

.812


.910

.796

.204


2 (Natural Quality
conscious)

.692

.266


.229

.347

.861


.649


Pertaining to consumer food preferences, the results presented in Table 7-6 were

attained for the combined apple and tomato data set. This time the extraction rate was

slightly lower, at 68 percent. The KMO measure of sampling adequacy was 0.863 and a

total of three factors were extracted using the scree plot decision rule. Most of the

loadings were similar to those in the tomato data set, with the majority clearly loading on

a single factor. Only "I will eat almost anything" and "If I don't know what is in a food, I

won't try it" did not, possibly indicating the vagueness of these statements.










Table 7-6. Rotated component matrix(a) for food preference proxy variables-tomato data
Factor


1(Open to different 2(Choosey) 3 (afraid of unfamiliar
foods) foods)

Proxy Variable

I like foods from different countries .689 .003 -.261

Ethnic food looks too weird to eat -.398 .006 .752

I like to try new ethnic restaurants .808 -.161 -.292

At parties, I will try a new food .832 -.286 -.115

I am very particular about the foods I -.140 .723 .272
will eat

I am constantly sampling new and .669 -.380 -.229
different foods

I don't trust new foods -.263 .305 .703

I will eat almost anything .462 -.766 .040

If I don't know what is in a food, I won't -.036 .633 .446
try it

I am afraid to eat things I have never -.167 .440 .671
eaten before



On the whole, the factor analyses presented here performed well with acceptable

levels of significance and extraction rates. The only case that warranted discarding of the

procedure was that of food safety variables which had low correlations. Perhaps more

questions on food safety should have been asked and framing them in a better way may

have improved the survey design to yield better results.














CHAPTER 8
ECONOMETRIC MODELING WITH FACTOR SCORES

This chapter presents the models with factor scores derived from the factor analyses

alluded to in Chapter 7. As in the case of Chapter 6, three models are developed (i.e. for

apple data, tomato data, and combined apple and tomato data) and their results are

discussed in detail.

Apples Model (Model 4)

When factor scores for food quality and food preferences were incorporated in

place of single variables in the apples probit model, the explanatory power increased

marginally (significance level of 0.94). However, this model was still insignificant as

shown in Table 8-1.

Despite this result, the truncated tobit estimation is presented below in Table 8-2

and as is shown, age, gender, income and location demographics were significant at 0.1

significance level. Surprisingly, the quality variables (QUAL1 and QUAL2) were both

insignificant in the truncated model. Nevertheless, the signs of these quality coefficients

were positive as expected, implying that quality conscious consumers would pay more

for the labeled apples. The signs of the estimated coefficients for demographics were also

as expected save for age, which was negative, suggesting that older consumers were

unwilling to pay more for labeled apples. With respect to income, consumers with an

annual pre-tax income of less than $50,000 were willing to pay more for COOL. The

other income group dummy variables were insignificant suggesting that they have the

same effect as the base ($100,000 and more).











Table 8-1. Apples probit model with factor scores (Model 4)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0031 0.0146 0.00 -0.215 0.8301 45.22
GENDER 0.2107 0.4408 0.06 0.478 0.6326 0.90
EDU1 -0.0799 0.3850 -0.02 -0.208 0.8356 0.35
EDU2 0.2305 0.3508 0.06 0.657 0.5112 0.38
LOC1 -0.1216 0.3710 -0.03 -0.328 0.7432 0.41
LOC2 0.5014 0.4963 0.11 1.010 0.3124 0.13
INC1 0.1824 0.3905 0.05 0.467 0.6403 0.26
INC2 -0.0315 0.3903 -0.01 -0.081 0.9356 0.19
INC3 0.2275 0.3890 0.06 0.585 0.5586 0.22
EXPOSE 0.0010 0.2178 0.00 0.004 0.9964 1.90
PC1 0.0820 0.4663 0.02 0.176 0.8604 0.26
PC2 0.6342 0.4935 0.14 1.285 0.1987 0.15
PC3 -0.0463 0.3560 -0.01 -0.130 0.8965 0.37
TRUST 0.1461 0.0962 0.04 1.520 0.1286 3.69
SAFE 0.0091 0.0911 0.00 0.100 0.9206 3.70
PFR1 -0.1714 0.1640 -0.05 -1.045 0.2960 0.00
PFR2 -0.0112 0.1525 0.00 -0.074 0.9412 0.00
PFR3 0.1364 0.1456 0.04 0.937 0.3487 0.00
QUAL1 0.1339 0.1511 0.04 0.886 0.3754 0.00
QUAL2 0.3361 0.1592 0.09 2.111 0.0348 0.00
Restricted log likelihood value, In L10 = -69.15 R2 (McFadden, 1973) = .0760
Maximum unrestricted log likelihood value, In L1 = -63.90 R2 (Estrella, 1998)= .0772
Log likelihood (df=19)= 10.50 (p = .9394) % of correct predictions = 80.1
Number of observations = 136


Table 8-2. Apples truncated tobit model with factor scores (Model 4)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0249 0.0128 -0.01 -1.955 0.0506 44.64
GENDER 1.0991 0.5256 0.36 2.091 0.0365 0.91
EDU1 0.4658 0.3691 0.15 1.262 0.2070 0.35
EDU2 0.2450 0.3275 0.08 0.748 0.4543 0.38
LOC1 -0.2101 0.3222 -0.07 -0.652 0.5143 0.40
LOC2 -1.9737 0.6091 -0.64 -3.240 0.0012 0.14
INC1 0.8316 0.3345 0.27 2.486 0.0129 0.27
INC2 0.0412 0.3532 0.01 0.117 0.9072 0.19
INC3 -0.1450 0.3608 -0.05 -0.402 0.6877 0.23
EXPOSE -0.2030 0.1976 -0.07 -1.027 0.3045 1.90
PC1 -0.1521 0.3628 -0.05 -0.419 0.6750 0.26
PC2 -0.1248 0.4131 -0.04 -0.302 0.7625 0.17
PC3 0.3012 0.3179 0.10 0.947 0.3435 0.36
TRUST 0.0565 0.0783 0.02 0.722 0.4704 3.74
SAFE -0.0040 0.0941 0.00 -0.042 0.9662 3.72
PFR1 -0.0071 0.1314 0.00 -0.054 0.9570 0.00
PFR2 0.1620 0.1269 0.05 1.277 0.2015 0.02
PFR3 0.0728 0.1236 0.02 0.589 0.5558 0.05
QUAL1 0.1897 0.1600 0.06 1.185 0.2359 0.04
QUAL2 0.1987 0.1379 0.06 1.441 0.1495 0.08
Sigma 0.7057 0.0904 7.804 0.0000
Number of observations = 136 Log likelihood function = -32.87
Observation after truncation = 108 Threshold values for model: Lower = 0 Upper = +oo











In terms of location, it was once again revealed that consumers in Lansing MI are

not willing to pay that much for produce labeled "U.S.A. Grown" unlike consumers in

Gainesville FL or Atlanta GA.



Tomatoes Model (Model 5)

When it came to the tomato model estimation, a slightly different picture was

found. Table 8-3 shows that most of the estimated coefficients for the tomatoes probit

model with factor scores had the expected signs even though only quality and location

were significant at 0.1 significance level.

Table 8-3. Tomatoes probit model with factor scores (Model 5)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0052 0.0131 0.00 -0.396 0.6918 44.21
GENDER -0.0592 0.3266 -0.02 -0.181 0.8561 0.87
EDU1 0.0248 0.3043 0.01 0.082 0.935 0.33
EDU2 0.3589 0.2900 0.13 1.238 0.2159 0.43
LOC1 0.6895 0.3008 0.26 2.292 0.0219 0.38
LOC2 -0.2145 0.2882 -0.08 -0.744 0.4566 0.28
INC1 -0.0578 0.3141 -0.02 -0.184 0.8539 0.27
INC2 -0.2076 0.3093 -0.07 -0.671 0.5022 0.28
INC3 -0.0952 0.3703 -0.03 -0.257 0.7972 0.14
EXPOSE 0.1879 0.1462 0.07 1.285 0.1989 1.94
PC1 -0.2490 0.4260 -0.09 -0.584 0.5589 0.23
PC2 -0.1610 0.3894 -0.06 -0.414 0.6792 0.22
PC3 -0.2101 0.3176 -0.08 -0.662 0.5083 0.37
SAFE 0.0742 0.0765 0.03 0.97 0.3321 3.41
TRUST 0.0589 0.0748 0.02 0.787 0.4314 3.71
PFR1 -0.1192 0.1240 -0.04 -0.961 0.3363 0.00
PFR2 0.0271 0.1175 0.01 0.231 0.8174 0.00
PFR3 0.0819 0.1132 0.03 0.723 0.4695 0.00
QUAL1 0.3140 0.1289 0.12 2.436 0.0149 0.00
QUAL2 0.1378 0.1217 0.05 1.132 0.2576 0.00
Restricted log likelihood value, In Lio = -104.70 R2 (McFadden, 1973) = .14248
Maximum unrestricted log likelihood value, In L1 = -89.77989 R2 (Estrella, 1998)= .16800
Log likelihood X2(df19)= 29.83458 (p = .0539) % of correct predictions = 74.3
Number of observations = 175


The model predicted correctly for 74.3 percent of the actual responses recorded.

Marginal effects indicated that being a Gainesville, FL (LOC1) consumer would increase







73


the probability of being willing to pay a premium for tomatoes labeled "U.S.A. Grown"

by approximately 26 percent as compared to a consumer from Atlanta, GA, ceteris

paribus. In contrast, being a Lansing, MI (LOC2) consumer did not change the

probability of being willing to pay for tomatoes labeled "U.S.A. Grown" when compared

to the base (consumers in Atlanta, GA).

Also according to model 5, consumers' perceptions about food quality in general

were shown to have a positive impact on the likelihood of being willing to pay a premium

for COOL in fresh tomatoes. As displayed in Table 8-3, the marginal probability thereof,

ceterisparibus, is 12 percent. The truncated estimation in model 5 is shown in Table 8-4.

Table 8-4. Tomatoes truncated tobit model with factor scores (Model 5)
Variable Estimated Standard Marginal Standardized p-value Mean of
Coefficient Error Effects Coefficient Regressor
AGE -0.0011 0.0096 0.00 -0.118 0.9057 44.26
GENDER 0.1995 0.2600 0.09 0.767 0.4429 0.86
EDU1 -0.3170 0.2201 -0.15 -1.440 0.1498 0.32
EDU2 -0.2297 0.2078 -0.11 -1.105 0.2690 0.46
LOC1 0.4261 0.2074 0.20 2.054 0.0400 0.45
LOC2 -0.6325 0.3224 -0.30 -1.962 0.0498 0.22
INC1 0.2769 0.2399 0.13 1.154 0.2484 0.27
INC2 0.3925 0.2384 0.19 1.647 0.0996 0.27
INC3 0.3744 0.2767 0.18 1.353 0.1761 0.14
EXPOSE -0.1740 0.1170 -0.08 -1.487 0.1371 1.94
PC1 -0.2476 0.2976 -0.12 -0.832 0.4054 0.22
PC2 -0.0260 0.2644 -0.01 -0.098 0.9218 0.22
PC3 -0.1813 0.2360 -0.09 -0.768 0.4424 0.36
TRUST 0.0374 0.0562 0.02 0.664 0.5066 3.46
SAFE 0.0987 0.0591 0.05 1.670 0.0950 3.86
PFR1 0.0535 0.1022 0.03 0.524 0.6004 0.02
PFR2 -0.1542 0.0878 -0.07 -1.756 0.0791 -0.01
PFR3 -0.0004 0.0895 0.00 -0.004 0.9968 0.00
QUAL1 0.1021 0.1145 0.05 0.892 0.3723 0.12
QUAL2 0.0916 0.0935 0.04 0.979 0.3274 0.09
Sigma 0.6093 0.0717 8.497 0.0000
Number of observations = 175 Log likelihood function = -47.01
Observation after truncation = 125 Threshold values for model: Lower = 0 Upper = +o


It was once again noted that location is a significant factor influencing how much

consumers will pay for COOL in tomatoes. Consumers in Gainesville, FL will pay on









average $0.20 more than Atlanta consumers while Lansing consumers will pay

approximately $0.30 less on average. Also shown is that household income level affects

the premium that consumers will pay for fresh tomatoes labeled "U.S.A. Grown."

Consumers in the $50,000 to $74,999 (INC2) income bracket will pay on average $0.19

more than consumers in the $100,000 or more income bracket (the base). Generally,

consumers with incomes of less than $100,000 are shown to be willing to pay a larger

premium for COOL in tomatoes.

An additional finding of model 5 is that consumers' views about food safety have a

bearing on how much they will pay as a premium for "U.S.A. Grown" labeling in fresh

tomatoes. This was strangely not so in the apple model (i.e. model 4). Consumers who

said they think about food safety when purchasing fruits and vegetables are found to be

willing to pay more for tomatoes labeled "U.S.A. Grown." Perhaps food safety concerns

of the surveyed consumers are somewhat greater in fresh tomatoes than in fresh apples.

Model 5 also estimates that consumers' food preferences have some bearing on

how much consumers will pay for COOL in fresh tomatoes. However, only the second

factor score for food preferences, "choosey," is significant at a = 0.1 and it negatively

impacts how much consumers will pay as a premium for tomatoes labeled "U.S.A.

Grown." The negative impact is quite unexpected because one would think that

consumers who are particular about the food they eat will pay more to know the origin of

their tomatoes. In this case, the model predicts the opposite; particular consumers are

unwilling to pay more for tomatoes labeled "U.S.A. Grown." It is possible that these

consumers trust the marketing system to ensure that tomatoes made available are of

acceptable quality and standards, thus they are not too concerned about country of origin.










This may also be the same for high income level consumers who also proved to be less

concerned about country of origin.

Combined Apples and Tomatoes Model (Model 6)

Table 8-5 shows the probit estimation for the combined apple and tomato data. It

shows that both food quality factor scores were significant at a = 0.05.

Table 8-5. Combined apples and tomatoes probit model with factor scores (Model 6)
Varible Coefficient Standard Marginal Standardized p-value Mean of
Error Effects Coefficient Regressor
AGE -0.0077 0.0088 0.00 -0.874 0.382 44.65
GENDER -0.0655 0.2415 -0.02 -0.271 0.7863 0.88
EDU1 0.1513 0.2219 0.05 0.682 0.4953 0.34
EDU2 0.2548 0.2105 0.08 1.211 0.226 0.41
LOC1 0.2909 0.2140 0.09 1.359 0.1741 0.40
LOC2 -0.2283 0.2195 -0.08 -1.04 0.2982 0.21
INC1 -0.0207 0.2280 -0.01 -0.091 0.9276 0.27
INC2 -0.1404 0.2328 -0.05 -0.603 0.5464 0.24
INC3 0.0822 0.2566 0.03 0.32 0.7489 0.17
EXPOSE 0.1826 0.1130 0.06 1.615 0.1062 1.91
PC1 -0.1269 0.2760 -0.04 -0.46 0.6457 0.24
PC2 0.0933 0.2733 0.03 0.341 0.7328 0.19
PC3 -0.0860 0.2209 -0.03 -0.389 0.6971 0.37
SAFE 0.0474 0.0538 0.02 0.881 0.3781 3.71
TRUST 0.1133 0.0545 0.04 2.08 0.0375 3.53
PFR1 -0.1170 0.0890 -0.04 -1.314 0.1889 0.00
PFR2 -0.0192 0.0848 -0.01 -0.226 0.8212 0.00
PFR3 0.0502 0.0849 0.02 0.592 0.554 0.00
QUAL1 0.1762 0.0859 0.06 2.052 0.0402 0.00
QUAL2 0.2264 0.0902 0.07 2.509 0.0121 0.00
Restricted log likelihood value, In L10 = -179.40 R2 (McFadden, 1973) = .07925
Maximum unrestricted log likelihood value, In L1 = -165.1843 R2 (Estrella, 1998)= .09086
Log likelihood X(df-19)= 28.43597 (p = .0754) % of correct predictions = 74.0
Number of observations = 311




Consumers who were more conscious about food quality (be it quality in general or

quality associated with natural foods) were found to be more likely to pay a premium for

fresh apples or tomatoes labeled "U.S.A. Grown." The level of trust that consumers have

for information they receive from U.S. government agencies (e.g. USDA, FDA, EPA

etc.) was the only other significant variable (at a = 0.05). Here it was found that

consumers who were more trusting of the information they receive from U.S. government









agencies were more likely to pay for COOL. Surprisingly, all demographics turned out to

be insignificant in the participation decision making process, suggesting that it does not

matter if one is male or female or if income is high or low. Perhaps the participation

decision is simply not a function of demographics. Overall, the model was significant (p

= 0.075) and had a 74.0 percent correct prediction rate.

The truncated tobit estimation is presented in Table 8-6, where it is reported that

gender and location significantly determine how much the consumers are willing to pay

once they have decided they are willing to pay for COOL. Marginal effects show the

expected result that female consumers are willing to pay 20 cents more per pound than

males. In terms of location, the base used was Atlanta, GA and consumers in Lansing, MI

were found to be willing to pay substantially less than those in Atlanta (approximately 46

cents per pound less). In contrast, consumers in Gainesville, FL were willing to pay

approximately 4 cents per pound more than those in Atlanta, GA. A reason for this could

be that MCOOL policy has been prevalent for the past 26 years at the state level in

Florida. Thus, shoppers in Gainesville FL could be accustomed to MCOOL and therefore

are willing to pay for COOL. Conversely, Michigan is geographically far from either

Georgia or Florida. Thus it could be the case that Michigan consumers are less exposed to

COOL and are therefore less willing to pay for COOL. Another possible explanation is

that Michigan borders Canada, and perhaps Michigan consumers are accustomed to

buying and consuming food from Canada. For more details on location-based premium

differentials refer to Table 5-3 in Chapter 5.

Income level is another demographic that seemed to have an impact on the amount

consumers are willing to pay. The greater-than-$100,000 income group was used as the











base in the model. Findings suggest that consumers with an income level of less than

$50,000 are the only group with a significantly greater WTP than the base (15 cents per

pound more). This finding is similar to that in Loureiro and Umberger (2002), implying

that more affluent consumers consider it unimportant to know where their apples or

tomatoes come from and thus do not value COOL.

Table 8-6. Combined apples and tomatoes truncated tobit model with factor scores


(Model 6)
Variable Estimated
Coefficient
AGE -0.0156
GENDER 0.5704
EDU1 -0.0216
EDU2 -0.0498
LOC1 0.1206
LOC2 -1.3283
INC1 0.4493
INC2 0.2555
INC3 0.0596
EXPOSE -0.2253
PC1 -0.1400
PC2 -0.0887
PC3 0.0583
SAFE 0.1204
TRUST 0.0306
PFR1 -0.0965
PFR2 0.0556
PFR3 -0.0695
QUAL1 0.1201
QUAL2 0.1701
Sigma 0.7937
Number of observations =


Standard
Error
0.0106
0.3406
0.2406
0.2276
0.2082
0.4310
0.2492
0.2476
0.2817
0.1327
0.2921
0.2785
0.2351
0.0667
0.0570
0.1051
0.0862
0.0938
0.1188
0.1042
0.0892
311


Marginal
Effects
-0.01
0.20
-0.01
-0.02
0.04
-0.46
0.15
0.09
0.02
-0.08
-0.05
-0.03
0.02
0.04
0.01
-0.03
0.02
-0.02
0.04
0.06


Standardized
Coefficent
-1.478
1.675
-0.09
-0.219
0.579
-3.082
1.803
1.032
0.212
-1.697
-0.479
-0.319
0.248
1.804
0.537
-0.918
0.645
-0.741
1.011
1.633
8.901


p-value

0.1394
0.0940
0.9285
0.8266
0.5625
0.0021
0.0714
0.3023
0.8324
0.0897
0.6319
0.7500
0.8041
0.0712
0.5911
0.3587
0.5191
0.4588
0.3120
0.1026
0.0000


Mean of
Regressor
44.46
0.89
0.34
0.42
0.43
0.17
0.27
0.23
0.18
1.93
0.24
0.20
0.36
3.80
3.62
0.00
0.01
0.01
0.08
0.09


Log likelihood function = -105.90


Observation after truncation = 233 Threshold values for model: Lower = 0 Upper = +oo



The truncated estimation also displayed the food quality factor score that is

associated with natural foods as having a positive impact on the price premium

consumers are willing to pay for the "U.S.A. Grown" labeled fresh apples and tomatoes.

In other words, consumers who are more natural-quality-conscious would pay more for

COOL. However, the factor score for quality in general, had no significant impact on the


amount they were willing to pay.


I









The food safety variable is the other factor that was significant at a significance

level of 0.1. Consumers that think about food safety when purchasing fruits and

vegetables were found to be willing to pay more for the label "U.S.A. Grown."

Unexpectedly, this food safety variable had not been significant in the participation stage,

suggesting that it only plays a role on the consumption decision. The consumers' self-

rated exposure to food safety information in fresh fruits and vegetable is another variable

that was significant in model 6 and only with respect to the consumption decision.

Consumers that rated themselves as having seen, read or heard less about food safety in

fresh fruits and vegetables were willing to pay less for the label "U.S.A. Grown." This

suggests that consumers may be likely to pay more for produce labeled "U.S.A. Grown"

if they are more informed about food safety issues in fresh fruits and vegetables.














CHAPTER 9
CONCLUSIONS AND IMPLICATIONS

Summary

This thesis researches consumers' WTP for COOL in fresh apples and tomatoes,

specifically the label attribute "U.S.A. Grown." Previous research on WTP for COOL has

focused primarily on the beef sector with very little work being done on COOL in the

produce sector. Thus, this thesis looks at the situation in the produce sector in order to

provide empirical information that can be used by policy makers and decision makers in

the produce sector. Given rising import competition in the produce sector and imminent

changes in COOL legislation, which will see current voluntary COOL becoming

mandatory, this information is timely and can contribute to a more informed decision

making process.

The study determines the nature of consumers' WTP for COOL in fresh apples and

tomatoes, ascertaining how much consumers are willing to pay and whether the WTP is

product specific or not. Factors that influence the WTP are researched by applying

Cragg's double hurdle model to data collected from primary shoppers in Gainesville, FL,

Lansing, MI and Atlanta, GA. In total, six separate double hurdle models are estimated

using the same model specification. Three of these estimations incorporate factor scores

of food quality and food preference variables as regressors, while the other three do not.

In each set of three models, one analyzes fresh apple data, another fresh tomato data and

a third combined fresh apple and tomato data.









Findings show that consumers from Michigan, Florida and Georgia are willing to

pay a premium for the COOL attribute, "U.S.A. Grown," in fresh apples and tomatoes. 79

percent of the consumers sampled are found to be willing to pay a premium for fresh

apples labeled "U.S.A. Grown" while 72 percent are willing to pay a premium in the case

of fresh tomatoes labeled "U.S.A. Grown." The mean WTP for COOL is calculated to be

approximately $0.49 and $0.48 per pound of U.S. labeled apples and tomatoes

respectively. These findings lead to the failure to reject the first hypothesis of the study,

thereby accomplishing the first specific objective outlined in Chapter 1.

The thesis also ascertains that the premium consumers are willing to pay for fresh

apples labeled "U.S.A. Grown" is statistically equivalent to the premium they are willing

to pay for fresh tomatoes labeled "U.S.A. Grown." This results in the rejection of the

second hypothesis of the study and achieves the second specific objective.

Concerning the factors that influence the WTP for the COOL attribute, "U.S.A.

Grown" in fresh apples and tomatoes, different models are developed, which come up

with slightly different results, possibly indicating that though the premiums may be

equivalent, the factors affecting them may differ slightly. Nonetheless, most of the

estimated coefficients are found to have the expected signs on them and similar findings

are made in all models.

It is established that consumer food quality perception is a strong predictor of WTP

for COOL in fresh apples and tomatoes. Location is also a significant explanatory

variable, with consumers from Lansing, MI being the least willing to pay for the label

"U.S.A. Grown" in both fresh apples and fresh tomatoes. Consumers from Gainesville,

FL turn out to be willing to pay the highest average price premium for COOL in fresh









tomatoes while consumers from Atlanta, GA are willing to pay the highest average price

premium for COOL in fresh apples. Consumer food preferences (as defined according to

the questions in the questionnaire in Appendix B) are found to be insignificant in most

models. Whether consumers are choosey or not about the food they eat does not seem to

affect their WTP for the label "U.S.A. Grown" in fresh apples and tomatoes. Similarly,

the gender, level of education and the presence of children in a household seem not to

have much of an effect on consumers' WTP for COOL in fresh apples and tomatoes.

Also, consumers' level of exposure to food safety information in fruits and vegetables

seem not to have much of an effect either. Most psychographics and demographics such

as gender, age, education, and food safety concerns seem to have minor effect depending

on the product under consideration. This is particularly so, in the truncated models where

the consumption decision (i.e. how much to pay) is estimated. These findings relate to the

third hypothesis of this study and lead to the rejection of some aspects of the hypothesis.

Consumer perceptions about food preferences and food safety are not key factors of WTP

for COOL in fresh apples and tomatoes, as demonstrated by the lack of significance at

0.1 levels. However, consumer perceptions about food quality are found to be statistically

significant, suggesting that they are key factors of WTP for COOL in fresh apples and

tomatoes.

Implications

There are several implications that arise from the findings made in this thesis. First,

by establishing that consumers are willing to pay for the label "U.S.A. Grown" in both

fresh apples and fresh tomatoes, the study implies that U.S. consumers want to know the

country of origin of the apples and tomatoes they consume. This contributes to

justification (on the consumers' side) of MCOOL or at the very least COOL on a









voluntary basis, in the apple and tomato markets; either mandatory or voluntary COOL

legislation would facilitate the provision of country-of-origin information as desired by

consumers at the final point of purchase.

Also, the findings suggest that it may be possible for producers and marketers to

use the label "U.S.A. Grown" in order to garner a competitive advantage over import

substitutes in the U.S. market. Consumers surveyed in this study were willing to pay

approximately $0.49 and $0.48 more for a pound of U.S. labeled apples and tomatoes

respectively, over unlabeled produce. However, the costs associated with incorporating

labels have to be calculated before an additional price mark up can be considered.

Furthermore, it would be necessary to compare the net price premium (i.e. the price

premium after deducting costs of incorporating labeling) for "U.S.A. Grown" produce

with those of other countries.

The thesis also establishes premium equivalency in fresh apples and tomatoes

labeled "U.S.A. Grown," possibly implying that there is potential for generic promotion

of the label "U.S.A. Grown" to enhance overall demand for U.S. produce over imports.

This is, however, not conclusive from the findings of this study, and complementary and

extensive research is required to make more than tentative claims about this.

By making use of Cragg's double hurdle model to estimate the WTP for COOL,

this thesis finds that consumers' food quality perceptions are critical factors in both the

participation and consumption decision making processes. This implies that before

COOL can be used as a tool for enhancing demand for U.S. produce; produce quality will

have to be ensured, otherwise, COOL is likely to induce a negative market response. It is

apparent from the research that the label "U.S.A. Grown" may serve to inform consumers









not only the country-of-origin but also the quality of the produce. This is also an

important implication with respect to prospects of generic promotion of U.S. labeled

produce. If any form of generic promotion is to be chosen by industry players, then

ensuring consistent quality will be most critical. Poor quality in any part of the industry

may tarnish the image of the label, consequently causing a decline in the price consumers

are willing to pay for produce labeled "U.S.A. Grown."

Location is another factor that is clearly shown to influence how much consumers

will pay for produce labeled "U.S.A. Grown." An implication from this is that if a

marketing strategy is to be developed for fresh tomatoes and apples labeled "U.S.A.

Grown," it will have to be sensitive to the location of the targeted consumers. If pricing

and promotion strategies are to be successful, they may need to be different in accordance

with the nuance of dominant consumer preferences at each location. Given that in Florida

(where MCOOL has been in place for 26 years) consumers were willing to pay more for

COOL, it is implied that U.S. consumers can become loyal to the label "U.S.A. Grown"

if exposed to it for a prolonged time. Perhaps, Floridian consumers may have learned to

expect labeling over the years.

Other factors that the thesis finds to be of relative importance include the extent to

which consumers trust the information they receive from U.S. government agencies (e.g.

USDA, FDA and EPA) and the consumers' food safety concerns. While most models

find these to be insignificant factors, it is clear that they have some bearing on how much

consumers will pay for COOL. Concerning the former, it is found to be a significant

determinant of the participation decision in one of the models with factor scores (Model

6). This is likely due to the fact that agencies are responsible for regulating and enforcing









produce labeling laws. If consumers trust the information they are getting from these

agencies, then they are more likely to value COOL because they would believe that there

is a trustworthy labeling verification system in place. Since COOL is a credence-attribute

label that consumers cannot verify except through a third party (i.e. an agency), this

would be important.

As pertaining to food safety concerns, these seem to mainly affect the WTP in

tomatoes and not in apples. The truncated estimations of the tomatoes models (Models 2

and 5) show that consumers who take food safety concerns into consideration when

making the decision to purchase fruits and vegetables will pay more money for the label

"U.S.A. Grown" in fresh tomatoes. This is somewhat expected, since fresh tomatoes are

relatively more perishable and prone to food safety problems. Another possible

implication here is that consumers regard U.S. tomatoes to be safer implying that U.S.

fresh tomato producers may be able to capitalize on this to garner competitive advantage

over import substitutes.

With respect to demographics, most are found to be insignificant predictor

variables, with gender and income level variables turning out to be mildly important. It is

critical to note that the data analyzed in this study was drawn from primarily high income

earners and highly educated individuals. Thus, for purposes of marketing the "U.S.A.

Grown" label, information in this thesis could be used to target this particular

demographic in the U.S. primary shopper population.

All in all, this thesis reports the findings that consumers' food quality perceptions

and consumers' location are important determinants of WTP for COOL in apples and

tomatoes. Most demographics are also somewhat important factors, especially in the









consumption decision on how much to pay, (i.e. once the consumer is willing to pay a

premium). Though food safety, trust and preferences are psychographics included in the

models estimated, these are relatively less important.

Areas for Further Research

The scope of this thesis, like all others, is and had to be limited. Several aspects

relating to the research topic are simply not covered, leaving some areas for further study.

Though this research finds consumers to be willing to pay for fresh apples and tomatoes

labeled "U.S.A. Grown" there is little knowledge on how the U.S. label would fare

against other country-of-origin labels. Researching this topic is imperative before

substantive conclusions can be drawn about promoting the generic "U.S.A. Grown" label.

The ramifications of such a promotional program on import competition in the U.S.

produce industry can only be estimated if competitor country labels' price premiums are

calculated and compared to that for the "U.S.A. Grown" label.

Using the same data that were collected and analyzed for this thesis, a forthcoming

study addresses some of these issues. Preliminary findings from the study, suggest that

the "U.S.A. Grown" label fares relatively well against several competitor country labels.

Table 9-1 presents a synopsis of this and shows the price premiums that consumers were

willing to pay to give up one pound of foreign fresh apples or fresh tomatoes, (which they

had been endowed with) in exchange for identical produce labeled "U.S.A. Grown." As

described in Chapter 3, each participating consumer had initially been endowed with one

pound of unlabeled fresh apples (tomatoes) and was then asked to bid for identical

produce labeled "U.S.A. Grown." What is not mentioned in Chapter 3 is that after the

fourth round of bidding had occurred, consumers were informed the origin of their fresh

apples (tomatoes), which they were initially endowed with.










Table 9-1. Comparison of mean bids: U.S.A. Grown versus Other Country labels
Paired
Samples Number of
WTP for U.S. Label Mean Test Observations
Difference t-value
Before After
Information Information
Apples
U.S.A. Grown versus No label 0.49 136
U.S.A. Grown versus Chile 0.42 0.41 -0.01 -0.240 59
U.S.A. Grown versus China 0.37 0.46 0.09 2.658 39
U.S.A. Grown versus New
Zealand 0.71 0.88 0.17 2.043 38
Tomatoes
U.S.A. Grown versus No label 0.48 175
U.S.A. Grown versus Canada 0.34 0.38 0.04 1.475 67
U.S.A. Grown versus Mexico 0.58 0.93 0.35 4.432 86
U.S.A. Grown versus the
Netherlands 0.56 0.67 0.11 1.941 22

Thus, in Table 9-1, the first two columns titled "before information" and "after

information," show the mean WTP for the fresh produce labeled "U.S.A. Grown" before

the foreign country information was revealed and after the foreign country information

was revealed to the consumers, respectively. Since each survey site had been told a

different foreign country of origin, the numbers of observations vary though the totals

sum up to the sample sizes of 136 for apples and 175 for tomatoes. Further comparative

analysis of the WTP by location, and with larger sample sizes, would be useful in future

studies if this is to better inform U.S. producers and marketers of fresh produce.

It would also be interesting to find out if U.S. producers would be better off using

generic promotion of the label "U.S.A. Grown" in an effort to enhance market demand

for U.S. grown produce. Currently, it is unclear if a firm level, sub-sector level or

national level approach to promote the "U.S.A. Grown" label would be appropriate.

Establishing this would be of keen interest to several players in the produce market who









may want to enhance demand for U.S. grown fresh tomatoes and apples, given the rising

import competition recorded in the last few years.

Concerning producers' and marketers' profitability, it would be useful to find out if

costs associated with implementing COOL can be offset by what consumers are willing

to pay for it. With the coming of MCOOL, this would be an important question to answer

since MCOOL would certainly introduce new costs. Though this thesis makes findings

that add to the justification for MCOOL or at the very least voluntary COOL, it does so

by establishing whether or not consumers are willing to pay for the label "U.S.A.

Grown." The cost implications thereof are not addressed.














APPENDIX A
EXPERIMENTAL AUCTIONS INSTRUCTIONS

Procedures for Experimental WTP Auctions

Instructions

Thank you for agreeing to participate in today's session. As you entered the room, you
should have been given $10.00 and a packet. You should also have been assigned an ID
number, which is located on the upper right hand corer of the packet. You will use this
ID number to identify yourself during this research session. We use random numbers in
order to ensure confidentiality.

Before we begin, I want to emphasize that your participation in this session is completely
voluntary. If you do not wish to participate in the experiment, please say so at any time.
Non-participants will not be penalized in any way. I want to assure you that the
information you provide will be kept strictly confidential and used only for the purposes
of this research. Please complete the consent form at this time.

In today's session, we are ultimately interested in your preferences for several different
types of foods. First, we will conduct a "food evaluation" exercise, followed by a
"consumer survey". Before we begin, I would like you all to open your packets and take
a minute to fill out the consent form.

I will now begin going through a set of instructions with you and will read from this
script so that I am able to clearly convey the procedures. Importantly, from this point
forward, I ask that there be no talking among participants.

Are there any questions before we begin?

In today's session, we are ultimately interested in your preferences for fruits and
vegetables. To determine how much these foods may be worth to you, we are going to
conduct an auction. To begin, however, we will conduct a candy bar auction so that you
can learn how the auction procedures work.

First, let me say that I realize the following instructions might be a bit confusing at first.
Don't get frustrated. We are conducting the candy bar auction first so that you will have
a chance to learn how things work. I will also provide an example that should help to
further clarify any confusion you might have.