Citation
Behavioral Economics, Nutrition Education, and Access to Markets

Material Information

Title:
Behavioral Economics, Nutrition Education, and Access to Markets Experimental Evidence and a Theoretical Framework for Improving Dietary Diversity
Creator:
Davidson, Kelly A
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (243 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Food and Resource Economics
Committee Chair:
KROPP,JACLYN DONNA
Committee Co-Chair:
MULLALLY,CONNER
Committee Members:
USECHE,MARIA DEL-PILAR
RUSSO,SANDRA L

Subjects

Subjects / Keywords:
access -- agriculture -- bangladesh -- behavioral -- development -- dietary -- diversity -- economics -- experimental -- market -- nudging -- nutrition
Food and Resource Economics -- Dissertations, Academic -- UF
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Food and Resource Economics thesis, Ph.D.

Notes

Abstract:
More than two billion people worldwide suffer from micronutrient deficiencies, a problem referred to as "hidden hunger." To combat this problem, behavior change communication (BCC) strategies are commonly used in developing countries to change knowledge, beliefs, or values surrounding nutrition. Behavioral economics (BE) is an alternative approach that uses subtle cues to nudge individuals toward healthier food choices. BE strategies are used extensively in the developed world to promote nutrition. However, no study to date has documented the effectiveness of nudges for improving nutrition in development. This dissertation investigates whether a food-based dietary guidelines (FBDG) nudge, similar to MyPlate, and a BCC strategy in the form of nutrition education can improve nutrition in rural Bangladesh. Research in the US suggests MyPlate nudges are effective at encouraging individuals to consume nutrient-rich foods. In a randomized experiment, we measure the individual and combined treatment effects of the FBDG nudge and nutrition education. To measure the short-term impacts, we observe participants' food choices in a lunch buffet meal where constraints to food access, such as income and availability, are removed. The long-term effects of the interventions are measured in the home, where constraints to food access are restored, using pre-and post-intervention survey data. The meal observation results suggest that the FBDG icon alone does not lead to healthier food choices in the short-term; however, combining the FBDG nudge with nutrition education encourages participants to consume a wider variety of foods. Furthermore, repeated exposure to the FBDG icon in the home environment modestly improves the food consumption score, a 7-day measure of dietary variety. Our findings suggest that other factors besides knowledge may constrain household dietary quality. Hence, we explore the relationships between market participation, production diversity, and household dietary diversity by expanding Barnum and Squire's (1979) agricultural household model to include health, nutrition, and market transaction costs. Empirical analyses suggest a positive correlation between participation in markets for selling products and production diversity, but a negative correlation between participation in markets for selling products and dietary diversity. Thus, household constraints such as market access should be considered in the design of nutrition interventions. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2017.
Local:
Adviser: KROPP,JACLYN DONNA.
Local:
Co-adviser: MULLALLY,CONNER.
Statement of Responsibility:
by Kelly A Davidson.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
LD1780 2017 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

BEHAVIORAL ECONOMICS, NUTRITION EDUCATION, AND ACCESS TO MARKETS: EXPERIMENTAL EVIDENCE AND A THEORETICAL FRAMEWORK FOR IMPROVING DIETARY DIVERSITY By KELLY A. DAVIDSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2017

PAGE 2

2017 Kelly A. Davidson

PAGE 3

To women and girls in Bangladesh and all around the world. Be who you were created to be, and you will set the world on fire. St. Catherine of Sienna

PAGE 4

4 ACKNOWLEDGMENTS The proverb I t takes a village usually references child rearing, yet it rings uncannily true for my journey to doctor. Several people offered support and direction on this dissertation. A multitude of individuals contributed to the research fieldwork. A few significant mentors have shaped my academic pursuits, and I could n o t have finished this degree without the support of friends and family. I want to express deepest gratitude to my village. Thank you to my advisor, Dr. Jaclyn Kropp, for the countless hours you have invested in my success. You made me a better researcher. Your dedication and words of encouragement motivated me to finish this dissertation. The process would not have been the same without you and our adventures in Bangladesh. I have gaine d a true friend and an inc redible mentor. Dr. Conner Mullally, thank you for your contributions to this research. I especially appreciate your tolerance of my many questions. Your words of advice often inspired me to believe in myself, and I am grateful fo r your commitment to my academic development. Dr. Pilar Useche, thank you for sharing your expertise and words of wisdom. Our conversations about development economics ultimately inspired this work. Several individuals helped bring this project to fruitio n. The USAID Feed the Future initiative Integrating Gender and Nutrition within Agricultural Extension Services (INGENAES) generously funded the research. Dr. Sandra Russo, thank you for investing in my exploration of dissertation topics You taught me t he value of interdisciplinary work. Many thanks to you and Dr. Kathy Colverson for encouraging me to work in Bangladesh and connecting me to INGENAES. Ms. Andrea Bohn, thank you for investing in this project and having confidence in its success. The resear ch would not have been possible without the generous support from our partner agencies, the Bangladesh Agricultural University Extension Center (BAUEC) and Shushilan. I extend a special note of appreciation to the program directors,

PAGE 5

5 coordinating field officers, our study participants, and their families. A few faculty members at BAU were fundamental to the success of this project: Dr. M. Wakilur Rahman, Dr. Md. Salauddin Palash, Ms. Nishith Zahan Tanny, and Mr. Kazi Farid. Thank you for your patience, time, energy, and dedication to the success of this project. Dr. M. Wakilur Rahman, I especially appreciate your contribution the research would not have been possible without your help. I had the true pleasure of working with an incredible team of young, mo tivated masters students from BAU and Patuakhali Technology and Science University (PTSU) during my fieldwork. Thank you for your friendship, hospitality, enthusiasm and time: Sambhu Singha, Shanta Islam, A. M. Shoaib, Shaon Afroz, Tania Akter, Md. Wahidu l Islam, Nusrat Farah, Md. Mahmudur Rahman, Md. Azihullah, Afroza Ferdousi, Ayesha Siddika, Sadia Afrin Mumu, Umme Salma Ami, Rubaiya Islam Eva, Rubyeat Hossain, Humayan Ahmed, Sadia Sharmin Hoque, Md. Al Amin Sheikh, Nazmul Kaysar, Md. Shakir Miah, Anik R oy, Md. Suzauddoula Sharkar, Afia Jahin, Anup Das, Golam Rabbani, Prodip Kumar Poddar, Md. Nosibul Momin, Shatabdi Roy, Nilima Ahmed, Tapoti Biswas, Mahbuba Ferdous, and Shirin Akter. To our collaborating team at the Bangladesh Institute for ICT Developme nt (BIID), thank you for coordinating the research, my travel, and taking care of many behind the scenes tasks about which I will never know. I especially thank Md. Shahid Uddin Akbar and Md. Selim Akbar for your time, hospitality and friendship. Finally, I offer sincerest gratitude to Olyul Islam and Asma Parvin for facilitating over 60 days of participant trainings. Thank you for the sacrifice of time away from your families and for your dedication to the participants. I will always treasure your friendsh ip and the memories we made. To my Dhaka family, Afrina and Loban, thank you for taking me to Bengali weddings, showing me around the city, and helping me understand cultural barriers. I also appreciate the

PAGE 6

6 cultural advice I received from my fellow graduate student, Marup. Thank you for helping with early translations, teaching me about Bengali markets, and offering project management suggestions. I have been blessed with amazing friendships and mentors throughout my academic career. I wish to acknowledge a few individuals who have kept me on track. Dr. Rick Weldon, thank you for being my sounding board and reminding me to make time for college sports. Dr. Lisa House, you are an inspiration to women in our field. Dr. Rod Clouser, thank you for your unwaver ing support. In addition to my fellow Gators, I would be remiss not to thank my University of Kentucky College of Agriculture family. You believed in me and afforded opportunities I thought were far beyond reach. Susan Skees, Jamie Dunn, and Emily Morgan, thank you for fostering my passion for higher education. Dr. Jerry Skees, thank you for introducing me to international development and policy analysis. You inspired me to pursue graduate studies. Ive always appreciated your advice and encouragement. Ms. Brenda Oldfield, you helped me get here by investing in me when I needed it the most. Several friends kept me afloat throughout the dissertation. Yuan, my cohort, we survived! We make a great team. Ricky, thanks for always checking in on me. Joanna, I am so glad were in this together. Ill miss our coffee and lunch rendezvous. Thank you to the St. As community, especially Vincent, Gail, Sarah, Mary, Elyssa, and Kim for your prayers. I also extend a special thank you to Fr. David for your prayers, spirit ual guidance, and words of encouragement. To Alex and my LIV family, thanks for showing me Im capable of way more than I think! Amber, thank you for always being the friend I could call when I thought no one else would understand.

PAGE 7

7 Finally, I thank my family. To my mom, Kathy, you inspired me to be a strong, independent woman. You taught me to value education and to dream big. I appreciate all the sacrifices you made. To Kerry, thank you for always reminding me that I was pursuing my passion You kept m y spirits high by sending pictures of the sweetest little boy in the world (my nephew, Ryan). My dad, Rick, taught me to work hard and never give up. It paid off. Thank you for encouraging me to fight for my dreams. My sister Katrina once told me I was smart and I should be something cool, like a marine biologist when I grew up. Im convinced thats why I became an economist. To Koa, my not so sweet cat, I apologize for neglecting you these past few years. Thanks for your unconditional cuddles and dissert ation supervision. To my in laws, Bill and Nancy, thank you for your constant support and for raising an incredible son who has seen me through this process. My husband, Kyle, deserves much more praise than I can write. You have shown me the true meaning o f love through out this process. Thank you for your patience while I left you for months at a time to do fieldwork. Thank you for your forgiveness when I often forgot to tell you Id be working late. Thank you for telling me I wasnt allowed to quit when th ings got tough. I could not have done this without you. At the end of the day, my endeavors are all in Gods glory. Our Lady of La Leche claimed this dissertation as I defended it on her feast day. She led me on this great journey facilitating my spirit ual and intellectual development in tandem. Thanks be to God for the many blessings He ha s bestowed upon me and for the fruits that will come from this work.

PAGE 8

8 TABLE OF CONTENTS page ACKNOWLEDGMENTS ........................................................................................................... 4 LIST OF TABLES .................................................................................................................... 10 LIST OF FIGURE S .................................................................................................................. 12 LIST OF ABBREVIATIONS .................................................................................................... 13 ABSTRACT ............................................................................................................................. 14 CHAPTER 1 INTRODUCTION ............................................................................................................. 16 2 AN EXPERIMENTAL ANALYSIS OF THE IMPACTS OF A BEHAVIORAL NUDGE AND NUTRITION EDUCATION ON MEAL DIVERSITY ............................... 22 Motivation ......................................................................................................................... 22 Experimental Design .......................................................................................................... 26 Data Collection .................................................................................................................. 31 Empirical Model ................................................................................................................ 34 Results ............................................................................................................................... 39 Conclusion ......................................................................................................................... 44 3 REPEATED EXPOSURE TO A BEHAVIORAL NUDGE IN THE HOME AND NUTRITION EDUCATION: LONGTERM IMPACTS ON DIETARY DIVERSITY ...... 60 Motivation ......................................................................................................................... 60 Experimental Desig n .......................................................................................................... 64 Empirical Analysis ............................................................................................................. 68 Results ............................................................................................................................... 71 Discussion .......................................................................................................................... 74 4 MARKET ACCESS AND NUTRITIONSENSITIVE AGRICULTURE: A CONCEPTUAL FRAMEWORK AND EMPIRICAL FINDINGS ..................................... 93 Background ........................................................................................................................ 93 Food Insecurity in Bangladesh ........................................................................................... 95 Nutrition, Markets and the Agricultural Household Model ................................................. 96 Methodology .................................................................................................................... 103 Data Collection ................................................................................................................ 105 Regression R esults ........................................................................................................... 109 Discussion ........................................................................................................................ 112 5 CONCLUDING REMARKS ............................................................................................ 120

PAGE 9

9 Contributions ................................................................................................................... 120 Findings ........................................................................................................................... 121 Policy Implications ........................................................................................................... 123 Future Research ............................................................................................................... 124 APPENDIX A SUPPLEMENTARY TABLES ........................................................................................ 126 B BASELINE QUESTIONNAIRE ...................................................................................... 138 C ENDLINE SURVEY ........................................................................................................ 176 LIST OF REFERENCES ........................................................................................................ 235 BIOGRAPHICAL SKETCH ................................................................................................... 242

PAGE 10

10 LIST OF TABLES Table page 21 Experimental design: treatment groups .......................................................................... 48 22 Empirical analysis: treatment groups and indicator variables by intervention ................. 49 23 Descriptive statistics: covariates .................................................................................... 50 24 Mean consumption of food items a nd diversity score ..................................................... 51 25 Treatment effects on meal diversity score by weight ...................................................... 52 26 Treatment effects on meal diversity score by volume ..................................................... 54 27 Treatment effects on food item consumption without covariates .................................... 56 28 Treatment effects on food it em consumption with covariates ......................................... 57 31 Pre and post intervention mean 24 hour consumption of food groups by treatment and control+ ................................................................................................................... 78 32 Test of differences in preintervention and post intervention mean 24 hour consumption of food groups by treatment+ ..................................................................... 79 33 At home use of BPP by participants ............................................................................... 81 34 Reason for noncompliance among participants who never use the BPP .......................... 82 35 Estimated marginal effects from Poisson regression of treatments on 24 hour individual dietary diversity score (IDDS)+ ..................................................................... 83 36 Logistic regression marginal effects results of treatment on 24 hour consumption of fruit and vegetable groups without covariates ................................................................ 85 37 Logistic regression marginal effects results of treatment on 24 hour consumption of other food groups without covariates ............................................................................. 86 38 Logistic regression marginal effects results of treatment on 24 hour consumption of fruit and vegetable groups including covariates .............................................................. 87 39 Logistic regression marginal effects results of treatment on 24 hour consumption of other food groups including covariates ........................................................................... 89 310 Estimated treatment effects on 7 day food consumption score (FCS) ............................. 91 41 FAO food groups and examples ................................................................................... 115

PAGE 11

11 42 Descriptive statistics covariates, by district .................................................................. 116 43 Mean household consumption and production of food groups ...................................... 117 44 Marginal effects from Poisson regression on farm diversity score ................................ 118 45 Marginal effects from Poisson regression on household dietary diversity score ............ 119 A 1 Covariate summary statistics by behavior change communication intervention (nutrition education) .................................................................................................... 126 A 2 Covariate summary statistics by behavioral economics intervention (Bengali Portion Plate) ........................................................................................................................... 127 A 3 Summary statistics by treatment groups defined in Table 21 ....................................... 128 A 4 Summary statistics by treatment groups defined in Table 2 2 ....................................... 129 A 5 T test on mean meal diversity score (MDS) by meal for BPP treatment groups ............ 130 A 6 T test for order effects: mean difference in MDSBPP and MDSRegular by order of exposure to the BPP ..................................................................................................... 131 A 7 Coefficient estimates from Poisson regression of treatments on individual dietary diversity score (IDDS) ................................................................................................. 132 A 8 Results from logit regressions on food group production (odds ratios) ......................... 134 A 9 Results from logit regressions on food group consumption (odds ratios) ...................... 135 A 10 Coefficient estimates from Poisson regression on farm diversity score ......................... 136 A 11 Coefficient estimates from Poisson regression on household dietary diversity score..... 137

PAGE 12

12 LIST OF FIGURES Figure page 21 Bengali Portion Plate and regular plate used in the plate intervention ............................. 47

PAGE 13

13 LIST OF ABBREVIATIONS BAU Bangladesh Agricultural University BAUEC Bangladesh Agricultural University Extension Center BBP B engali Portion P late BBS Bangladesh Bureau of Statistics BCC Behavior Change Communication BDT Bangladesh Taka BE Behavioral Economics FAO Food and Agriculture Organization FBDG Food Based Dietary Guidelines F C S Food Consumption Score HDDS Household Dietary Diversity Score HFIAS Household Food Insecurity Access Score IDDS Individual Dietary Diversity Score IFAD International Fund for Agricultural Development INGENAES Integrating Gender and Nutrition within Agricultural Extension Services MaNaR Managing Natural Resources by the Coastal Community MDS Meal Diversity Score NGO Non governmental Organization RCT Randomized Controlled Trial SSC Staff Selection Commission USAID United States Agency for International Development USDA United States Department of Agriculture WFP World Food Programme

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BEHAVIORAL ECONOMICS, NUTRITION EDUCATION, AND ACCESS TO MARKETS: EXPERIMENTAL EVIDENCE AND A THEORETICAL FRAMEWORK FOR IMPROVI NG DIETARY DIVERSITY By Kelly A. Davidson December 2017 Chair: Jaclyn D. Kropp Major: Food and Resource Economics More than two billion people worldwide suffer from micronutrient deficiencies, a problem referred to as hidden hunger. To combat this problem, behavior change communication (BCC) strategies are commonly used in developing countries to change knowledge, beliefs, or values surrounding nutrition. Behavioral economics (BE) is an alternative approach that uses subtle cues to nudge individuals toward healthier food choices. BE strategies are used extensively in the developed world to promote nutrition. However, no study to date has documented the effectiveness of nudges for improving nutrition in development. This dissertation investigates whether a food based dietary guidelines (FBDG) nudge similar to MyPlate and a BCC strategy in the form of nutrition educat ion can improve nutrition in rural Bangladesh. R esearch in the US suggests MyPlate nudges are effectiv e at encouragin g individuals to consume nutrient rich foods In a randomized experiment we measure the individual and com bined treatment effects of the F BDG nudge and nutrition education To measure the short term impacts, we observe participants food choices in a lunch buffet meal where constraints to food access, such as income and availability, are removed. The long term

PAGE 15

15 effects of the interventions ar e measured in the home, where constraints to food access are restored, using pre and post intervention survey data. The meal observation results suggest that the FBDG icon alone does not lead to healthier food choices in the short term; however, combining the FBDG nudge with nutrition education encourages participants to consume a wider variety of foods. Furthermore, repeated exposure to the FBDG icon in the home environment modestly improves the food consumption score, a 7day measure of dietary variety Our findings suggest that other factors besides knowledge may constrain household dietary quality. Hence, we explore the relationships between market participation, production diversity, and household dietary diversity by expand ing Barnum and Squires (1979) agricultural household model to include health, nutrition, and market transaction costs. Empirical analyses suggest a positive correlation between participation in markets for selling products and production diversity, but a negative correlation between participation in markets for selling products and dietary diversity. Thus, household constraints such as market access should be considered in the design of nutrition interventions.

PAGE 16

16 CHAPTER 1 INTRODUCTION T raditional food security initiatives in low income countries have improved farm productivity and profitability. As such, the world has seen great technological advances in the production of cereal s and grains and an increase in the availability of calories per capita (IFPRI, 2011). However, the persistent prevalence of global hunger, malnutrition, and chronic disease signals that increasing production alone will not solve the food security problem (Thompson and Amoroso, 2011). M ore than two billion people w orldwide suffer from micronutrient deficiencies, a problem referred to as hidden hunger ( Iannottie et al., 2009; Micronutrient Initiative/World Bank/UNICEF 2009). In response international development agencies propose a more comprehensive approach to i mproving food systems, one which targets not only agricultural productivi ty but also improves nutrition and health outcomes through nutritionsensitive agriculture initiatives 1 (IFPRI, 2011; Thompson and Amoroso, 2011; Townsend, 2015; World Bank, 2006). A targeted outcome for many nutrition sensitive agriculture initiatives is to encourage the consum ption of a wider variety of food groups as consuming a more diverse diet increases nutrient intake and reduces malnutrition ( Arimond and Ruel, 2004 ; Hatloy et al., 1998, Torheim et al., 2004; Ruel and Alderman, 2013) To facilitate nutrition sensitive agriculture initiatives, a gricultural extension agents or field officers are increasingly tasked with delivering nutrition information alongside technical trainings As such, there is a growing dialogue about the best practices for communicating nutrition guidelines. Development agencies commonly use b ehavior change communication (BCC) to promote health and nutrition in developing countries particularly for child and 1 Agricult ural projects with nutrition -based outcomes are often referred to as nutrition -sensitive agriculture initiatives.

PAGE 17

17 maternal care ( Alderman, 2007; Fitzsim ons et al., 2016; Haider et al., 2000; Linnemayr and Alderman, 2011; Roy et al., 2007; Wong et al., 2014). BCC encompasses a variety of messaging techniques intended to change knowledge, beliefs, and values in an individual or a community. Examples of BCC include peer counseling or community health worker visits. Behavioral economics (BE) is an alternative approach for health promotion that does not rely on changing the knowledge, beliefs or values of an individual but rather implements a strategy to discreetly encourage, or nudge, the individual to choose a healthier option In the United States, behavioral economics interventions are increasingly used to promote healthy eating habits (Hanks et al., 2012; Just e t al., 2007; Just and Payne, 2009; Miller et al., 2016; Thorndike et al. 2014). However, the application of behavioral economics to nudge nutrition decisions in the developing world is novel and evidence of its effectiveness has yet to be documented In a collection of three essays, this dissertation investigates the extent to which the aforementioned tools for behavior change can influence food choice and explores factors that may constrain dietary diversity among agricultural households The essays are set in the context of Bangladesh, a country that continues to battle some of the highest rate s of malnutrition in the world despite impressive levels of economic growth and poverty reduction ( World Bank, 2013). W e conduct ed a randomized controlled trial to measure the potential for BE and BCC strategies to promote diverse, nutrient rich food choices among men and women in rural Bangladesh. Based on the findings from our experiment this dissertation also explore s the role of income and access constraints on household nutrition decisions. Our experiment al analysis measures the impact of two randomly assigned nutrition interventions : exposure to a plate printed with an icon conveying foodbased dietary guidelines (FBDG) and nut rition education in a participatory workshop. The printed plate, which we refer to

PAGE 18

18 as the Bengali Portion Plate (BPP), uses images of food common in the Bengali diet to promote dietary diversity and proper portioning. In its design and intended messaging, the BPP resembles the USDA MyPlate, which is an effective nutrition nudge ( Brown et al., 2014; Miller et al., 2016). The nutrition education intervention, on the other hand, exemplifies a strategy for behavior change communication. Participants were random ly assigned to these two interventions, such that some individuals were exposed to the BPP with no nutrition education some received nutrition education but we re not exposed to the BPP, others received both interventions and those assigned to t he control group did not participate in nutrition education and never s aw the BPP. Thus, our analysis measures the individual treatment effects of the behavioral economics nutrition nudge and the behavior change communication nutrition education intervention a s well as the effects of combin ing these methods. To implement the experiment, we invited men and women from agricultural households to the field office of our partnering agency to partake in a lunch buffet If the participant was assigned to BCC treatment he o r she received nutrition education prior to eating his or her meal. The BPP intervention was introduced during the meal via the plate on which participants served themselves P articipants assigned to the nudge treatment used the BPP and non treated partici pants used a regular plate. We collected meal observation data b y discreetly recording participants food choices during the lunch buffet using data collection methods inspired by U.S. research on behavioral economics in nutrition. T he meal observation dat a are analyzed to evaluate the short term impacts of these interventions on food choices In addition to analyzing the short term effects in the lab setting, we also measure the long term effects of the interventions in the home environment. Each participant who was exposed to the BPP during the buffet meal was given one BPP to take home for each member of his or her f amily. Thus we

PAGE 19

19 analyze p re and post intervention household survey data to measure the long term effects of the BPP and nutrition education interventions on dietary diversity at home. By a nalyzing the meal observation data as well as survey data, we captur e both the short term and long term effects of a nutrition nudge and nutrition education on individuals consumption of healthier more nutrient rich food items. Furthermore, t his unique experimental design allows us to measure the impact s of the two nutri tion interventions both in a lab setting, where the income and access constraints to eating nutrient rich foods are removed, and also in the home environment, where those constraints are restored. C hapter 2 present s the short term, unconstrained effect s of the BE and BCC interventions on food choices and meal diversity in the lab setting when income and access constraints are removed The findings suggest the BPP nudge alone does not affect food choices or increase meal diversity H owever nutrition educa tion and the BPP nudge combined with nutrition education encourage the selection of more diverse foods Chapter 3 examines the potential long term impact s of the BE and BCC interventions by measuring the change in dietary diversity at home. The essay applies a differencein difference approach to baseline and endline survey data. The analysis measures changes to individual dietary diversity controlling for individual, household, and farm characteristics. The baseline survey was conducted at least one day prior to the meal observation, and the endline survey was collected at least one month following the participants meal observation. In addition to individual dietary diversity over a 24 hour period, we also analyze change s to the food consumption scor e to assess potential dietary changes over a longer reference period. The results show little to no evidence that either intervention the BPP nudge or nutrition education impact

PAGE 20

20 24hour dietary diversity at home. However, there is evidence that the BPP nud ge increases the food consumption score, a 7 day measure of dietary diversity. Given the weak evidence that theses nutrition interventions produce long term effects in the home, C hapter 4 investigates the relationship between household dietary diversity a nd access to markets for buying food and selling agricultural products. Two distinct pathways exist through which changes to agricultural practices, inputs, or the food value chain can lead to improved nutrition ( Carletto, 2015; Chung, 2012). Agricultural h ouseholds that increase their farm production can either use additional income from the sale of these agricultural items to purchase food or they can simply increase consumption by consum ing the food s produced on the farm. The latter directly impacts nutr ition through food consumption, whereas the pathway to nutrition through the sale of agricultural products assumes that the additional income will be used to purchase nutritious food ( Carletto, 2015; Chung, 2012). C hapter 4 investigates whether market part icipation influences these pathways from agriculture to nutrition thereby affect ing the demand for various nutrients Specifically, in Chapter 4, we explore the role of market participation in food production and consumption decisions among our participa nt households Our conceptual framework expand s Beckers (196 5) household production and Barnum and Squires (1979) agricultural household model s to include health, nutrition, and market transaction costs where production and consumption decisions are nonseparable. E mpirical analys e s using the baseline survey data examine the relationship between market participation agricultural production diversity and household dietary diversity. These analyses provi de evidence that the off farm sale of agricultural products is positively correlated with farm diversity, but inversely related to household dietary diversity.

PAGE 21

21 The findings presented in this dissertation justify further investigation into the use of beha vioral economics to improve nutrition in a development context The novel protocol from our experiment can be extended to other countries to explore the use of behavioral economics to combat malnutrition and hidden hunger. We also recommend further exploration of the constraints that bind household nutrition decisions in developing countries. Potential p olicy implications of our findings the limitations of the current study, and an agenda for future research are presented in C hapter 5.

PAGE 22

22 CHAPTER 2 AN EXPERIMENTAL ANALYSIS OF THE IMPACTS OF A BEHAVIORAL NUD G E AND NUTRITION EDUCATION ON MEAL DIVERSITY Motivation In South Asia, 336 million people suffer from extreme hunger (World Bank, 2011). However, estimates of food insecurity such as this one do not include micronutrient deficiencies, a problem referred to as hidden hunger and hence underreport the true scope of nutrition challenges (Iannottie, 2009; World Bank, 2011) Bangladesh like many countries in South Asia, has experienced impressive levels of economic growth and poverty reduction, yet it continues to battle some of the highest rates of malnutritio n in the world (World Bank, 2013). While the access and afforda bility of food has increased, a lack of dietary diversity has perpetuated micronutrient deficiencies and hindered progress toward better nutrition. In Bangladesh, r ice consumption accounts for mor e than 70% of daily per capita calorie intake amongst rural households (Ahm ed et al., 2013). Economic theory suggests that increases in income would lead to the consumption of a more diverse bundle of foods as households move away from the consumption of staple foods However, households across Bangladesh consume similar food baskets, with rice being the prominent food item, regardless of income or poverty level (Rabbani, 2014). As a result of limited dietary variety, high rates of vitamin A and iron defici encies persist (World Bank, 2013). I mprov ed dietary divers ity has been shown to increase nutrient adequacy thereby reducing the prevalence of stunting and wasting (Arimond and Ruel, 2004; Hatl oy et al., 1998; Rah et al., 2010; Torheim et al., 2004) Agencies commonly use behavior change communication (BCC) strat egies t o promote dietary diversity, particularly among wome n and children ( Alderman, 2007; Fitzsim ons et al., 2016; Haider et al., 2000; Linnemayr and Alderman, 2011; Roy et al., 2007) BCC includes various messaging mechanisms intended to change knowledge

PAGE 23

23 beliefs, or values (e.g. peer counseling, community health visits, etc.) Behavioral economics (BE) is an alternative approach to behavior change that relies on subtle cues to nudge individuals toward healthier choices rather than changing the knowledge, beliefs or values of the individual. In developed countries behavioral economics interventions such as icons that promote food based dietary guidelines (FBDG), are used to nudge the consumption of nutrient rich foods (Hanks et al., 2012; Just et al., 2007; Kongsbak et al., 2016; Miller et al., 2016; Thorndike et al., 2014). However, no study to date has documented the effectiveness of nutrition nudges in the developing world. This essay investigates whether a FBDG icon can nudge individuals in Bangladesh to diversify their diets such that a higher portion of their calories come from nutrient rich foods instead of rice. In addition, we investigate the impact s of nutrition education as a form of BCC and the effects of combining BCC with a nutrition nudge. Specifically, we estimate the individual and combined treatment effects of two interventions : exposure to a FBDG icon and nutrition education, using regression analysis on meal observation data collected in a cluster randomized controlled trial experiment The FBDG icon in our experiment is a plate designed by the SHIKHA1 project to promote dietary diversity in Bangladesh T he plate is printed with a pictorial diagram of proper meal portioning and nutrition guidelines We refer to this plate as the Bengal i P ortion P late (BPP). The BPP is similar to the USDA MyPlate in that it uses a simple plate diagram to educate people about the national guidelines for nutrition (USDA, 2016; FHI360/USAID, 2016) Using images of locally prepared food the BPP portrays inf ormation to encourage dietary variety, proper portioning, maternal nutrition, and hand sanitation, while informing users about the types of food that constitute a healthy diet. The key message of the plate is to consume a half plate of 1 The SHIKHA project is a USAID initiative facilitated by FHI360 and BRAC to promote dietary diversity and improved nutrition among pregnant a nd lactating women in Bangladesh

PAGE 24

24 rice and at least f our other varieties of food ( FHI360/ USAID, 2015). In 2016, the Bangladesh Ministry of Health and Family Welfare formally adopted the BPP for use in their nutrition education efforts targeting pregnant and lactating women ( FHI360/ USAID, 2016). Although the BPP design targets pregnant and lactating women, the guidelines mirror national dietary recommendations for men and women. In our experiment, the two interventions were implemented in a lab setting at the field office of our partner agency The BPP interv ention was i ntroduced during a lunch buffet whe re participants food choices wer e discreetly observed and nutrition education treatments wer e administered prior to the buffet. This essay analyze s meal observation and survey data to measure the extent to w hich the interventions encourage men and women in rural Bangladesh to make healthier more diverse food choices during a buffet meal when food access and income constraints are removed The results show that the BPP alone has no impact; however combining the BPP with nutrition education is an effective strategy for encouraging more diverse food choices We are specifically interested in whether the BPP can nudge individuals in rural Bangladesh to reduce rice consumption and diversify diets. A nu dge is a subtle cue that influences an individuals behavior without removing that individuals freedom of choice (Thaler and Sunstein, 2009). In the case of encouraging better nutrition, a nudge could be the introduction of a nutritional information image like the MyPlate icon, conveniently packaging healthy food items, or changing the choice architecture by arranging food items in a way that encourages the selection of healthier options. In the US, research suggests that the MyPlate icon effectively nudge s individuals toward healthy food choices (Brown et al., 2014; Mill er et al., 2016). Evaluating the National School Lunch Program (NSLP), Miller et al. (2016) find

PAGE 25

25 elementary and middle school students are 27.7% 15.8% and 16.3% more likely to select the fruit, vegetable, and low fat milk meal components, respectively, when pre ordering their lunch in the morning using a computer program than when ordering in the normal lunch line. The study shows even larger increases in the likelihood of selecting fruits vegetables, and low fat milk of 51.4% 29.7 % and 37.3% respectively, when students are prompted with MyPlate messages during the pre ordering process. Brown et al. (2014) finds similar results exposing college students to the MyPlate via text message i ncreases the consumption of fruits by 13 % and vegetables by 8 % compared to the control group. We apply techniques similar to the above cited literature to test whether participants in Bangladesh can be nudge d toward healthier choices to combat malnutrition N udges may be especially effective in the case of preventative care ( Banerjee and Duflo 2011) Some o rganizations have successfully implemented nudging to promote activities for which households do not receive an immediate reward, but instead receive a future pay off ( e.g bed nets to prevent malaria). Good nutrition is an example of preventative care as m any of the benefits of adopting healthy food choices may only be realized in the future. For example, improving nutrition leads to decreases in child malnutrition, increases in labor productivity, improvements in cognitive development, higher rates of educational attainment, and reduced costs of medical care (Belli et al., 2005). While several studies have examined the use of behavioral economics in nu trition and food policy in the developed world, the application of behavioral economics to combat malnutrition is just starting to gain traction in low income economies. The World Bank is currently implementing a set of behavioral economic experiments in M adagascar to nudge women toward purchasing healthier food ( World Bank 2016a ). Thus, o ur experimental design is

PAGE 26

26 novel as it applies methods used in behavioral economics and nutrition research in the US to the context of a developing country. In its unique design this study develops protocol for future research on nudging nutrition in development. The fusion of behavioral economics with behavior change communication is a unique approach to nutrition interventions in agricultural development. This multifacet ed experimental design allows us to evaluate the individual impac t s of the BPP nudge and nutrition education, as well as the combined effects The results from this study will inform agricultural extension providers and community health workers about the b est practices for promoting nutrition information using FBDG icons. The following section provides the details of experimental protocol and meal observation data collection procedures We then present the empirical model and define the construction of our meal diversity index which we use to measure the diversity of food choices selected during the buffet meal Our results are then presented, followed by a discussion of our findings implications of the findings, and further research. Experimental Design To facilitate the recruitment of participants into our field experiment, we partnered with two separate institutions : the Bangladesh Agricultural University (BAU) and Shushilan, a local NGO. The research was conducted in two project areas where the Bangladesh Agricultural University Extension Center (BAUEC) and Shushilan provide agricultural extension services to rural households. For the purpose of this research we engaged Shushilan beneficiaries from the Managing Natural Resources by the Coastal Communi ty (MaNaR) project. BAUEC and MaNaR were selected as partner programs based on the criteria that the beneficiaries had not received prior nutrition training. The research was conducted in the Mymensingh and Borguna districts, the beneficiary sites for BAU EC and Shushilan MaNaR, respectively. Mymensingh is located in the northern part of Bangladesh, while Borguna offers a stark contrast in southern

PAGE 27

27 Bangladesh. The BAUEC farmer membership is 55 % male and 45% female and spans across 20 unions in one upazila2. Membership is open to farmers in the communities surrounding BAU BAUEC operates as a traditional extension program, disseminating technical agricultural advice for activities such as livestock vaccination. The beneficiary roster for MaNaR is 94 % female a nd 5% male, and the project encompasses three unions in one upazila. B eneficiaries in the MaNaR project were selected on the criteria of extreme poverty and limited access to land or capital. The project promotes production activities such as vegetable mul ti cropping and floating gardens while educating beneficiaries on strategies for climate change resilience. Our target number of participants totaled 1,200 participating in two meal observations each, for a total sample of 2,400 meal observations. Stratif ying by location, we targeted 600 randomly selected participants from each district. The sample included 18 communities in the S h ushilan project area and 35 communities in the BAUEC project area, where the communities are smaller. Following a fully factor ized randomized controlled trial (RCT) design, participants were randomly assigned to a combination of two interventions : 1) exposure to the BPP and 2) nutrition education. The BPP intervention was introduced as the plate on which participants served thems elves during a lunch buffet. The education treatments were implemented through a participatory workshop prior to the lunch buffet. Table 21 outlines the possible combinations of the BPP and nutrition education treatment assignments. All participants were invited to two lunch buffet meals one month apart, where their food choices were discreetly observed. W e invited each participant to two meals to observe individual preferences and eating patterns This was particularly important since the participants were not accustomed to selecting food from a buffet. Individuals assigned to one of the nudge treatment s used the BPP during one of the two 2 An upazila is equivalent to a county -level administrative area ; a union is the sub -county level comprised of a cluster of villages

PAGE 28

28 meals, while the control group used a regular plate without the BPP diagram at both meals We varied the order of expo sure to the BPP nudge to test for potential order effects, hence the two meal design. In the absence of order effects, the observations can be pooled for analysis. As depicted in T able 2 1, some p articipants ( treatment B ) used a regular plate during the f irst meal observation and the BPP i n the second meal observation. Other p articipants (treatment C) used the BPP in the first meal observation and a regular plate in the second meal observation. The remaining p articipants ( treatment A ) used a regular plate for both buffet meal observations an d were never exposed to the BPP. P articipants in treatment A serve as the control for the BPP intervention. All of the plates used in the experiment, both the regular plate and the BPP, were ordered from the same factory made of melamine, and equivalent in size Although the plates have the same background color, the BPP was printed with the proper portioning diagram while the regular plate was printed with flowers (F igure 21). The flower print pat tern resembles a standard plate commonly used in Bangladesh. Since this experiment was part of a larger study on the effectiveness of the BPP on dietary diversity, participants assigned to a BPP treatment were also given BPPs to take home. Each participa nt received one BPP for each family member in his or her household Researchers distributed these BPPs at the end of the meal in which the treated participant was first exposed to the BPP To reduce potential spillover effects and contamination that would occur if participants receiving the BPP showed the plate to participants in the control group, the BPP intervention was cluster randomized at the community level. All participants in a particular community were assigned to the same plate treatment. The nut rition education intervention was randomly assigned at the individual level. In T able 21, group 1 serves as the control group P articipants assigned to group 1 were not exposed to the BPP at either meal observation and received no

PAGE 29

29 nutrition education. To satisfy the BCC component of this experiment, each participant was assigned to one of two nutrition education treatments or the control : 1) no nutrition education, None which served as the control 2) nutrition education, N, or 3) nutrition education with an additional gender component, NG. Participants received the same BCC assignment, if any, at both meal events. To implement the experiment, all participants were invited to the field office of the ir respective partner organization on two occasions, approximately one month apart, for lunch and the nutrition workshop if the participant was assigned to a BCC treatment. Each meal was served at 1:30pm, the typical lunch hour in Bangladesh. Ten food items corresponding to the BPP food groups were arranged in a buffet line: rice (cereals), chicken (meat), fish, mixed vegetables (white potatoes and tubers; orange vegetables), leafy green vegetables, cucumber and onion salad (other vegetables), dal (lentils), boiled eggs, bananas (fruit), and yogurt (dairy). A local chef w as hired in each study location. The same chef prepared all dishes in the same fashion according to local cuisine, for each meal throughout the duration of the study The items were arranged in the same order in both locations for each meal, and the menu remained the same over the course of the study. E numerators discreetly observed individuals food choices on each occasion as participants served themselves from the lunch buffet. Participants were encouraged to take as much food as they desired. If the participant was assigned to one of the two nutrition education treatments he or she attended the respective participatory workshop before the meal. Facilitators who w ere trained to follow the methods described in the Introductory Workshop on Integrating Gender and Nutrition within Agricultural Extension Services (INGENAES) Facilitators Guide by Henderson (2016) conducted the workshops. Specifically, the nutrition educ ation component followed the What

PAGE 30

30 Goes on the Plate activity, where small groups illustrated and discussed the components of a healthy diet. The session concluded with a presentation of the national food based dietary guidelines. The concluding informati on corresponded to the messages on the BPP ; however the facilitator did not explicitly introduce the BPP during the training. Participants who were assigned to the NG treatment also participated in the Who Gets What to Eat exercise from Henderson (2016) The activity facilitated a role play on intra household food distribution and invoke d discussion on womens nutrition and the cultural norms surrounding intr a household allocation of food. Each participant was assigned the role of a family member in the household (i.e. husband, wife, mother in law, son, etc.) The person assigned to the wife role was given a variety of food item props to reflect a household meal. T he facilitator then ask ed the person assigned to the wife role to distribute the food items to the other participants according to their assigned family member roles following s ocial and cultural norms in their community. Once the food ha d been distributed, the facilitator led participants in a discussion about gender based dietary needs and the importance of allocating nutrient rich food, such as protein, among all members of the household. Individuals assigned to the N and NG education treatment groups arrived at the research site at 10:00am. All workshop attendees participated in the same nutrition education session, the What Goes on the Plate activity. Individuals assigned to treatment N received this training only and were then asked to leave the training room for a break. Individuals assigned to treatment NG stayed for the gender c omponent of the training. On the day of the workshop, the treatment groups were denoted by colored nametags, prepared in advance by the facilitators. The workshop concluded at approximately 1:00pm. Participants who were assigned to the nutrition education control group, None, did not receive nutrition training. T he se

PAGE 31

31 participants were invited to come to the research site at 1:00pm, after the participatory workshop was finished. All participants were invited to eat the buffet meal at 1:30pm, where the BPP experiment was implemented. Individuals received the same assigned education treatment and followed the same schedule for the second meal observation one month later. The nutrition trainings were repeated because frequent exposure to nutrition messages reinforces the information (Brown et al. 2014). Data Collection The data collect ed for this study were gathered through a baseline survey, the field experiment incorporating the two meal observations, and a short post meal survey following each meal. The baseline survey contained questions on household demographics, agricultural production, household dietary diversity, and prior nutrition knowledge. The survey was conducted as a faceto face interview at the respondents home by trained enumerators. All interview s w ere conducted in Bengali, the national language. Each participant was surveyed at least one business day prior to his or her visit to the field office for the first meal observation The survey enumerator invited the participant to attend two meals and the nutrition workshops (if applicable) as compensation for his or her ti me spent completing the survey. The participant was given two meal tickets with the dates, location, and time to arrive at the field office. The meals served as partial compensation for the participant s time, but also provided an opportunity to collect meal observation data. Participants were also given a small monetary stipend, 100 BDT3 in Borguna and 400 BDT in Mymensingh, to cover their lost wages and transportation costs to the field office at the end of each meal. The respective stipend amounts reflect the differences in local wages and standard rates for research participation across the two districts. 3 US $1 is approximately 78 BDT

PAGE 32

32 During each meal, enumerators recorded the amount of each food item a person took from the buffet. When the participant had finished eating, enumerator s also documented the amount of waste of each food item left on the plate. Selection and plate waste of n ine of the ten food items on the lunch buffet were measured using visual inspection This method of food measurement is frequently used to measure plat e waste in school cafeterias and other institutional setting, and has also been used in various nutrition intervention evaluation s ( Buzby and Guthrie, 2002; Friedman and Hurd Crixell, 1999). Visual inspection has been validated as a cost effective, reliabl e alternative to weighted measurement (Buzby and Guthrie, 2002; Hanks et al., 2014; Richter et al., 2012 ; Shatenstein et al., 2002 ). Our method of visual inspection is an extension of the quarter waste method also known as the Comstock method. Under the quarter waste method, the researcher estimates the proportion of an items standard serving size that remains on the food tray or plate following consumption The estimated proportion of a wasted food item is recorded in quarter increments using a five po int scale ( e.g. 0, , or 1; where 0 indicates that nothing is waste and 1 indicates that the entire serving is wasted ) (Buzby and Guthrie, 2002; Comstock et. al., 1981; Hanks et al., 2014; Kropp et al., 2017). The method is typically used in cafeteri a settings where portion sizes are standardized ; however Richter et al. (2012) also validated the reliability of visual inspection for nonstandard portions. Our data collection methods apply the Comstock scale to spoonfuls of the food items selected by th e participant rather than standard servings Since we were interested in selection decisions, we allowed participants to serve themselves from the buffet. Identical serving spoons were used for all food item s except the fruit offering (banana), which was m easured on a piecebasis. Thus one spoonful constituted a standard serving. Data collectors were thoroughly

PAGE 33

33 trained to visually estimate spoonfuls of each item on a five point scale (0, , or 1 ). As is standard practice when applying the Comstock method, all of the data collectors participated in establishing measurement standards. Specifically, the data collectors weigh ed five spoonfuls of each food item at the beginning of each meal observation day t o obtain a baseline measure for comparison In this process, the researchers agreed on the visual estimation of quarter spoonfuls for each item and o ne researcher recorded t he weights of the five standard spoonfuls for each food item. For the analysis, the standardized weight of a spoonful of a food item is calculated as the average weight of a spoonful on the observation day During the meal, the enumerators recorded the number of spoonfuls of each food item taken by a participant in quarter increments (0, 0.25, 1.25, etc.) A zero indicated that the food item was not selected The method allowed for cases in which the participant took multiple spoonfuls of an item One of the main intended outcomes of the BPP and nutrition education treatments was to reduce rice consumption. Thus, to get a more precise measurement, rice selection was measured by weight using a hidden food scale placed under the serving dish for rice. Specifically, the platform of the scale was placed under the serving dish and t he digital display faced away from the participant. When the participant finished selecting rice, the enumerator recorded the change in weight of the serving dish. The researchers used electronic tablets to record all data for all food items Selection information was link ed to the respective participant using his or her identification number which w as noted on the participant s nametag To ensure consistency in the visual inspection for each food item, one observer was assigned to record selection of a particular food item. Enumerators remain ed the same throughout the study period and were thoroughly trained at practice meals to ensure inter rater reliability.

PAGE 34

34 After the participant finished eating, he or she completed a short post meal questionnaire about 1) how hungry he or she was before the meal and 2) whether he or she disliked any of the food items on the buffet. At the second meal observation, the post meal questionnaire also included the four week household food insecurity questions to account for the fact that more than four weeks had p assed since the baseline questionnaire. Empirical Model To determine whether these interventions nudge participants toward a more diverse plate of food, we construct an outcome variable to measure participants meal diversity, which we refer to as the mea l diversity score (MDS). The meal diversity score is modeled after the Simpsons index, which is commonly used to measure biodiversity and has also been applied to farm diversity (Jones et al., 2014). MDS is defined as where is the portion, or share, of the individuals plate allocated to food item j The portion is calculated as where is the weight (kg) of food item j consumed by the individual, i and is the total weight (kg) all food items ( ) consumed by the individual such that As a robustness check we also calculate in terms of the volume of each food item consumed relative to the total volume of food consumed at the meal. This latter measure accounts for the fact that different food offerings have different densities F or example a quarter of a plate full of boiled eggs weighs more than a quarter of a plate full of rice. To calculate the portion of a food item consumed in terms of volume, where is the volume of a particular item, j consumed by individual i and is the sum of the volume of all food items ( ) consumed by individual i such that In both cases, the MDS takes a value between

PAGE 35

35 [0,1], where a smaller number represents low meal diversity and a larger number represents high meal diversity. Prior to regression analysis, we verify the absence of order effects from the two meal experimental design. W e use a paired t test to determine if the mean difference in MDSBPP (the MDS when a participant is exposed to the BPP) and MDSRegular (the MDS when the same participant is exposed to the regular plate) is statistically significantly different for participants in plate treatment B versus C. If the test indicates that the mean difference i s not statistically different by treatment groups, then we can conclude that order effects are not present and the observations from both meals can be pooled. Pooling the meal observations reduces our treatment and control groups to the six combinations pr esented in T able 22. The following equation represents the model to be used to estimate the impacts of the nutrition education and BPP interventions on the meal diversity score for individual i during meal m assuming that order effects are not present. (2 1) In Equation (2 1) BPP is a dummy variable that takes the value of 1 if the individual was given the BPP to use during the meal observation but received no nutrition education and 0 otherwise N and NG are dummy variables in dicating whether the individual received nutrition education or nutrition education with a gender component, respectively, but was not exposed to the BPP during the meal observation The estimated value of 1 represents the effect of the BPP nudge on MDS when the participant is nudged by the BPP alone, 2 is the impact of nutrition education on MDS and 3 is the impact of nutrition and gender education on MDS Through interaction terms, 4 and 5 measure the combined effects of the BPP nudge with the nut rition or nutrition and gender education, respectively.

PAGE 36

36 The model also includes district level, household level, and individual level explanatory variables contained in Xi. We control for the district where the participant lives using a dummy variable equ al to 1 for the Mymensingh district and 0 if the participant lives in the Borguna district. The district covariate captures general differences in socioeconomic status, tastes, and preferences between the two geographical districts. At the household level we include covariates for household monthly income, a poverty scorecard index, household food insecurity access score (HFIAS), and farm diversity to control for exogenous factors that affect food preferences such as familiarity with various food items. Hou sehold monthly income is a continuous variable measured in the local currency, Bangladeshi Taka (BDT). The poverty scorecard index follows Schreiner (2013), taking a value from 0 to 100 to quantify the likelihood that household expenditures are below the n ational poverty line for Bangladesh. A lower poverty scorecard index represents a higher likelihood that a household falls below the poverty line. The household food insecurity access score (HFIAS) measur es participants perceptions of food vulnerability a nd responses to food insecurity (Coates et al., 2007). The HFIAS is a continuous variable rang ing from 0 to 27, where a higher score represents a higher degree of food insecurity. Farm diversity is a count variable measuring the number of food groups a hou sehold produces, where individual crops are classified in food groups following the FAO guidelines for dietary diversity (Kennedy et al., 2010). Studies have shown farm diversity is positively correlated with household dietary diversity (Jones et al., 2014). W e expect an individual from a household that produce s a larger variety of food may have a higher meal diversity score if he or she is accustomed to eating diverse meals. Cultural norms such as religion and gender inequities can dictate individuals con sumption of food items. For example, women in Bangladesh generally consume less protein

PAGE 37

37 than men, and traditional hierarchies influence the allocation of food items within the household (Ahmed et al., 2013). Thus, for each individual we also include covari ates for religion, sex, age, and relationship to the head of household. Religion is a dummy variable equal to 1 if the participant ident ifies as Muslim and 0 otherwise Muslim is the majority religion in Bangladesh. The sex of the individual is a dummy var iable equal to 1 if the individual is female and 0 if the participant is male. Age is a continuous variable. The relationship to the head of household is a set of categorical variables for the individuals position in the family equal to 1 when true and 0 otherwise where the categories include household head, spouse of household head, and other. Household head is the base category and is omitted from the regression analysis Education is included as a set of dummy variables where 1 indicates the highest l evel of education that the individual has completed (none, primary, secondary, junior secondary, SSC pass, or postsecondary) and 0 otherwise, where no education is the omitted category. We expect more educated individuals to be familiar with the health ben efits of consuming a variety of food items and be more receptive to the nutrition messages. We also include covariates for baseline nutrition knowledge and hunger at the time of the meal. Nutrition knowledge is a continuous score [0,36] based on responses to a set of questions about dietary recommendations, nutrient content of familiar food items, diet disease relationships, and child and maternal nutrition. The questions follow the validated methods of Parmenter and Wardle (1999). Hunger at the time of the meal is a Likert scale variable [1,5] ranging from not hungry at all to extremely hungry. Observations about hunger were collected after the participant had finished his or her meal. In addition to measuring the treatment effects on meal diversity, we also seek to understand how the BPP and nutrition education interventions impact individual consumption behavior for specific food items. Our experimental design provides a unique opportunity to

PAGE 38

38 investigate which types of food items and how much of each food item participants consume, given nutrition information, when income and food access constraints are removed. Thus, in addition to the empirical analysis on meal diversity, we also meas ure the impacts of the BPP and nutrition education interventions on the consumption of each of the ten food items using linear regressions. The general form of the food item consumption model is: (2 2) where yijm is the amount (kg) of food item j consumed by individual i during meal m The treatment variables and covariates are the same as defined in E quation ( 21). Thus, the estimated value of 1 represents the effect of the BPP nudge on the consumption of food i tem j when the participant is nudged by the plate alone, 2 is the impact of nutrition education, N on the consumption of food item j and 3 is the effect of nutrition education with a gender component, NG on the consumption of food item j The combined effects of the BPP nudge with N and NG are measured through the estimated values of the coefficients on the interaction terms, 4 and 5, respectively. We hypothesize that the BPP will be most effective at improving meal diversity when combined with nutri tion education. We also estimate the impact of adding a gender component to the training, which we expect will have a larger impact on MDS than nutrition education alone since basic dietary recommendations are reiterated during the discussion about intrahousehold allocation of food. We expect participants in the BPP treatment group will consume more nutritious food items and less rice than the participants who were not e xposed to the plate. Following our hypotheses on meal diversity, we also expect the impact on the consumption of nutritious food items to be largest when the plate is combined with nutrition or nutrition and gender education.

PAGE 39

39 The models are estimated usin g OLS regression. Due to the cluster randomized design, we cluster our standard errors at the community level. However clustered standard errors may result in downward biased estimates and over rejection of the null if the number of clusters is small ( Cam eron and Miller, 2014; Cameron et al., 2008 ; Duflo et al., 2008; Woo ldridge, 2004). With 53 unbalanced clusters divided into five treatments and the control, we have potential ly too few clusters. Following Mackinnon and Webb (2017) we estimate wild cluster bootstrapt p values to correctly test our hypotheses given the few clusters problem. Results Table 2 3 presents the descriptive statistics for all participants by district. Our sample includes 1,105 individuals in the first meal, 1,077 of whom returned f or meal two, for a total of 2,182 meal observations. Thus, the attrition rate was approximately 2.5 % We use a balanced sample of 2, 227 observations in all estimation results. The Mymensingh district is slightly overrepresented with 54 % of participants res iding in this district. Our sample is 70 % female, primarily due to the woman centric nature of the Shushilan MaNaR project. When possible, we recruited male members of the Shushilan beneficiary households. The mean household income is 11,001 BDT, or approximately USD $140, per month F ollowing the estimates of household poverty likelihoods in Schreiner (2013), the mean poverty score of 4 9 implies that 33.5 % of the sample lives on less than USD $1.25 per day. Participants from the Borguna district have a hig her mean poverty score and higher mean HFIAS than participants in Mymensingh. Thirty six percent of participants have no formal education. This is not surprising since we recruited participants through organizations who serve rural households. On average, the households in our sample produce 3.4 different food groups. To account for spiritual norms surrounding food consumption, we also asked participants to identify the religious affiliation of their household. The majority of households in Bangladesh ident ify as Muslim, which is reflected in our sample.

PAGE 40

40 On a scale of 36, t he average nutrition knowledge score is 18.9, just over 50% T his confirms that our sample had a low comprehension of nutrition recommendations prior to the field experiment. Tables A 1 th rough A 4 in A ppen dix A present further analyses of summary statistics for individual and household characteristics by treatment group. Table A 4 also contains Pearsons Chi squared statistics which test the differences in covariate distribution between treatment groups. Table 24 shows mean consumption (kg) of each of the ten food items per meal As one would expect according to the traditional Bengali diet, rice is the most consumed food item. On average, participants consumed 0.38 kg of rice per meal This is consistent with the amount of rice one would expect participants to consume if the lunch buffet is the primary source of food for the day. The average daily per capita rice consumption in Bangladesh is 495.5 grams, or 0.495 kg (Ahmed et al., 2013). Since we informed participants at least one day prior that a meal would be served at the event, we expect ed the participants would consider the lunch buffet as their primary source of calories for the day. As noted in the methodology section, our model co ntrols for the individuals level of hunger at the time of the meal. Table 23 shows the average hunger score was 3.7, or very hungry. By design, this field experiment removes income and access constraints to nutritious food items such as meat and dairy, which may be too expensive for many of the participants to purchase for home consumption. It is not surprising then, that the mean consumption of protein sources, chicken, fish, yogurt, and lentils, are also relatively high. The average meal diversity sco re, MDS, is 0.77, which indicates that participants consumed relatively diverse meals at the experiment. Participants consumed 7.8 different food items on average.

PAGE 41

41 W e conduct a cluster adjusted paired t test to verify the absence of order effects. The t te st is clustered since BPP treatment was cluster randomized at the community level. The clustered t test results confirm that the mean difference in MDSBPP and MDSRegular is not statistically significantly different for participants in plate treatments B ve rsus C (p = 0.179; T able A 6). Thus, we determine no order effects exist, and we analyze the pooled meal observation data. Table 25 presents the results of the regression analyses on the meal diversity score by weight. The table reports standard errors c lustered by community as well as w ild cluster bootstrap t p values. The coefficients on the education treatment variables, N and NG, as well as the interaction between the BPP treatment and nutrition education, BPP x N are positive and statistically significant. These results suggest t he individuals who received nutrition education ( N ) consumed a larger variety of food items than participants that did not receive nutrition education and were not exposed to the BPP T he MDS w as 0.016 higher on average for this group a 2.1 % increase on the mean. The MDS for participants in the NG treatment was 0.017 higher on average than participants that did not receive the education treatment and were not exposed to the BPP Although the treatment effect of the comb ined nutrition and gender ( NG ) education is larger in magnitude than the effect of nutrition education alone ( N), a Wald test on the two parameters shows there is no statistically significant difference in the coefficients (p = 0.4371). Confirming our hypo thesis, the largest treatment effect comes from the combination of the what goes on the plate activity in the nutrition workshop and exposure to the BPP. Individuals received nutrition education and were also exposed to the BPP during the lunch buffet co nsumed a more diverse meal on average than the participants in the control group O n

PAGE 42

42 average their MDS were 0.0 2 points higher, a 2. 6 % increase on the mean. However, the estimated coefficient on the interaction between exposure to the BPP and the combined trainings on nutrition and gender, BPP x NG is not statistically significant. This is likely due to the baseline level of nutrition knowledge of participants in the NG treatment group. Although the interventions were randomly assigned, the group assigned to nutrition and gender training had a higher baseline understanding about nutrition according to the nutrition knowledge assessment scores. A Pearsons chi squared test confirmed the statistically significant difference in nutrition knowledge between nutr ition education treatment groups. Thus, the effect of prior nutrition knowledge of these participants captured in the nutrition knowledge assessment scores potentially dampened the effect of the effect of BPP x NG rendering it insignificant. While the effects of the nutrition education treatments fit our hypothesis, the BPP alone did not statistically significantly impact meal diversity. We suspect the plate alone is not effective because the key messages on the plate are written in Bengali ( F igure 21). While we did not measure participant literacy directly, our sample is largely uneducated and hence likely illiterate. Therefore, it is reasonable to assume participants who have had no formal education or only completed primary school are unabl e to interpret the written messages on the BPP. The food images on the BPP were not sufficient to nudge individuals toward more diverse food choices. However, when the key messages we re verbally communicated during the participatory training on nutrition, the plate nudge d participants to consume a more diverse meal. While the results of the paired t tests imply that order of exposure to the BPP did not affect the plates impact on MDS, we repeat the regression analysis with the inclusion of a dummy variable indicating the participant was exposed to the BPP at the second meal as a ro bustness check. Model 3 in T able 25 shows the indicator variable for order of exposure is not

PAGE 43

43 statistically significant, verifying the results of the paired tests. Thus, we confirm no order effects exist. Table 2 6 present s the regression results when MDS is constructed based on the relative volume of each food item consumed to the total volume of food consumed. The results are similar in magnitude and statistical significance, verifying the robustness of our model. At the aggregate level, we are interested whether participants consume a more diverse set of foods given nutrition information. Howe ver, this study also employs a unique approach to understand consumption behavior with respect to specific food items. The increase in meal diversity may be driven by a preference for certain nutrients or food items. Thus, in addition to the analysis on me al diversity, we also evaluate the treatment effects on the consumption of each of the ten food items. Table 27 presents these regression results and T able 2 8 presents the regression results with covariates Contrary to our hypothesis, the results show no statistically significant treatment effect s on rice consumption. The consumption of lentils (dal) is 0.014 kg higher on average among individuals who were assigned to the nutrition education treatment and 0.016 kg higher on average for participants who received nutrition education with the gender component compared to those who did not receive nutrition or gender education and were not exposed to the BBP. Additionally, participants who used the BPP and received nutrition education, with or without the g ender component, consumed more lentils on average compared to the control group, 0.02 9 kg and 0.020 kg, respectively. Furthermore, the average consumption of leafy green vegetables is slightly higher, 0.00 5 kg, for participants who received education treat ments compared to those who did not receive nutrition or nutrition and gender education and were not exposed to the BPP T here is no statistically significant effect however, on the

PAGE 44

44 consumption of leafy green vegetables when the trainings are combined wit h the BPP intervention. The results of the regressions on specific food items suggest that the treatment effects on meal diversity may be driven by lentil consumption. This may be due to the pronounced image of lentils on the BPP. Alternatively, it is als o possible that our participants were more familiar with lentils as a source of protein compared to other food items, as lentils are a more affordable source of protein and hence might be consumed in the home more frequently than other protein sources. Co nclusion Several development initiatives aim to promote nutrition among agricultural households in Bangladesh and the greater South Asia region. Food based diagrams and food plate tools are especially popular instruments to communicate nutrition informatio n I n certain cases these tools can be used to nudge individuals toward healthier food choices as recent studies have shown their effectiveness in the US ( Brown et al., 2014; Miller e t al., 2016). However, initiatives for communicating any type of behavio r change can be costly in developing countries. For an initiative to be successful, it is critical that the characteristics of the target population be taken into account in the design and implementation of behavioral change strategies. Furthermore, eviden ce based pilot testing and impact evaluation should be implemented prior to scaling up the initiative. In this study we conduct ed an experiment using meal observation data to measure the impacts of a melamine plate printed with Bengali food based dietary guidelines on food choice and meal diversity. In addition to the BPP, a behavioral nudge our study measure d the effect s of behavior change communication through participatory education on nutrition and gender norms surrounding nutrition. The results of this randomized controlled trial show that exposure to the

PAGE 45

45 printed BPP alone is not enough to encourage individuals to consume a more diverse diet. However, when the BPP is combined with nutrition education individuals consume a more diverse meal. In Bangladesh, nutrition guidelines especially emphasize a reduction in rice consumption. Thus, we also use d the meal observation data to test the effect of the BPP and education interventions o n rice consumption as well as the consumption of other food items. W e f ound the interventions are not successful at decreasing rice consumption A caveat to this research is that our sample is a subset of the membership rosters for our partner organizatio ns. A nationally representative sample would allow for a more robust analysis of the treatment effects to better inform the scaleup ability of the tested interventions. Furthermore, the preexisting relationship between participants and the respective par tner organization may influence these results. Participants are potentially more receptive to the information treatment due to their relationship with and trust in the partner organization. The fact that participants were compensated to complete the study also generates potential bias in choices, as the level of compensation might bias the extent to which they internalize the program information. Further research should explore the role of trust in nutrit ion information dissemination. This study employed a novel approach to evaluating nutrition sensitive initiatives in agricultural development. Drawing upon research from the developed world, this randomized controlled trial used meal observation data collection methods to investigate the use of behavioral ec onomics strategies for improved nutrition in developing countries. The methods used in this study can be extended to other countries to evaluate similar strategies for nudging individuals to improve nutrition and dietary diversity.

PAGE 46

46 In addition to research protocol, this study offers direct policy implications. Several organizations in Bangladesh plan to use the BPP in their nutrition programs, some of which emphasize agricultural interventions ( FHI360/ USAID, 2016). Thus, this study specifically target ed hou seholds in rural Bangladesh to evaluate the effectiveness of the BPP as a tool for improving dietary diversity in agricultural communities. Low levels of education characterize our target population Assuming education is a proxy for literacy, we suspect l ow literacy levels in our sample influence our findings, which show that the BPP is not an effective nudge toward dietary diversity. The key messages on the BPP are written in Bengali, thus the information might not be conveyed to illiterate participants u nless it is verbally communicated. The findings suggest participatory training at the community level is an effective approach to communicating information where literacy is a constraint. Thus, we encourage policy makers to consider the appropriateness of nutrition training tools for the education level and literacy of the target population. Further research is necessary to determine if the results would differ if participants had a higher rate of literacy. The evidencebased results from this field experim ent inform government and nongovernment agencies in Bangladesh about the effectiveness of the BPP. Agencies in other developing countries can replicate the methods used in this study to pilot test similar nutrition interventions

PAGE 47

47 (a) (b) Figure 21. Bengali Portion Plate and regular plate used in the plate intervention Messages: Half plate of rice and at least four other varieties of food, Eating a variety of food in appropriate amounts keeps mothers and children healthy, Eat a little more f ood during pregnancy and Wash both hands and soap with running water before preparing and eating food. Source: FHI360 / USAID SHIKHA Food Plate (Bengali Portion Plate)

PAGE 48

48 Table 21. Experimental design: treatment groups Behavioral Economics Interventio n Behavior Change Communication Intervention Treatment (Meal 1, Meal 2) No Nutrition Education Nutrition Education Nutrition and Gender Education Treatment A (Regular, Regular) Group 1 Group 2 Group 3 Treatment B (Regular, BPP) Group 4 Group 5 Group 6 Treatment C (BPP, Regular) Group 7 Group 8 Group 9

PAGE 49

49 Table 22. Empirical analysis: treatment groups and indicator variables by intervention Behavioral Economics Intervention Behavior Change Communication Intervention No Nutrition Education Nutrition Education Nutrition and Gender Education Regular plate (Control) N NG BPP BPP BPP x N BPP x NG

PAGE 50

50 Table 23. Descriptive statistics: covariates All Borguna Mymensingh VARIABLES Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Nutrition education 0.23 0.42 0.22 0.415 0.23 0.42 Nutrition and gender education 0.22 0.41 0.22 0.413 0.22 0.42 BPP only 0.11 0.31 0.12 0.319 0.10 0.30 BPP x nutrition education 0.13 0.33 0.13 0.337 0.12 0.33 BPP x nutrition and gender education 0.12 0.32 0.12 0.322 0.11 0.32 Mymensingh district 0.54 0.50 Household monthly income (BDT) 11001.44 9466.80 6159.34 4885.36 15130.50 10425.22 Poverty score [0,100] 48.59 16.00 38.27 11.04 57.42 14.20 Number of food groups produced [0,12] 3.40 2.23 2.16 1.83 4.47 1.98 Household food insecurity access score [0,27] 1.92 3.68 3.53 4.40 0.53 2.09 Religion (Muslim = 1) 0.95 0.21 0.91 0.28 0.99 0.11 Female 0.70 0.46 0.92 0.27 0.52 0.50 Age 38.03 12.50 38.54 11.67 37.59 13.17 Primary school (highest level) 0.33 0.47 0.44 0.50 0.24 0.43 Junior secondary school (highest level) 0.11 0.31 0.06 0.24 0.15 0.36 Secondary school (highest level) 0.06 0.24 0.02 0.14 0.10 0.30 SSC pass (highest level) 0.06 0.23 0.01 0.08 0.10 0.30 Postsecondary school (highest level) 0.07 0.25 0.00 0.06 0.12 0.33 Spouse of household head 0.51 0.50 0.66 0.47 0.39 0.49 Other relationship to household head 0.13 0.33 0.06 0.23 0.19 0.39 Hunger at time of meal event [1,5] 3.71 1.17 3.13 1.16 4.21 0.93 Baseline nutrition knowledge score [0,36] 18.86 5.10 17.38 5.23 20.12 4.63 N = 2,227

PAGE 51

51 Table 24. Mean consumption of food items and diversity score Dependent variable Mean St. dev. Rice (kg) 0.38 0.17 Chicken (kg) 0.08 0.04 Fish (kg) 0.08 0.04 Salad (kg) 0.05 0.05 Egg (kg) 0.06 0.02 Mixed vegetables (kg) 0.07 0.05 Leafy green vegetables (kg) 0.05 0.04 Lentils (dal) (kg) 0.09 0.09 Fruit (banana) (kg) 0.07 0.05 Yogurt (kg) 0.08 0.06 Meal diversity score by weight [0,1] 0.77 0.08 Meal diversity score by volume [0,1] 0.75 0.08 Number of food items consumed [0,10] 7.80 1.54 N = 2,22 7

PAGE 52

52 Table 25. Treatment effects on m eal diversity score by weight (1) ( 2 ) (3 ) VARIABLES MDS (weight) MDS (weight) MDS (weight) Nutrition education 0.016*** 0.013*** 0.013*** [0.005] [0.005] [0.005] p<0.001 <0.014> <0.020> Nutrition and gender education 0.017*** 0.016*** 0.016*** [0.005] [0.005] [0.005] <0.002> <0.006> <0.004> BPP intervention 0.001 0.004 0.005 [0.010] [0.007] [0.007] <0.932> <0.616> <0.462> BPP x nutrition education 0.020* 0.016* 0.014* [0.010] [0.008] [0.008] <0.076> <0.054> <0.098> BPP x nutrition and gender education 0.010 0.007 0.006 [0.010] [0.007] [0.007] <0.312> <0.276> <0.350> Mymensingh district 0.045*** 0.045*** [0.009] [0.009] <0.002> p<0.001 Household monthly income 2.94E 07 3.06E 07* [1.54E 07] [1.58E 07] <0.078> <0.104> Poverty score 1.66E 04 1.66E 04 [1.44E 04] [1.44E 04] <0.252> <0.284> Number of food groups produced 0.001 0.001 [0.001] [0.001] <0.648> <0.674> Household food insecurity access score 2.07E 04 2.23E 04 [0.001] [0.001] <0.694> <0.648> Religion (Muslim / non Muslim) 0.070*** 0.069*** [0.019] [0.019] <0.048> <0.052> Female 0.006 0.006 [0.006] [0.006] <0.334> <0.330> Age 1.35E 04 1.38E 04 [1.56E 04] [1.57E 04] <0.402> <0.394>

PAGE 53

53 Table 25. Continued. (1) ( 2 ) (3 ) VARIABLES MDS (weight) MDS (weight) MDS (weight) Primary school (highest level) 0.006 0.006 [0.004] [0.004] <0.166> <0.168> Junior secondary school (highest level) 0.002 0.002 [0.005] [0.005] <0.718> <0.696> Secondary school (highest level) 0.017*** 0.017*** [0.006] [0.006] <0.012> <0.012> SSC pass (highest level) 0.012* 0.012* [0.007] [0.007] <0.102> <0.074> Postsecondary education 0.017** 0.017** [0.008] [0.008] <0.052> <0.040> Spouse of household head 0.002 0.002 [0.006] [0.006] <0.732> <0.666> Other relationship to household head 0.001 0.002 [0.006] [0.006] <0.866> <0.830> Hunger at time of meal event 0.003 0.003 [0.002] [0.002] <0.210> <0.214> Baseline nutrition knowledge 4.00E 04 3.97E 04 [3.59E 04] [3.60E 04] <0.268> <0.290> Order effect ( BPP in first meal) 0.004 [0.006] <0.472> Constant 0.764*** 0.658*** 0.658*** [0.008] [0.023] [0.023] Number of observations 2,227 2,137 2,137 R squared 0.010 0.214 0.214 Robust standard errors in brackets; Wild cl uster bootstrap t p value in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 54

54 Table 26. Treatment effects on m eal diversity score by volume (1) ( 2 ) (3 ) VARIABLES MDS (volume) MDS ( volume ) MDS ( volume) Nutrition education 0.012** 0.008 0.008* [0.006] [0.005] [0.005] <0.044> <0.110> <0.090> Nutrition and gender education 0.014** 0.013** 0.014** [0.006] [0.006] [0.006] <0.018> <0.032> <0.042> BPP intervention 3.62E 04 0.006 0.005 [0.010] [0.008] [0.008] <0.954> <0.414> <0.450> BPP x nutrition education 0.023*** 0.016** 0.017** [0.008] [0.008] [0.008] <0.010> <0.062> <0.042> BPP x nutrition and gender education 0.015* 0.010 0.011 [0.008] [0.007] [0.007] <0.082> <0.148> <0.140> Mymensingh district 0.014* 0.014* [0.008] [0.008] <0.082> <0.094> Household monthly income 4.58E 08 3.62E 08 [1.98E 07] [1.94E 07] <0.804> <0.862> Poverty score 1.96E 04 1.97E 04 [1.67E 04] [1.68E 04] <0.260> <0.262> Number of food groups produced 0.001 0.00E+00 [0.001] [0.001] <0.602> <0.640> Household food insecurity access score 3.18E 04 3.02E 04 [0.001] [0.001] <0.562> <0.572> Religion (Muslim / non Muslim) 0.076*** 0.077*** [0.015] [0.015] <0.016> <0.001> Female 0.008 0.008 [0.007] [0.007] <0.246> <0.228> Age 4.35E 05 4.63E 05 [1.64E 04] [1.62E 04] <0.776> <0.790>

PAGE 55

55 Table 26. Continued. (1) ( 2 ) (3 ) VARIABLES MDS (volume) MDS ( volume ) MDS ( volume) Primary school (highest level) 0.005 0.005 [0.005] [0.005] <0.316> <0.300> Junior secondary school (highest level) 0.002 0.002 [0.006] [0.006] <0.700> <0.704> Secondary school (highest level) 0.021*** 0.021*** [0.006] [0.006] <0.040> p<0.001 SSC pass (highest level) 0.008 0.008 [0.007] [0.007] <0.262> <0.280> Postsecondary education 0.013 0.013 [0.009] [0.009] <0.124> <0.146> Spouse of household head 0.002 0.002 [0.005] [0.005] <0.666> <0.750> Other relationship to household head 0.003 0.002 [0.006] [0.006] <0.662> <0.704> Hunger at time of meal event 0.006** 0.006** [0.003] [0.003] <0.068> <0.480> Baseline nutrition knowledge 3.45E 04 3.49E 04 [3.93E 04] [3.91E 04] <0.338> <0.352> Order effect ( BPP in first meal) 0.005 [0.006] <0.464> Constant 0.743*** 0.634*** 0.634*** [0.007] [0.019] [0.019] Number of observations 2,227 2,137 2,137 R squared 0.009 0.113 0.114 Robust standard errors in brackets; Wild cl uster bootstrap t p value in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 56

56 Table 27. Treatment effects on f ood item consumption without covariates (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) VARIABLES Rice Chicken Fish Egg Lentil Salad Mixed veg Leafy veg Fruit Yogurt Nutrition education 0.007 0.007** 0.001 0.003 0.014** 0.004 0.004 0.005* 0.002 0.005 [0.013] [0.003] [0.002] [0.002] [0.006] [0.003] [0.004] [0.003] [0.003] [0.003] <0.608> <0.200> <0.586> <0.326> <0.596> <0.436> <0.416> <0.320> <0.452> <0.328> Nutrition and gender education 0.013 0.003 0.002 0.001 0.016*** 0.002 0.003 0.005** 0.004 0.005 [0.013] [0.002] [0.003] [0.002] [0.005] [0.003] [0.004] [0.003] [0.003] [0.004] <0.386> <0.256> <0.418> <0.452> <0.270> <0.556> <0.442> <0.278> <0.290> <0.466> BPP intervention 0.010 0.001 0.004 0.003 0.011 0.003 0.003 0.002 0.004 0.004 [0.015] [0.005] [0.005] [0.003] [0.008] [0.007] [0.005] [0.003] [0.006] [0.006] <0.544> <0.922> <0.470> <0.370> <0.600> <0.732> <0.616> <0.608> <0.626> <0.600> BPP x nutrition education 0.012 2.05E 04 0.002 0.003 0.029*** 0.004 0.003 0.003 0.007 0.005 [0.018] [0.004] [0.006] [0.003] [0.010] [0.007] [0.006] [0.004] [0.007] [0.007] <0.534> <0.948> <0.708> <0.428> <0.840> <0.718> <0.634> <0.546> <0.670> <0.560> BPP x nutrition and gender education 0.006 0.001 0.003 0.001 0.020** 0.001 0.002 0.005 0.007 0.003 [0.017] [0.004] [0.005] [0.003] [0.010] [0.005] [0.005] [0.005] [0.007] [0.007] <0.722> <0.722> <0.640> <0.800> <0.768> <0.888> <0.680> <0.568> <0.640> <0.688> Constant 0.386*** 0.074*** 0.076*** 0.058*** 0.079*** 0.046*** 0.068*** 0.051*** 0.066*** 0.079*** [0.011] [0.003] [0.003] [0.002] [0.006] [0.004] [0.003] [0.002] [0.005] [0.005] Number of observations 2,227 2,227 2,227 2,227 2,227 2,227 2,227 2,227 2,227 2,227 R squared 0.002 0.004 0.003 0.004 0.010 0.001 0.002 0.002 0.004 0.003 Robust standard errors in brackets; Wild cl uster bootstrap t p value in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 57

57 Table 28. Treatment effects on f ood item consumption with covariates (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) VARIABLES Rice Chicken Fish Egg Lentil Salad Mixed veg Leafy veg Fruit Yogurt Nutrition education 0.009 0.005 0.001 0.001 0.013** 0.004 0.004 0.005** 0.003 0.005 [0.013] [0.003] [0.002] [0.002] [0.005] [0.003] [0.003] [0.002] [0.003] [0.003] <0.530> <0.336> <0.676> <0.606> <0.402> <0.446> <0.428> <0.234> <0.500> <0.268> Nutrition and gender education 0.015 0.002 0.002 2.50E 04 0.017*** 0.002 0.003 0.005* 0.004* 0.004 [0.013] [0.003] [0.003] [0.002] [0.004] [0.003] [0.004] [0.003] [0.002] [0.004] <0.346> <0.452> <0.558> <0.880> <0.114> <0.536> <0.498> <0.290> <0.254> <0.428> BPP intervention 0.002 0.002 0.007 0.001 0.006 0.002 0.004 3.07E 04 0.001 0.004 [0.014] [0.004] [0.004] [0.002] [0.005] [0.006] [0.005] [0.003] [0.006] [0.005] <0.886> <0.572> <0.356> <0.680> <0.406> <0.722> <0.528> <0.936> <0.882> <0.526> BPP x nutrition education 0.022 0.002 0.001 0.001 0.023*** 0.004 0.003 0.002 0.009 0.005 [0.017] [0.004] [0.005] [0.003] [0.007] [0.005] [0.005] [0.003] [0.007] [0.006] <0.332> <0.652> <0.898> <0.778> <0.386> <0.646> <0.634> <0.522> <0.652> <0.532> BPP x nutrition and gender education 0.009 2.10E 04 0.005 0.002 0.017** 1.84E 05 0.001 0.003 0.008 0.002 [0.014] [0.004] [0.004] [0.003] [0.008] [0.004] [0.004] [0.004] [0.007] [0.006] <0.534> <0.976> <0.378> <0.538> <0.472> <0.996> <0.836> <0.476> <0.612> <0.846> Mymensingh district 0.054** 0.009* 0.019*** 0.008*** 0.050*** 0.020*** 0.017*** 0.036*** 0.045*** 0.043*** [0.020] [0.005] [0.003] [0.002] [0.007] [0.007] [0.006] [0.004] [0.007] [0.007] <0.346> <0.460> <0.044> <0.102> <0.648> <0.992> <0.760> <0.532> <0.984> <0.796> Household monthly income 3.27E 07 1.28E 08 4.95E 08 6.15E 08 3.41E 07 1.21E 07 8.86E 08 4.50E 08 5.00E 08 1.74E 08 [3.49E 07] [1.50E 07] [7.34E 08] [5.10E 08] [3.21E 07] [1.25E 07] [1.60E 07] [9.86E 08] [1.67E 07] [1.86E 07] <0.422> <0.926> <0.534> <0.286> <0.586> <0.494> <0.624> <0.668> <0.792> <0.938> Poverty score 0.001* 2.32E 05 4.86E 04 3.96E 05 8.46E 06 3.99E 05 1.86E 04* 3.07E 05 2.01E 04 2.04E 05 [3.98E 04] [8.64E 05] [7.50E 05] [4.46E 05] [1.87E 04] [7.87E 05] [1.11E 04] [8.46E 05] [1.03E 04] [1.23E 04] <0.212> <0.804> <0.338> <0.398> <0.960> <0.664> <0.372> <0.738> <0.830> <0.872> Number of food groups produced 0.004* 1.85E 04 4.32E 04 7.46E 05 2.84E 04 7.42E 05 6.37E 06 2.65E 04 0.001 4.72E 04 [0.002] [0.001] [4.39E 04] [3.00E 03] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] <0.226> <0.780> <0.390> <0.802> <0.772> <0.894> <0.992> <0.664> <0.552> <0.576>

PAGE 58

58 Table 28. Continued. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) VARIABLES Rice Chicken Fish Egg Lentil Salad Mixed veg Leafy veg Fruit Yogurt Household food insecurity access score 0.002 0.001** 1.52E 04 3.39E 05 4.23E 04 2.04E 04 0.001* 1.88E 04 1.91E 04 0.001* [0.002] [4.21E 04] [2.30E 04] [4.46E 05] [3.81E 04] [3.43E 03] [4.56E 04] [2.96E 04] [2.96E 04] [4.77E 04] <0.352> <0.386> <0.556> <0.862> <0.370> <0.656> <0.478> <0.590> <0.598> <0.482> Religion (Muslim / non Muslim) 0.116*** 0.019** 0.012*** 0.021*** 0.004 0.004 0.017* 0.020*** 0.015** 0.021** [0.033] [0.008] [0.004] [0.006] [0.009] [0.016] [0.010] [0.006] [0.006] [0.009] <0.658> <0.824> <0.316> <0.884> <0.744> <0.878> <0.840> <0.760> <0.466> <0.658> Female 0.055*** 0.011*** 0.015*** 0.006*** 0.050*** 0.014*** 0.007 0.003 0.010 0.004 [0.014] [0.004] [0.004] [0.002] [0.011] [0.005] [0.005] [0.003] [0.007] [0.005] <0.040> <0.212> <0.138> <0.680> <0.978> <0.880> <0.520> <0.542> <0.736> <0.572> Age 7.02E 05 1.51E 03* 3.46E 05 9.41E 05** 1.61E 04 3.41E 05 1.76E 04 1.52E 04** 2.38E 04** 1.41E 04 [3.57E 03] [8.77E 05] [9.26E 05] [4.53E 05] [1.93E 04] [9.17E 05] [1.29E 04] [6.82E 04] [1.07E 04] [1.06E 04] <0.848> <0.230> <0.736> <0.100> <0.556> <0.764> <0.460> <0.232> <0.328> <0.354> Primary school (highest level) 1.98E 03 0.001 0.003 4.43E 04 0.005 4.04E 05 0.003 0.001 0.002 0.002 [0.011] [0.002] [0.002] [0.001] [0.004] [0.003] [0.003] [0.002] [0.002] [0.002] <0.990> <0.602> <0.320> <0.648> <0.458> <0.982> <0.484> <0.638> <0.554> <0.404> Junior secondary school (highest level) 0.006 0.005 0.004 0.002 0.011 0.003 0.003 0.002 0.005 0.007 [0.016] [0.004] [0.003] [0.001] [0.010] [0.004] [0.004] [0.003] [0.005] [0.004] <0.704> <0.308> <0.238> <0.202> <0.610> <0.578> <0.512> <0.536> <0.480> <0.292> Secondary school (highest level) 0.015 0.001 0.004 0.001 0.002 1.75E 04 0.014** 0.006 0.003 0.006 [0.021] [0.005] [0.003] [0.002] [0.009] [0.005] [0.006] [0.005] [0.006] [0.006] <0.538> <0.778> <0.244> <0.736> <0.846> <0.968> <0.352> <0.426> <0.646> <0.420> SSC pass (highest level) 0.000 0.004 0.004 0.001 0.000 0.005 0.001 1.30E 03 0.001 0.008 [0.014] [0.004] [0.004] [0.002] [0.009] [0.006] [0.006] [0.004] [0.007] [0.005] <0.970> <0.358> <0.460> <0.706> <0.978> <0.530> <0.894> <0.974> <0.952> <0.200> Postsecondary education (highest level) 0.054** 0.005 0.006 0.002 0.002 0.014** 0.005 0.006 0.002 0.003 [0.024] [0.005] [0.004] [0.002] [0.014] [0.006] [0.006] [0.004] [0.007] [0.007] <0.278> <0.350> <0.222> <0.450> <0.906> <0.470> <0.512> <0.454> <0.724> <0.648>

PAGE 59

59 Table 28. Continued. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) VARIABLES Rice Chicken Fish Egg Lentil Salad Mixed veg Leafy veg Fruit Yogurt Spouse of household head 0.032** 2.64E 04 0.003 1.77E 05 0.003 0.004 0.005* 0.003 0.005 0.004 [0.013] [0.003] [0.002] [0.001] [0.005] [0.004] [0.003] [0.003] [0.003] [0.004] <0.136> <0.934> <0.288> <0.988> <0.620> <0.630> <0.272> <0.462> <0.414> <0.460> Other relationship to household head 0.005 1.01E 05 0.001 0.001 0.010 0.002 0.001 1.68E 05 0.004 0.001 [0.016] [0.004] [0.003] [0.002] [0.007] [0.004] [0.005] [0.003] [0.005] [0.005] <0.746> <0.996> <0.656> <0.622> <0.434> <0.686> <0.902> <0.998> <0.518> <0.908> Hunger at time of meal event 0.001 0.002 0.004*** 2.77E 04 0.001 0.002 3.38E 04 0.003** 0.001 0.004** [0.006] [0.002] [0.002] [0.001] [0.002] [0.002] [0.002] [0.001] [0.002] [0.002] <0.904> <0.510> <0.524> <0.798> <0.650> <0.760> <0.850> <0.438> <0.770> <0.658> Baseline nutrition knowledge 4.01E 04 1.40E 04 1.02E 04 1.36E 05 4.46E 04 2.05E 04 0.001* 9.50E 06 3.42E 04 0.001** [0.001] [1.95E 04] [1.73E 04] [1.07E 04] [3.85E 04] [2.07E 04] [2.72E 04] [1.90E 04] [2.19E 04] [2.96E 04] <0.654> <0.544> <0.580> <0.902> <0.418> <0.516> <0.384> <0.948> <0.296> <0.342> Constant 0.370*** 0.052*** 0.053*** 0.050*** 0.087*** 0.040*** 0.045*** 0.047*** 0.031** 0.042*** [0.039] [0.012] [0.010] [0.008] [0.021] [0.015] [0.017] [0.008] [0.015] [0.013] Number of observations 2,137 2,137 2,137 2,137 2,137 2,137 2,137 2,137 2,137 2,137 R squared 0.097 0.076 0.161 0.072 0.190 0.098 0.029 0.144 0.146 0.177 Robust standard errors in brackets; Wild cl uster bootstrap t p value in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 60

60 CHAPTER 3 REPEATED EXPOSURE TO A BEHAVIORAL NUDGE IN THE HOME AND NUTRITION EDUCATION: LONGTERM IMPACTS ON DIETARY DIVERSITY Motivation M any developing countries have made progress in food security, yet malnutrition persists in terms of micronutrient deficiencies due to low dietary quality (Iannottie et al., 2009 ; World Bank, 2013). This problem is often referred to as hidden hunger (Iannottie et al., 2009). In particular, vitamin A and iron deficiencies plague the adult population in several low income countries (Iannottie et al., 2009). Increasing dietary diversity can improve micronutrient deficiencies since consuming a wider variety of food increases nutrient intake (Iannottie et al., 2009; Hatloy et al., 1998; Torheim et al., 2004). Thus, to combat the hidden hunger problem, many development initiatives promote diet diversification through activities such as homestead gardening, community health campaigns, and cooking demonstrations, designed to help disseminate nutrition guidelines. Sever al countries have adopted sciencebased tools known as food based dietary guidelines (FBDG) to deliver country specific nutrition messages with a focus on locally available food items (FAO, 2016). The USDA MyPlate is the official FBDG diagram used in the U nited States. Studies have shown the USDA MyPlate is an effective nudge that encourages individuals to choose healthier food items (Brown et al., 2014; Miller et al., 2016). A nudge is a subtle cue that encourages an individual to change behavior without restricting his or her choice set (Thaler and Sunstein, 2009). Miller et al. (2016) employ two behavioral economics strategies, pre ordering and MyPlate prompts, to the selection of food items among elementary and middle school students enrolled in the U. S. National School Lunch Program (NSLP). Prompting, or nudging, students with MyPlate messages during the pre ordering process results in larger increases in the selection of fruits, vegetables, and low fat milk of 51.4% 29.7 % and 37.3%

PAGE 61

61 respectively co mpared to pre ordering alone Brown et al. (2014) measures changes in self reported food frequency when college students are exposed to MyPlate prompts via text message. The study finds exposure to MyPlate messages increases the consumption of fruits by 13% and vegetables by 8 % compared to the control group. Leak et al. ( 2015) introduces protocol for parents to use behavioral economics strategies in the home to nudge children toward eating more vegetables with dinner. One strategy instructs parents to serve dinner on a disposable plate outlining proper vegetable portioning according to MyPlate guidelines. Contrary to Brown et al. (2014) and Miller et al. (2016), an evaluation of the se strategies finds weak evidence that a MyPlate nudge increases reported frequency of vegetable intake in the home (Leak, 201 7). In 2013 a plate based diagram similar to the MyPlate was developed by the SHIKHA1 project in Bangladesh. The diagram illustrates dietary recommendations and proper portioning using pictures of locally so urced food items. As a tool for disseminating nutrition messages under the SHIKHA project, the diagram and key written messages2 were printed on a melamine plate, which we will refer to as the Bengali Portion Plate (BPP). The Bangladesh Ministry of Health and Family Welfare adopted the BPP for use as a counseling tool to improve maternal and child nutrition ( FHI360/USAID 2016). Although the BPP was designed to target pregnant and lactating women, the guidelines portrayed by the plate fit the dietary recomm endations for the general Bangladeshi population, male or female. This essay investigates whether the use of the BPP in the home environment effectively nudges nutrition decisions among rural households. 1 The SHIKHA project is a USAID initiative implemented by FHI360 and BRAC to promote dietary diversity and improved nutrition among pregnant and lactating women in Bangladesh 2 The food plate follows the national dietary guidelines for Bangladesh and also promotes messages for child and maternal nutrition including: Eating a variety of food in appropriate amounts keeps mothers and children healthy and Eat a little more food during pregnancy.

PAGE 62

62 Behavioral economics is increasingly applied to research in development economics ; however its application to nutrition in low income countries is relatively new. The World Bank recently implemented a study on nudging mothers in Madagascar to purchase healthier food ( World Bank 2016a ). At this time, the e xperiment is ongoing. Thus, our research is one of the first contributions to the literature on nutrition nudging in development. To date, most nutrition and health interventions in developing countries have focused on tools for behavior change communicati on (BCC). Behavioral economics (BE) differs from BCC in that BE discreetly encourages individuals to make decisions without influencing their beliefs or values, whereas BCC uses various messaging sources to specifically promote change in the knowledge, beliefs, and values of an individual or a community A common example of BCC is the promotion of health and nutrition by community health workers. In a randomized controlled trial in Malawi, Fitzsimons et al. (2016) finds that visits from c ounselors on child nutrition increase per capita monthly food consumption of protein rich foods, fruits, and vegetables in the household. Similarly, a community growth promotion program evaluation in Uganda finds higher consumption of legumes, milk, fruits and vegetables among children in households visited by community health workers ( Alderma n, 2007). The interventions in the aforementioned studies promote best practices for child nutrition. Our research departs from the nutrition development literature in that it focuses on nutrition education and nudging to improve individual dietary diversity among adults. Encouraging adults consumption of fruits and vegetables is particularly important for socioeconomically disadvantaged households (Ansem et al., 201 4; Ball et al. 2006; Campbell et al., 2013). Educating adults can lead to better feeding practices for children since parents make the rules

PAGE 63

63 and decisions about food consumed in the household ( Wong et al., 2014; van Ansem et al., 2014). In a study of schoo lchildren in China, Wong et al. (2014) find child hemoglobin levels improve when parents engage in participatory training on nutrition and anemia. In a Dutch study, van Ansem et al. (2014) finds the home availability of fruits and vegetables increases when adults are encouraged to consume healthy food. The presence of fruits and vegetables improves the home food environment such that children have more access to healthy foods ( van Ansem et al., 2014). Acknowledging adult behavior as an agent for change in h ousehold nutrition, the intervention in this study combines behavioral economic strategies with behavior change communication to promote healthy food choices among adults in rural Bangladesh. To our knowledge, there have been no studies on the use of platebased food diagrams to nudge the consumption of healthy foods and dietary diversity in developing countries. Thus, we make a novel contribution to research on nutrition nudging in development by applying behavioral economics methods to dietary diversity i n Bangladesh. Even within the U.S. literature on behavioral economics in nutrition, few studies evaluate the use of such strategies in the home environment, which is the focus of this analysis. Our experiment uniquely combines behavioral economics with BCC in the random assignment of both exposure to the BPP, a BE tool, and nutrition education through participatory training, a method of BCC. Because of its unique design and the focus on the home environment, the experimental protocol in this research is imp ortant to inform future studies on nutrition nudges in both the developed and developing world. By r andomly assigning exposure to the BPP, this study investigates the effectiveness of the BPP at nudging both men and women in rural Bangladesh toward more d iversified diets. Treated participants were first exposed to the BPP during a lunch buffet where food choices were

PAGE 64

64 discreetly observed. After the meal BPPs were distributed to treated participants for use at home. Each participant received one BPP for each member of his or her family. In this essay we measure whether exposure to the BPP in the home environment impacts individual dietary diversity. Our experiment also includes a second level of treatment, nutrition education through a participatory workshop. This research aims to measure the impact of repeated exposure to nutrition guidelines via a nutrition education and a plate printed with a FBDG icon Thus, we measure the impacts of the BPP alone and when combined with nutrition education. We also measure the individual impacts of nutrition education. In a randomized controlled trial we evaluate whether exposure to the BPP will nudge individuals in rural Bangladesh toward more diverse diets at home. The results from a difference in differen ce analys is on the 24 hour individual dietary diversity score fail to provide evidence that the BPP nudges study participants toward dietary diversity. However, over a longer reference period, we see a positive impact of the BPP nudge on dietary diversity measured by the food consumption score. T he study advocates a need for future research on nutrition nudges in developing countries and develops protocol for conducting such research. The next section of this essay describes the experimental design and trea tment assignment. The empirical methodology is then defined, followed by results and possible explanations for our findings. The essay concludes with a discussion on the relevance of our study and suggestions for future research. Experimental Design A randomized experiment to evaluate the impacts of two nutrition interventions, satisfying both BE and BCC properties, was conducted from August 2016 to January 2017 in two districts of Bangladesh. Data from the experiment include a baseline survey, two meal observations, and an endline survey for 1,099 individuals from 53 villages across two districts in

PAGE 65

65 Bangladesh. In the baseline survey, the participant provided information on demographics in a household roster, individual and household 24hour food group consumption, 7 day food group consumption, nutrition knowledge, poverty, food insecurity, and agricultural production. The endline survey mirrored the baseline questionnaire, but also included questions about the nutrition workshop (i.e. the field experime nt described below) and the receipt and use of the BPP at home. In addition to the surveys, each individual participated in a field experiment, which involved attending two meal events, approximately one month apart. The baseline survey was conducted as a faceto face interview at the participants household at least one day prior to the first meal event. The endline survey, also a faceto face interview, was conducted at the participants household at least one month after the second meal event. Participan ts for the study were randomly selected from the membership rosters of the Bangladesh Agricultural University Extension Center (BAUEC) in the Mymensingh district and Shushilan a NGO in the Borguna district. The research was funded by the USAID initiative Integrating Gender and Nutrition within Agricultural Extension Services (INGENAES), thus one objective was to coordinate research alongside local extension providers. The specific partner agencies were selected on the criteria that they provide technical agricultural assistance to their respective beneficiaries, but had not established protocol for disseminating nutrition guidelines to rural households at the time of the study. BAUEC serves male and female farmers, 55% and 45% respectively, in the Mymens ingh district. Shush ilan is a national organization; however we partnered specifically with the Managing Natural Resources by the Coastal Community (MaNaR) project in the Borguna district. The MaNaR project is targeted specifically toward rural women, t hus its beneficiary roster is 94 % female.

PAGE 66

66 The primary objective of this experiment was to investigate the impact of repeated exposure to nutrition guidelines using different mediums. As such, participants were randomly assigned to two nutrition interventi ons : 1) exposure to the BPP and 2) nutrition education in a participatory workshop. Participants who were assigned to the BPP treatment were first exposed to the tool during a lunch buffet where they were also given the BPP to use at home. Thus, the BPP in tervention was cluster randomized at the village level to reduce potential spillover effects during at home use that would contaminate the control group. Nutrition education was randomly assigned at the individual participant level. Each participant was invited to his or her respective field office, BAEUC or S h ushilan, for a lunch buffet on two occasions, one month apart. All attendees were from the same village on any given day. Enumerators conducted a baseline survey at the participants household at le ast one day prior to the first meal. The BPP intervention was implemented in terms of the plate that participants used during the lunch buffet. The control group used a standard plate during both buffet meals. Individuals from treated villages served thems elves using the BPP during one of the two events and a standard plate durin g the other event ( F igure 21). Some villages used the BPP during the first meal while others used the BPP in the second meal. During the meal, data collectors discreetly observed p articipants food choices. The meal observation data collection protocol and analysis are described in Chapter 2 If the participant used a BPP during the meal, he or she was given one BPP for each member of his or her household to use at home after the me al The nutrition education intervention was implemented in the morning, prior to each lunch buffet. Individuals who were assigned to the nutrition education treatment arrived early to participate in the workshop(s) described below. The nutrition education assignment remained the

PAGE 67

67 same during both of the two meals. Participants who were assigned to the control group for nutrition education simply attended the lunch buffet and did not attend a ny nutrition workshop s The nutrition education intervention encom passed two levels of treatment : 1) nutrition education only and 2) nutrition education with a gender component Individuals were randomly assigned to one of the two treatments or the control group of no nutrition education The trainings followed the small group participatory methods of two activities described in Henderson (2016). Specifically, the nutrition education intervention followed the What Goes on the Plate activity in which participants were asked to draw a circle representing a plate and illus trate food items that characterize a balanced diet within that plate. A group representative then shared and described the illustration to open a discussion among the participants. In a complementary activity, groups were given a budget and asked to prepar e a shopping list for a balanced meal within that budget. To conclude the session the facilitator summarized the national dietary guidelines, presented examples of healthy food items, and encouraged the consumption of healthy foods such as fruits and veget ables. The national dietary guidelines coincide with the information on the BPP, however the facilitator did not explicitly show the BPP during the workshop. Individuals who were also assigned to the gender component of the BCC intervention participated in the Who Gets What to Eat activity in addition to the nutrition education In this activity each participant was assigned to role play as a different member of the household (wife, husband, father in law, daughter, etc.). The person assigned to the role of wife distributed food items to each member, mimicking traditional household gender roles for food allocation. A discussion then ensued about the nutritional needs of men and women as well as the importance

PAGE 68

68 of including protein, fruits and vegetables in womens diets, particularly for adolescent, pregnant, and lactating women. Empirical Analysis The baseline and endline survey data are evaluated using differencein difference analysis to measure the impacts of the BPP and nutrition education interventions on dietary diversity at home. The differencein difference model estimates changes in the individual dietary diversity score IDDS, as a function of the interventions, controlling for changes in dietary diversity over time in the control grou p. Following FAO guidelines for dietary diversity, IDDS is a count of the number of food groups consumed by the individual in the last 24 hours, [0,15] (Kennedy, 2010) In addition to the IDDS, we estimate the change in the household food consumption score (FCS), a 7 day measure of dietary diversity and frequency of consumption. By design, the FCS weights different food groups by nutrient value in addition to recording frequency of consumption (WFP, 2008). Following the calculation steps in WFP (2008), we m ultiply the frequency of consumption of nine different food groups by the defined weight of that food group. The FCS is the sum of the weighted consumption frequencies. In this analysis, FCS remains a continuous variable ranging from 0 to 112 Evaluat ing both IDDS and FCS provides a more comprehensiv e picture of potential outcomes since the two measures vary by reference period and the FCS accounts for nutrient content (FAO and WFP, 2012). The following equation presents the general form of the model for i ndividual i in randomized cluster j at time t where D is the respective dietary diversity score, IDDS or FCS

PAGE 69

69 (3 1) In E quation (31) post is a dummy variable equal to 0 for pre intervention observations collected in the baseline survey and 1 for post intervention observations collected in the endline survey. The treatments are indicated in variables BPP, nutritionED and genderED Each indicator takes the value of 1 when true and 0 otherwise. BPP is a tr eatment variable equal to 1 for a participant from village j randomly assigned to the BPP treatment. NutritionED is a dummy variable equal to 1 if the participant was assigned to the nutrition education treatment. The genderED variable equal to 1 indicates the individual was assigned to the nutrition education with a gender component treatment. The interaction terms between BPP and the education treatments are dummy variables indicating assignment to both interventions. The n utritionED x BPP variable is equal to 1 when an individual was exposed to the BPP and also received nutrition education. Similarly, genderED x BPP is equal to 1 for an individual who was assigned nutrition education with the gender component and was also assi gned to the BPP treatment. The es timated coefficients 1, 2, 3, 4, a nd 5, represent the treatment effects of the BPP intervention, the nutrition education intervention, nutrition and gender education, the BPP combined with nutrition education, and the BPP combined with nutrition and gender education, respectively. The model includes time invariant individual and household level covariates, contained in Xij. To control for geographical differences in agricultural productivity and availability of food, district is a dummy variable equal to 1 if the participant lives in Mymensingh and 0 if the participant lives in Borguna. Demographic indices such as gender, a ge, and education affect food

PAGE 70

70 preferences as well as the intrahousehold allocation of food in Bangladesh. Thus, we include gender of the individual as a dummy variable equal to 1 if the participant is female, and 0 otherwise. Age of the individual is a co ntinuous variable. Education is a categorical variable equal to 1 for the highest level of education c ompleted by the individual (no education primary, junior secondary, secondary, SSC pass, or postsecondary education) where no education is the omitted va riable. We hypothesize that a nutrition intervention will have a greater impact if the person receiving the information has control over his or her food choices within the household. Thus, we include food preparer as a dummy variable equal to 1 if the part icipant is the primary person responsible for preparing meals for the household. Total cultivable land is a continuous variable measured in decimals. Land ownership serves as a proxy for the household poverty level, which impacts the individuals food availability and access. Studies have shown farm diversity is positively correlated with household dietary diversity ( Jones et al., 2014; Kumar et al., 2015; Rawlins et al., 2014; Sibhatu et al., 2015). Thus, we include farm diversity measured as the count of food groups produced by the household. Equati on (3 1) is estimated using Poisson regression on IDDS and ordinary least squares regression on FCS. S tandard errors are clustered at the villagelevel for all analyses To test for robustness we estimate both model s with and without covariates. Since IDDS is an aggregate measure of all food groups consumed, the outcome does not necessarily capture the treatment effects if a participant simultaneously increases and decreases his or her consumption of different food items. Thus, we also measure the impact of the interventions on the consumption of each food group in a series of logistic regressions. For each of the 15 food groups, the dependent variable is a binary variable indicating whether or not the individua l consumed that particular food group in the last 24 hours.

PAGE 71

71 The general form of the model for individual i in randomized cluster j at time t is: (3 2) where yfijt is vector of binary variables equal to 1 if the individual consumed food group f in the last 24 hours and 0 otherwise. The treatment varia bles and covariates in E quation (32) follow the definitions in E quation (31). Equation (32) is estimated using logi stic regression with standard errors clustered at the village level. Results Table 3 1 presents the pre and post intervention means for 24hour consumption of each food group mean IDDS, and mean FCS by treatment and control. Two sample test s of proportio ns ( T able 3 2) reveal the increase in mean consumption is statistically significant for several food groups including vegetables, fish, eggs, and fats/oils. There is a statistically significant decrease in meat consumption between the two time periods for two of our treatment groups. Table 3 2 also reports t wo sample t test s of the difference s in mean IDDS and mean FCS between pre and post interve ntion periods by treatment Mean IDDS and FCS are statistically significantly higher post intervention for parti cipants in the BPP only and the combined treatment groups. In several cases, statistically significant differences in consumption and IDDS occur within the control group as well as treated participants. The change in dietary diversity in the control group indicates seasonality in food consumption and validates the use of differencein difference analysis to control for natural trends in consumption pat terns. The endline survey include s a number of questions about BPP use at home. Prior to analyzing the tre atment effects, we first investigate the level of compliance with the BPP

PAGE 72

72 treatment (i.e. whether participants used the BPP at home). Table 33 reveals that 11 % of all individuals in villages assigned to the BPP treatment never use the BPP at home. Participants across treatment groups use the BPP at similar frequencies and for similar purposes ( T able 33) Approximately 47% of participants us e the BPP to eat meals sometimes (3 4 times in the last 4 weeks) or often (more than 10 times in the last 4 we eks) Similarly, 46 % and 48% of participants sa y they refer to the BPP to make decisions on the type and amount of food to prepare and consume, respectively. Table 34 tabulates noncomplying individuals with the reasons why they never use the BPP at home. The most commonly cited reason for not using the BPP at home is a preference to use other plates to eat meals. Some individuals d o not use the BPP due to a lack of access or affordability to food items promoted by the pictorial diagram. However, the numbe r of respondents who sa y that the BPP is more suitable for use when guests arrive i s higher than those who report income or access constraints to BPP use. Perhaps recipients of the BPP view the plate as special dinnerware, which may also explain why the ma jority of people say they prefer to use other plates to eat their meals. The results from the differencein difference regressions on IDDS are presented in T able 35. Model (1) excludes covariates, whereas model (2) includes individual and household characteristics to test for robustness. The results show no evidence of treatment effects for the BPP or nutrition education interventions. None of the treatment variables ( treatment x post) are statistically significant. The variable post captures the time eff ect; its statistical significan ce in model (2 ) reflects a seasonal trend in individual dietary diversity H owever, in the 24 hour consumption data there is no causal evidence to suggest the BPP or nutrition education improve s

PAGE 73

73 24hour individual dietary diversity T he results are fairly robust since only small changes in coefficient magnitude arise with the addition of covariates. It is not surprising to see a lack of treatment effect on IDDS since the aggregate measure of dietary diversity may not reflect simultaneous changes in different food groups If a participant decreases the consumption of one food group and increases consumption of another food group IDDS remains unchanged. Thus, we also i nvestigate treatment effects on the 24hour consumption of each individual food group. The r esults from the logistic regressions on food group consumption suggest treatment effects exist but the impact var ies by food group and treatment. Our sample contai ns little to no variation in cereals most individuals consume rice daily thus we do not include a regression on cereals. The marginal effects in T ables 3 6 and 3 7 indicate that the nutrition education treatment increases the likelihood of indi viduals cons uming tubers and roots while the combined treatment increases the likelihood of leafy green vegetable consumption. Respondents assigned to nutrition education with a gender component are 11% more likely to consume tubers and roots compared to individuals who did not participate in trainings or receive the BPP. The combined intervention, exposure to the BPP and nutrition education increase s the likelihood respondents consume leafy green vegetables by 16% compared t o the control. Contrary to our hypothesis, nutrition education decreases the likelihood of individuals consuming vitamin A rich vegetables by 12% compared to the control Combining nutrition and gender education with the BPP decreases the likelihood of in dividuals consuming vitamin A rich fruit and fish by 11% and 14% respectively, compared to the control As a robustness check, we also run logit regressions on each food group with individual and household level covariates specified in the previous sectio n The results in T ables 3 8 and 3 9 show the marginal effects of treatment are consist when covariates are included.

PAGE 74

74 In addition to the 24 hour IDDS model, we measure the treatment effects on the 7 day food consumption score to capture fluctuations in con sumption patterns and food availability throughout the week. As expected, we find different results in the 7 day FCS analysis. Taking into account frequency of consumption over a longer period of time, the re is some evidence that exposure to the BPP at hom e improves dietary diversity ( T able 310). T he results show a 3.5 point increase in mean FCS ( a 5.2% increase on the mean) among households that received the BPP only compared to households that were not assigned to either treatment. The combined treatmentBPP with nutrition education increased the FCS by 3.7 points ( a 5 .4 % increase on the mean ) compared to the control. The results also show a 3.4 point increase in mean FCS among households that received nutrition and gender education but were not assigned to the BPP treatment, compared to the control. Model (2) in T able 310 presents the estimated impact on FCS controlling for individual and householdlevel covariates. We find a slight ly lower magnitude of treatment effects when covariates are included Discussion The results from this study provide mixed evidence on the longterm effectiveness of the BPP and nutrition education interventions We find no statistically significant treatment effects on individual dietary diver sity. On the other hand, the analysis on the 7 day food consumption score presents some evidence that the BPP nudges households to consume a greater variety of food more frequently. The results show the BPP nudge increases the FCS, and the combined interve ntion of the BPP nudge with nutrition education also generate s a positive impact on the frequency of diverse food items consumed in the household. Furthermore, our findings suggest nutrition education with the gender component improves FCS, but nutrition e ducation alone is not effective. This discrepancy in education impact may be due to the reiteration of information. Some of the key nutrition messages were repeated during the gender training. Thus participants

PAGE 75

75 in the dual education treatment may have retained more information, and hence the long term effects were sustainable. Further research should measure retention rates of information disseminated via participatory workshops. I nvestigating the change in 24hour consumption for each food group reveals f ew significant treatment effects and no clear pattern The proportion of individuals consuming leafy green vegetables and white tubers is higher among some treated participants. Contrarily the proportion of individuals consuming fish and vitamin A rich fr uit is lower among some treated participants. There is not enough evidence from the analysis on 24 hour consumption to suggest the individuals respond to either the BPP or nutrition education interventions in a consistent manner. However, the results show that the combined intervention of the BPP with nutrition education promotes leafy green vegetable consumption. The effect may be explained by the prominence of leafy green vegetables on the BPP as well as the food groups affordability and availability as a source of nutrients in Bangladesh. A variety of products in the leafy green vegetable food group (i.e. red amaranth, spinach, jute leaves, pumpkin leaves) are available in the study areas. Thus, participants may have responded to the nutrition messages by increasing their consumption of nutrients in the most affordable and accessible way. The simultaneous decrease in the consumption of other food groups among the treated suggests the participants may be substituting one food group for another. Traditional constraints to accessing diverse food items, such as availability or affordability of healthy food, may explain the different outcomes by food group and frequency of consumption. Further research should explore whether the effectiveness of the BPP nudge v aries based on the number of times the treated household member goes to market over the 7 day time period.

PAGE 76

76 The rate of usage of the BPP at home may also contribute to the difference in impact on IDDS versus FCS. Eleven percent of our treated sample never uses the BPP at home. In other words, there is some lack of compliance with the randomly assigned BPP treatment. Participants who do not use the BPP despite receiving it claim to prefer using other plates during meals at home. Some non compliers repor t sav ing the BPP for use when guests arrive. Thus, rather than the BPP acting as a nudge to increase dietary diversity, households store the BPP as a treasured item. The frequency of BPP use may also explain the difference in treatment effects between IDDS and FCS. Twenty percent of individuals who received the BPP use it at least biweekly (often) to make decisions about food consumption, whereas 25 % or more report only referring to the BPP sometimes. Further analysis investigat ing the heterogeneity of treat ment effects by frequency of BPP use is needed Similarly, f uture research should investigate the characteristics of the individuals in the household who are using the BPP and explore the heterogeneity of treatment effects by user. In addition to BPP use, we plan to investigate the extent to which participants shared the information from the BPP and nutrition education with their family and community members Specifically, we will us e network analysis techniques to investigate the sp read of information since the endline survey asked participants to identify the individuals with whom they shared information about the interventions. A potential limitation of this study is the sampling strategy. Since the participants were randomly selec ted from the membership rosters of our partner agencies, the results are not representative of the national population. Participants membership in our partner organization might affect our results since extension activities vary by agency In particular, the Shushilan MaNaR project emphasizes climate change resilient food production methods such as floating gardens. The different nature of the partner agencies technical services influences the

PAGE 77

77 accessibility and availability of nutritious food items in the home. The pre existing relationship between the partner organization and the participant also potentially biases the responsiveness of participants to the information treatments. Furthermore, participants were compensated for their time in this study. Thi s compensation may have affected the receptiveness of participants to the information. The methods in this study can be applied to a nationally representative sample to measure the impacts of scaling up behavioral economics and behavior change com munication initiatives. Future studies should investigate the use of nutrition nudges using FBDG icons in the home environment in the context of other countries. Further investigation is also warranted to examine whether nutrition initiatives result in foo d group substitution by individuals in low income countries. The research should be extended to understand whether smallholder farmers production decisions change as a result of nutrition initiatives and the extent to which production decisions reflect po tential consumption substitution patterns

PAGE 78

78 Table 31. Pre and post intervention mean 24 hour consumption of food groups by treatment and control+ Control BPP only Nutrition education Nutrition & gender education BPP x nutrition BPP x nutrition & gender Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Cereals 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.99 1.00 Vitamin A rich vegetables 0.19 0.38 0.14 0.32 0.21 0.26 0.19 0.31 0.20 0.30 0.19 0.38 Tubers and roots 0.76 0.66 0.78 0.68 0.82 0.78 0.65 0.70 0.77 0.70 0.77 0.72 Leafy green vegetables 0.58 0.51 0.51 0.54 0.54 0.55 0.63 0.53 0.45 0.54 0.56 0.60 Other vegetables 0.50 0.68 0.49 0.72 0.58 0.69 0.52 0.77 0.53 0.65 0.55 0.68 Vitamin A rich fruit 0.04 0.17 0.04 0.11 0.04 0.15 0.07 0.09 0.04 0.10 0.11 0.13 Other fruit 0.26 0.19 0.29 0.16 0.20 0.18 0.20 0.18 0.25 0.19 0.29 0.21 Meat 0.40 0.29 0.29 0.26 0.39 0.31 0.37 0.23 0.30 0.21 0.33 0.27 Eggs 0.19 0.27 0.21 0.34 0.22 0.29 0.23 0.33 0.24 0.33 0.23 0.34 Fish 0.77 0.88 0.72 0.76 0.64 0.79 0.78 0.81 0.73 0.79 0.78 0.76 Pulses 0.53 0.55 0.52 0.48 0.54 0.50 0.45 0.54 0.57 0.50 0.50 0.58 Dairy 0.17 0.23 0.19 0.24 0.12 0.19 0.18 0.23 0.22 0.22 0.22 0.26 Fats and oils 0.93 0.97 0.92 0.98 0.89 0.98 0.98 1.00 0.89 0.96 0.89 0.97 Sugar 0.17 0.30 0.23 0.33 0.19 0.25 0.19 0.23 0.25 0.32 0.25 0.30 Spices 0.75 0.82 0.78 0.89 0.77 0.75 0.80 0.83 0.75 0.84 0.74 0.83 Individual dietary diversity score (IDDS) 7.25 7.92 7.11 7.82 7.15 7.68 7.25 7.78 7.19 7.66 7.39 8.00 (2.19) (3.39) (2.40) (2.90) (2.10) (3.09) (2.05) (2.87) (2.44) (2.95) (2.81) (2.91) Food consumption score (FCS) 67.92 67.89 67.56 71.28 66.06 69.68 66.28 69.19 68.44 72.01 69.15 72.62 (13.71) (17.73) (18.86) (17.84) (17.08) (18.63) (15.23) (16.67) (18.40) (18.80) (18.29) (17.29) Note: standard deviation of IDDS and FCS in ( ). +In this study, the BPP is a behavioral economics intervention and nutrition education is a behavior change communication strategy.

PAGE 79

79 Table 32. Test of differences in preintervention and post intervention mean 24 hour consumption of food groups by treatment+ Food group (Post Pre) Control BPP only Nutrition education Nutrition & g ender education BPP x nutrition BPP x nutrition & gender Vitami n A rich vegetables 3.09 ** 4.60 ** 0.90 1.92 2.69 ** 4.91 ** <0.002> p < 0.001 <0.368> <0.055> <0.007> p < 0.001 Tubers and roots 1.62 2.06 * 0.94 0.35 1.56 0.97 <0.106> <0.039> <0.349> <0.728> <0.118> <0.334> Leafy green vegetables 0.93 0.81 0.25 2.19 ** 2.29 ** 1.34 <0.350> <0.420> <0.799> <0.029> <0.022> <0.180> Other vegetables 2.71 ** 4.82 ** 1.97 ** 3.80 ** 2.72 ** 3.10 ** <0.007> p < 0.001 <0.049> p < 0.001 <0.007> <0.002> Vitamin A rich fruit 3.02 ** 2.80 ** 2.68 ** 0.73 2.79 ** 0.49 <0.003> <0.005> <0.007> <0.464> <0.005> <0.626> Other fruit 1.11 2.90 ** 0.80 0.91 1.55 1.60 <0.265> <0.004> <0.425> <0.361> <0.121> <0.111> Meat 1.68 1.03 1.33 2.14 ** 2.37 ** 1.59 <0.093> <0.304> <0.182> <0.032> <0.018> <0.112> Eggs 1.41 3.30 ** 1.15 1.43 2.40 ** 2.98 ** <0.158> <0.001> <0.250> <0.154> <0.016> <0.003> Fish 2.09 ** 1.27 2.80 ** 0.12 1.41 0.10 <0.036> <0.202> <0.005> <0.903> <0.158> <0.920> Pulses 0.27 0.71 1.02 1.30 1.32 1.78 <0.790> <0.478> <0.309> <0.192> <0.187> <0.075> Dairy 1.17 1.15 1.25 0.97 0.36 0.96 <0.244> <0.249> <0.212> <0.330> <0.716> <0.338> Fats and oils 1.55 2.77 ** 3.10 ** 2.25 ** 3.01 ** 3.02 ** <0.122> <0.006> <0.002> <0.024> <0.003> <0.003> Sugar 2.35 ** 2.41 ** 1.10 0.80 1.90 1.35 <0.019> <0.016> <0.271> <0.423> <0.058> <0.178> Spices 1.30 3.14 ** 0.15 1.08 2.25 ** 1.81 <0.193> <0.002> <0.880> <0.280> <0.024> <0.070>

PAGE 80

80 Table 3 2. Continued. Food group (Post Pre) Control BPP only Nutrition education Nutrition & g ender education BPP x nutrition BPP x nutrition & gender IDDS 1.773** 2.953** 1.490 1.365 2.123** 2.606** <0.078> <0.003> <0.138> <0.174> <0.034> <0.009> FCS 0.027 3.728** 3.617 2.912 3.566** 3.468** <0.990> <0.032> <0.112> <0.143> <0.030> <0.035> Note: Z -scores and t -score of (Post pre) reported; p -values are in <>. ** indicates significance at = 0.05 +In this study, the BPP is a behavioral economics intervention and nutrition education is a behavior change communication strategy.

PAGE 81

81 Table 33. At home use of BPP by participants All treated BPP only BPP x nutrition education BPP x nutrition and gender education Frequency of BPP use for participants who received the BPP N Percent N Percent N Percent N Percent Use BPP ever 642 89% 205 92% 226 88% 211 89% Never use BPP 77 11% 19 8% 32 12% 26 11% How often do you use the BPP to eat your meals? Never 104 10% 29 13% 38 15% 35 15% Rarely 97 9% 26 11% 29 11% 39 16% Sometimes 212 20% 66 29% 78 30% 64 27% Often 288 27% 95 41% 101 39% 90 38% Did not receive 356 34% 1 0.4% How often do you use the BPP to make decisions on the type of food to prepare? Never 113 10% 38 16% 41 15% 31 13% Rarely 101 9% 29 12% 32 12% 39 16% Sometimes 291 26% 89 38% 108 40% 92 37% Often 223 20% 68 29% 77 29% 74 30% Did not receive 358 33% 1 0.4% How often do you use the BPP to make decisions on the type of food to consume? Never 109 10% 37 16% 39 15% 30 12% Rarely 99 9% 27 11% 33 12% 38 15% Sometimes 292 27% 89 38% 104 39% 97 39% Often 226 21% 70 30% 81 30% 71 29% Did not receive 358 33% 1 0.4% 1 0.4% How often do you use the BPP to make decisions on the type of food to serve others? Never 113 10% 37 16% 41 15% 32 13% Rarely 94 9% 27 11% 29 11% 37 15% Sometimes 275 25% 83 35% 102 38% 88 36% Often 245 22% 77 32% 85 32% 79 32% Did not receive 358 33% 1 0.4% How often do you use the BPP to make your own decisions about the amount of food to consume? Never 106 10% 35 15% 39 15% 29 12% Rarely 96 9% 28 12% 29 11% 38 15% Sometimes 280 26% 84 35% 100 37% 94 38% Often 245 22% 76 32% 90 34% 75 30% Did not receive 358 33% 1 0.4% 1 0.4% Note: Never = not once in the last 4 weeks; rarely = 1 2 times in the last 4 weeks; sometimes = 3 4 times in the last 4 weeks; often = more than 10 times in the last 4 weeks

PAGE 82

82 Table 34. Reason for noncompliance among participants who never use the BPP All treated BPP only BPP x nutrition education BPP x nutrition and gender education N Percent N Percent N Percent N Percent I do not have access to the items on the BPP 3 4% 1 5% 2 6% --I cannot afford the food items on the BPP 3 4% 1 5% 2 6% --I prefer to use other plates 46 60% 13 68% 21 66% 12 46% I do not understand the BPP 1 1% ----1 4% I only use the BPP for guests 5 6% 1 5% 2 6% 2 8% Note: participants could select more than one reason

PAGE 83

83 Table 35. Estimated marginal effects from Poisson regression of treatments on 24 hour individual dietary diversity score (IDDS)+ (1) (2) VARIABLES IDDS IDDS BPP only 0.110 0.043 [0.549] [0.241] Nutrition education 0.133 0.029 [0.331] [0.295] Nutrition education with gender 0.066 0.221 [0.225] [0.193] BPP x nutrition education 0.040 0.012 [0.524] [0.233] BPP x nutrition education with gender 0.192 0.148 [0.576] [0.263] Post 0.669 0.707* [0.424] [0.406] BPP x post 0.083 0.048 [0.448] [0.439] Nutrition education x post 0.165 0.216 [0.486] [0.485] Nutrition education with gender x post 0.241 0.276 [0.462] [0.451] BPP x nutrition education x post 0.154 0.181 [0.462] [0.456] BPP x nutrition education with gender x post 0.004 0.047 [0.471] [0.462] Mymensingh district 3.070*** [0.138] Gender (female = 1) 0.136 [0.193] Age 0.005 [0.005] Primary school (highest level) 0.242** [0.121] Junior secondary school (highest level) 0.418** [0.197] Secondary school (highest level) 0.623*** [0.207] SSC pass (highest level) 0.509** [0.236] Postsecondary education 0.498* [0.257] Responsible for food preparation 0.280* [0.148]

PAGE 84

84 Table 35. Continued. (1) (2) VARIABLES IDDS IDDS Total cultivable land (decimals) 0.001*** [1.90E 04] Farm diversity (count of food groups produced) 0.128*** [0.038] Number of observations 2,164 2,164 Log Likelihood 5153 4704 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 85

85 Table 36. Logistic regression marginal effects results of treatment on 24 hour consumption of fruit and vegetable groups without covariates Treatment Vitamin A rich vegetables Tubers and roots Leafy green vegetables Other vegetables Vitamin A rich fruit Other fruit BPP only 0.069* 0.017 0.075 3.63E 05 0.001 0.025 <0.041> <0.061> <0.061> <0.101> <0.051> <0.061> Nutrition education 0.017 0.062 0.051 0.056 0.008 0.037 <0.041> <0.051> <0.061> <0.051> <0.051> <0.051> Nutrition education with gender 0.003 0.107** 0.065 0.003 0.028 0.039 <0.041> <0.051> <0.081> <0.051> <0.051> <0.041> BPP x nutrition education 0.008 0.011 0.124** 0.032 0.011 0.008 <0.041> <0.061> <0.061> <0.091> <0.041> <0.061> BPP x nutrition education with gender 0.011 0.017 0.022 0.049 0.086 0.023 <0.051> <0.061> <0.061> <0.091> <0.051> <0.061> Post 0.172*** 0.092* 0.062 0.171*** 0.119** 0.061 <0.051> <0.051> <0.081> <0.061> <0.051> <0.071> BPP x post 0.026 0.007 0.099 0.050 0.035 0.053 <0.061> <0.061> <0.091> <0.081> <0.061> <0.071> Nutrition education x post 0.122* 0.036 0.077 0.051 0.012 0.017 <0.071> <0.071> <0.081> <0.081> <0.061> <0.071> Nutrition education with gender x post 0.068 0.109* 0.075 0.058 0.091 0.012 <0.071> <0.061> <0.111> <0.071> <0.061> <0.091> BPP x nutrition education x post 0.070 0.031 0.161* 0.058 0.036 0.004 <0.061> <0.061> <0.091> <0.071> <0.061> <0.081> BPP x nutrition education with gender x post 0.019 0.052 0.123 0.034 0.108** 0.003 <0.071> <0.061> <0.091> <0.071> <0.061> <0.081> Covariates No No No No No No Number of observations 2,164 2,164 2,161 2,164 2,164 2,164 Log Likelihood 1197 1237 1442 1411 634.8 1139 Note: Marginal effects are reported; Cluster robust standard errors are in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 86

86 Table 37. Logistic regression marginal effects results of treatment on 24 hour consumption of other food groups without covariates Treatment Meat Eggs Fish Pulses Dairy Fats and oils Sugar Spices BPP only 0.090 0.018 0.048 0.017 0.041 0.008 0.091 0.022 <0.081> <0.061> <0.071> <0.071> <0.071> <0.031> <0.081> <0.051> Nutrition education 0.010 0.046 0.120*** 0.017 0.065 0.029 0.012 0.012 <0.071> <0.051> <0.041> <0.061> <0.061> <0.021> <0.081> <0.041> Nutrition education with gender 0.039 0.051 0.023 0.092* 0.001 0.037 0.022 0.029 <0.071> <0.061> <0.051> <0.051> <0.051> <0.041> <0.071> <0.041> BPP x nutrition education 0.087 0.049 0.028 0.031 0.057 0.027 0.102 0.003 <0.071> <0.061> <0.071> <0.071> <0.061> <0.031> <0.081> <0.051> BPP x nutrition education with gender 0.057 0.034 0.002 0.020 0.068 0.025 0.099 0.001 <0.081> <0.071> <0.071> <0.071> <0.071> <0.031> <0.081> <0.051> Post 0.097 0.088 0.133** 0.018 0.066 0.058 0.145** 0.068 <0.081> <0.071> <0.061> <0.061> <0.061> <0.041> <0.061> <0.061> BPP x post 0.052 0.051 0.083 0.051 0.022 0.018 0.049 0.063 <0.091> <0.081> <0.071> <0.071> <0.071> <0.051> <0.061> <0.081> Nutrition education x post 0.024 0.023 0.009 0.082 0.010 0.058 0.078 0.060 <0.081> <0.081> <0.071> <0.071> <0.081> <0.071> <0.111> <0.081> Nutrition education with gender x post 0.024 0.010 0.126 0.063 0.014 --0.097 0.013 <0.101> <0.091> <0.081> <0.071> <0.071> --<0.081> <0.071> BPP x nutrition education x post 0.002 0.005 0.080 0.076 0.053 0.004 0.074 0.011 <0.091> <0.081> <0.071> <0.071> <0.071> <0.051> <0.061> <0.071> BPP x nutrition education with gender x post 0.031 0.033 0.136*** 0.064 0.031 0.010 0.092 0.003 <0.091> <0.081> <0.081> <0.071> <0.061> <0.051> <0.071> <0.081> Covariates No No No No No No No No Number of observations 2,164 2,164 2,164 2,164 2,164 2,035 2,164 2,164 Log Likelihood 1297 1257 1165 1491 1122 450 1233 1073 Note: Marginal effects are reported; Cluster robust standard errors are in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 87

87 Table 38. Logistic regression marginal effects results of treatment on 24 hour consumption of fruit and vegetable groups including covariates Treatment Vitamin A rich vegetables Tubers and roots Leafy green vegetables Other vegetables Vitamin A rich fruit Other fruit BPP only 0.056 0.031 0.052 0.003 0.009 0.049 <0.051> <0.051> <0.061> <0.047> <0.041> <0.044> Nutrition education 0.024 0.068 0.043 0.065* 0.001 0.027 <0.044> <0.048> <0.060> <0.039> <0.053> <0.054> Nutrition education with gender 0.015 0.083* 0.088 0.040 0.045 0.010 <0.041> <0.045> <0.075> <0.045> <0.048> <0.043> BPP x nutrition education 0.017 0.021 0.105* 0.032 0.006 0.013 <0.049> <0.049> <0.061> <0.040> <0.041> <0.042> BPP x nutrition education with gender 0.011 0.020 0.003 0.042 0.0816** 0.037 <0.056> <0.050> <0.063> <0.042> <0.038> <0.043> Post 0.166*** 0.095** 0.059 0.170*** 0.118** 0.053 <0.049> <0.047> <0.082> <0.059> <0.048> <0.062> BPP x post 0.023 0.012 0.101 0.053 0.036 0.060 <0.062> <0.061> <0.093> <0.073> <0.058> <0.071> Nutrition education x post 0.127* 0.035 0.087 0.052 0.019 0.018 <0.068> <0.069> <0.082> <0.076> <0.057> <0.069> Nutrition education with gender x post 0.072 0.111* 0.074 0.045 0.094 0.011 <0.066> <0.067> <0.108> <0.070> <0.064> <0.085> BPP x nutrition education x post 0.066 0.033 0.163* 0.055 0.035 0.002 <0.064> <0.062> <0.092> <0.069> <0.056> <0.073> BPP x nutrition education with gender x post 0.022 0.052 0.118 0.034 0.104* 0.007 <0.069> <0.062> <0.093> <0.074> <0.054> <0.075> Mymensingh district 0.172*** 0.311*** 0.175*** 0.375*** 0.176*** 0.228*** <0.028> <0.027> <0.032> <0.025> <0.024> <0.036>

PAGE 88

88 Table 38. Continued. Treatment Vitamin A rich vegetables Tubers and roots Leafy green vegetables Other vegetables Vitamin A rich fruit Other fruit Gender (female = 1) 0.021 0.036 0.053 0.109** 0.024 0.0781** <0.049> <0.041> <0.045> <0.053> <0.024> <0.037> Age 0.001 0.001 4.00E 04 2.80E 04 0.001 0.001 <0.001> <0.001> <0.001> <0.001> <0.001> <0.001> Primary school (highest level) 0.0571** 0.022 0.059** 0.004 0.0592*** 0.003 <0.027> <0.023> <0.029> <0.023> <0.019> <0.028> Junior secondary school (highest level) 0.089** 0.002 0.040 0.025 0.0695*** 0.046 <0.036> <0.038> <0.033> <0.029> <0.025> <0.031> Secondary school (highest level) 0.093** 0.054 0.047 0.038 0.0593*** 0.063 <0.040> <0.043> <0.040> <0.041> <0.022> <0.043> SSC pass (highest level) 0.075 0.031 0.008 0.035 0.0558** 0.010 <0.047> <0.058> <0.050> <0.052> <0.027> <0.044> Postsecondary education 0.077 0.022 0.018 0.017 0.0677** 0.062 <0.047> <0.049> <0.047> <0.051> <0.031> <0.039> Responsible for food preparation 0.053 0.080** 0.005 0.0692** 0.011 0.002 <0.038> <0.040> <0.046> <0.034> <0.019> <0.034> Total cultivable land (decimals) 0.001* 9.57E 05 9.07E 06 2.65E 05 7.92e 05*** 1.13E 04** <0.006> <1.62E 04> <7.29E 05> <7.60E 05> <2.43E 05> <5.57E 05> Farm diversity (count of food groups produced) 0.001 0.006 0.005 0.001* 0.003 0.0176*** <0.006> <0.005> <0.005> <0.005> <0.004> <0.006> Number of observations 2,161 2,161 2,161 2,161 2,161 2,161 Log Likelihood 1149 1071 1442 1088 542.2 1003 Note: Marginal effects are reported; Cluster robust standard errors are in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 89

89 Table 39. Logistic regression marginal effects results of treatment on 24 hour consumption of other food groups including covariates Treatment Meat Eggs Fish Pulses Dairy Fats and oils Sugar Spices BPP only 0.099 0.041 0.034 0.007 0.041 0.001 0.041 0.020 <0.070> <0.056> <0.046> <0.049> <0.050> <0.022> <0.054> <0.044> Nutrition education 4.27E 04 0.047 0.112*** 0.039 0.043 0.022 0.007 0.001 <0.060> <0.049> <0.037> <0.063> <0.056> <0.023> <0.067> <0.042> Nutrition education with gender 0.022 0.068 0.047 0.053 0.029 0.045 0.033 0.014 <0.062> <0.061> <0.048> <0.052> <0.051> <0.035> <0.064> <0.038> BPP x nutrition education 0.101 0.069 0.017 0.053 0.055 0.021 0.049 0.040 <0.064> <0.056> <0.047> <0.046> <0.048> <0.021> <0.049> <0.045> BPP x nutrition education with gender 0.072 0.047 0.009 0.006 0.056 0.020 0.040 0.036 <0.067> <0.057> <0.048> <0.050> <0.044> <0.023> <0.056> <0.044> Post 0.090 0.093 0.130** 0.022 0.073 0.060 0.134*** 0.072 <0.077> <0.068> <0.059> <0.058> <0.052> <0.037> <0.051> <0.061> BPP x post 0.043 0.048 0.078 0.053 0.024 0.015 0.036 0.059 <0.094> <0.079> <0.070> <0.071> <0.064> <0.048> <0.055> <0.083> Nutrition education x post 0.012 0.019 0.019 0.097 0.007 0.053 0.080 0.057 <0.076> <0.081> <0.067> <0.066> <0.074> <0.067> <0.102> <0.080> Nutrition education with gender x post 0.033 0.014 0.122 0.061 0.015 0.094 0.014 <0.092> <0.093> <0.075> <0.066> <0.069> -<0.073> <0.071> BPP x nutrition education x post 0.007 0.002 0.078 0.076 0.054 0.002 0.066 0.007 <0.090> <0.079> <0.071> <0.065> <0.064> <0.044> <0.060> <0.073> BPP x nutrition education with gender x post 0.022 0.029 0.137* 0.057 0.035 0.006 0.080 0.005 <0.088> <0.080> <0.070> <0.070> <0.058> <0.045> <0.061> <0.083> Mymensingh district 0.145*** 0.139*** 0.301*** 0.320*** 0.263*** 0.0743*** 0.283*** 0.0732** <0.041> <0.023> <0.031> <0.025> <0.024> <0.022> <0.028> <0.035>

PAGE 90

90 Table 39. Continued. Treatment Meat Eggs Fish Pulses Dairy Fats and oils Sugar Spices Gender (female = 1) 0.010 0.0751* 0.022 0.061 0.037 0.011 0.115*** 0.152*** <0.043> <0.039> <0.041> <0.045> <0.036> <0.020> <0.041> <0.037> Age 0.00184** 0.001 1.95E 04 4.21E 04 0.00186** 4.85E 04 0.001 3.36E 04 <0.001> <0.001> <0.001> <0.001> <0.001> <4.45E 04> <0.001> <0.001> Primary school (highest level) 0.0445** 0.010 0.014 0.004 0.0944*** 0.002 0.0674** 0.014 <0.022> <0.028> <0.020> <0.028> <0.022> <0.011> <0.030> <0.021> Junior secondary school (highest level) 0.0693** 0.007 0.020 0.040 0.104*** 0.017 0.070 0.009 <0.034> <0.034> <0.033> <0.036> <0.025> <0.021> <0.043> <0.035> Secondary school (highest level) 0.101*** 0.001 0.010 0.017 0.149*** 0.009 0.108** 0.052 <0.038> <0.044> <0.041> <0.053> <0.029> <0.028> <0.050> <0.056> SSC pass (highest level) 0.060 0.033 0.029 0.0774* 0.0870*** 0.031 0.0614* 0.064 <0.038> <0.069> <0.064> <0.042> <0.033> <0.025> <0.034> <0.046> Postsecondary education 0.167*** 0.035 0.111** 0.025 0.057 0.034 0.067 0.013 <0.042> <0.051> <0.050> <0.053> <0.038> <0.028> <0.048> <0.041> Responsible for food preparation 0.012 0.008 0.021 0.045 0.005 0.005 0.039 0.004 <0.044> <0.035> <0.036> <0.035> <0.030> <0.015> <0.038> <0.033> Total cultivable land (decimals) 2.41E 04*** 9.71e 05** 1.86E 05 9.84E 06 1.35E 04** 1.25E 06 2.40E 04*** 3.25E 05 <7.53E 05> <4.37E 05> <5.61E 05> <7.34E 05> <6.09E 05> <3.60E 05> <7.54E 05> <8.77E 05> Farm diversity (count of food groups produced) 0.009 0.010 0.0108** 0.0107* 0.0270*** 0.00569* 0.005 0.0124** <0.007> <0.007> <0.005> <0.006> <0.005> <0.003> <0.006> <0.005> Number of observations 2,161 2,161 2,161 2,161 2,161 2,032 2,161 2,161 Log Likelihood 1201 1215 944 1313 903 418 986 1038 Note: Marginal effects are reported; Cluster robust standard errors are in <> *** p<0.01, ** p<0.05, p<0.1

PAGE 91

91 Table 310. Estimated treatment effects on 7 day food consumption score (FCS) (1) (2) VARIABLES FCS FCS BPP only 0.296 0.328 [3.879] [1.635] Nutrition education 1.666 0.685 [1.658] [1.482] Nutrition education with gender 2.116 0.182 [1.463] [1.312] BPP x nutrition education 0.599 0.876 [3.660] [1.491] BPP x nutrition education with gender 1.376 1.519 [3.755] [1.600] Post 0.175 0.680 [1.597] [1.461] BPP x post 3.500* 3.234* [1.950] [1.810] Nutrition education x post 3.454 2.690 [2.279] [2.237] Nutrition education with gender x post 3.412** 3.077* [1.669] [1.717] BPP x nutrition education x post 3.678* 3.384* [2.049] [1.850] BPP x nutrition education with gender x post 3.217 2.540 [2.052] [1.924] Mymensingh district 18.914*** [0.867] Gender (female = 1) 0.136 [1.169] Age 0.013 [0.029] Primary school (highest level) 0.333 [0.808] Junior secondary school (highest level) 3.151*** [1.097] Secondary school (highest level) 4.955*** [1.360] SSC pass (highest level) 1.976 [1.284] Postsecondary education 4.133*** [1.266] Responsible for food preparation 1.977* [1.078]

PAGE 92

92 Table 310. Continued. (1) (2) VARIABLES IDDS IDDS Total cultivable land (decimals) 0.006** [0.002] Farm diversity (count of food groups produced) 1.818*** [0.153] Constant 67.719*** 47.669*** [2.710] [2.787] Number of o bservations 2,130 2,127 R squared 0.014 0.523 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 93

93 CHAPTER 4 MARKET ACCESS AND NUTRITIONSENSITIVE AGRICULTURE: A CONCEPTUAL FRAMEWORK AND EMPIRICAL FINDINGS Background Despite progress in recent years, 795 million people in the world are classified as undernourished, or unable to meet the dietary energy requirements for a healthy and active life (FAO, IFAD and WFP, 2015). Micronutrient deficiencies due to low dietary quality, often called hidden hunger are even more prevalent (FAO, IFAD, and WFP, 2015). In the developing world, 40% of the population l acks sufficient iron in their diet, and over one in three children suffer from vitamin A deficiencies (World Bank, 2006). Improving dietary diversity by consuming a wider variety of food groups can lead to improved nutrition by increasing the intake of mic ronutrients such as vitamin A and iron ( Arimond and Ruel, 2004; Hatloy et al., 1998; Torheim et al., 2004). To encourage the consumption of a larger variety of food groups, many development initiatives promote the adoption of new technologies or practices to improve agricultural production. The linkages between agriculture and nutrition can be identified as two farm based pathways towards improved nutrition: 1) increasing income through the sale of agricultural products, thus allowing for the purchase of mo re diverse foods, and 2) household consumption of food produced on the farm (Carletto, 2015; Chung, 2012). There may be a tendency to think of these farm based pathways as mutually exclusive, where some households engage in commercial agriculture and others operate at subsistence farming levels. In reality, many smallholder farmers are faced with the decision to sell or to consume each product they produce. Oftentimes households supplement farm revenue with off farm income, and the earnings from off fa rm labor may be used to purchase food from the market. Thus, while farm production may impact nutrition decisions, the households access to markets for selling products, buying food, and earning off farm labor should also be considered.

PAGE 94

94 Recent empirical s tudies show a positive relationship between farm production diversity and household dietary diversity (Dillon et al., 2015; Jones et al., 2014; Keding et al., 2012; Kumar et al., 2015). Livestock production is also associated with higher consumption of ani mal source foods, thus increasing household dietary diversity (Azzari et al., 2015; Rawlins et al., 2014). However, few of these studies consider the role of market access on production and consumption decisions (Sibhatu et al., 2015). Jones et al. (2014) includes an indicator to account for whether the household is market oriented or engages in subsistence farming. The indicator is measured as the proportion of food consumed in the previous week sourced from own production. The results suggest that farm ho useholds who are more market oriented have higher dietary diversity. Chege et al. (2014) find farm household participation in supermarket contracts leads to an increase in nutrient intake for vitamin A, iron, and zinc due to increases in income and househo ld vegetable production. In a multicountry analysis, Sibhatu et al. (2015) find access to agricultural markets positively impacts household dietary diversity, even more so than production diversity. The study finds household diets are more diverse among farmers who produce commercial cash crops compared to subsistence farmers who produce a larger variety of products. The authors suggest that encouraging farm diversity is not always the solution to improving dietary diversity, and efforts to improve market access may be a better investment. This essay expands on the previous literature to investigate the relationship between agricultural production, household dietary diversity, and access to markets. First we present a theoretical framework to formalize th e pathways from agriculture to nutrition. Barnum and Squires (1979) agricultural household model is expanded to include market participation and nutrition. An empirical analysis of household level survey data then follows to investigate the

PAGE 95

95 effects of mar ket participation and farm diversity on the decisions to produce and consume a variety of food groups in rural Bangladesh. The results suggest households with greater levels of participation in buyers markets consume more micronutrient rich foods. Partici pation in markets for selling agricultural products increases farm diversity but decreases dietary diversity. Further investigation is needed to determine whether households are diversifying production and selling high value items such as fruits and vegeta bles in order to purchase more staple food items, namely rice. Food Insecurity in Bangladesh Rural areas in Bangladesh are particularly prone to undernourishment and insufficient dietary quality (Ahmed et al., 2013). In a nationally representative study, 35.5% of rural households are food energy deficient, meaning they cannot afford an adequate diet to meet daily energy requirements (Ahmed et al., 2013). Rural households spend the highest share of their total budget, 60% on food expenditures (Ahmed et al. 2013). On average, Bangladeshi households spend 20% to 50 % of their total food budget on rice, depending on income, and no other food item accounts for more than 10 % of the food budget (Ahmed et al., 2013). The primary strategy for achieving food securit y in Bangladesh is self sufficiency in rice production. With investments in agricultural technology and irrigation, the country has largely achieved this goal. Rice accounts for 77 % of total cropped area, yet micronutrient deficiencies persist due to the i nfrequent consumption of other food items such as fruit, vegetables, and meat (Ahmed et al., 2013). The dietary imbalance is worse for the poorest members of society and is driven by a lack of production diversity, income constraints, strong individual food preferences, and a lack of nutrition knowledge (Hossain et al., 2005). In addition to a nutrition poverty trap, a gendered nutrition gap exists in Bangladesh, where womens nutrient consumption is one and a half times lower than mens nutrient consumptio n (Hossain et al., 2005). Rashid et al. (2011) find lower

PAGE 96

96 dietary diversity and protein availability among female headed households than male headed households, and they go on to suggest that female headed households are at a dietary disadvantage due to th e limited mobility of women in Bangladesh. Typically, male family members go to the market, where they can access diverse foods. Femaleheaded households may not have adult male members, and may lack access to markets as a result. This paper aims to unders tand the means through which production diversity and market participation affect household consumption of protein, fruits, and vegetables in Bangladesh. Specifically, we explore household decisions to participate in markets for purchasing or selling nutri ent rich foods. Under perfect market conditions, an agricultural household would be indifferent between purchasing food at the market versus consuming food produced at home. The price of food from either source, purchased or produced, would be equal after accounting for the transaction costs associated with buying or selling the food item. In this case, a household would separately consider consumption decisions to maximize utility and production decisions to maximize profits. When market imperfections exis t, however, the consumption and production models are no longer separable ( Muller, 2009; Singh et al., 1986). Rather, the household makes simultaneous production and consumption decisions to maximize utility subject to a production function as well as time and income constraints. Nutrition, Markets and the Agricultural Household Model The nonseparable model where households jointly consider consumption and production decisions is referred to as the agricultural household model (Barnum and Squire, 1979; Sing h et al., 1986) and builds on the household production model of Becker (1965). Based on the framework for health and nutrition developed by Behrman and Deolalikar (1988), we expand the agricultural household model so that household utility is not only a fu nction of the goods and leisure consumed by the household, but also the health of its members. Nutrition is an input to

PAGE 97

97 health delivered through the consumption of food items, which contain a combination of macronutrients (carbohydrates, fats, and proteins ) and micronutrients (vitamins and minerals). To expand the agricultural household model such that it includes nutrition, we consider all goods, whether purchased or produced, to be a function of nutrients, N and other attributes,

PAGE 98

98 (4 9) In E quation (41), household utility, is a function of household health, consumed goods and services, that are produced at home or purchased from the market, leisure time consumed by the household, and other household characteristics and preferences, Household health, is a function of the nutrients and attributes consumed through goods and services produced at home or purchased from the market, and household characteristics that affect health, such as dwelling conditions, sanitation, or genetic endowment of family members. The consumption of goods and services that affect health include food items as well as nonfoo d items such as medicines, immunizations, and doctor visits. Food and nonfood items can affect health positively or negatively, depending on the good. For example, consuming sugary foods or cigarettes negatively impacts health, while consuming leafy gree n vegetables or proper doses of medicine has positive health benefits. As sh own in E quation (46), each good is a function of its nutrients, N and other attributes, j produced at home, there is a unique production function, which maps a set of inputs to ou tputs for that good ( E quation (42)). The inputs, required to produce output j include variable inputs such as seed or fertilizer, and quasi fixed assets such as land and capital. The production of good j also depends on labor, which includes the unpaid labor of family members working on the farm and hired labor. Following the original design of the agricultural household model in Barnum and Squire (1979), E quation ( 43) indicates that a household is a net seller or net buyer of time, T based on household labor allocation. Time, T is the sum of total labo r used in the production of household goods, wage labor, and leisure includes

PAGE 99

99 hired labor as well as unpaid time of the family members allocated t o own production. Lw reflects net changes to total labor that arise as a household buys and sells labor. If then wage labor is hired and external workers contribute to household production. If then wage labor is sold and family members earn wages for off farm labor. As noted in E quation (41), utility is a function of leisure, consumed, which is the difference between time, T and labor, A household can be a net seller or a net buyer of any good based on the difference between household production and consumption of that item. Based on the household surplus or shortage of a good, the household will enter the market to sell or buy good j E quation (47) defines as an indicator for market participation. A household whose optimal decision includes participation in the market for selling good j will produce a surplus, As the household enter s the market for selling, is equal to 1, indicating participation in the sellers market. A household facing a shortage, will enter the market for buying good j and equal to 1 indicates participation in the buyers market. For a household in autarky, who neither sells nor buys item j in the market, and are equal to zero, indicating no ma rke t participation. Per E quation (48), is defined as the quantity of units sold or bought in the market. Following Barrett (2008), we assume the household faces different prices, or for good j depending on whether the good is purchased in the market or sold in the m arket,

PAGE 100

100 respectively. Equation (4 9) shows the difference in price and reflects the transaction costs, associated with market participation to buy or sell good j The transaction costs depend on the households access to assets, A (i.e. land, labor, capital), public services, G such as roads and extension services, and off farm income, W (Barrett, 2008). The transact ion costs for a commodity vary depending on the scope of market transactions, the presence of intermediaries along the value chain for good j such as traders and brokers, and a set of household specific characteristics, such as gender, education, and social status of the head of household and other factors that affect the decision makers search costs and ability to participate in markets. Under perfect market conditions, there would be no difference in the buying and selling price, would be equal to and the household utility model would be separable. However, this is rarely the case in a developing country such as Bangladesh. Households face a variety of market failures such as a lack of infrastructure, information asymmetries, and restricted access to markets based on cultural norms such as gender roles. Thus, consumption and production decisions among rural households in Bangladesh are nonseparable. To model this, the household maximizes utility subject to the full income constraint in E quation (44). T he left hand side of E quation (44) includes the total cost for units of all goods and services j purchased in the market at price As defined above, indicates participa tion in the market for buying goods. when a household does not purchase good j from the market. Th e right hand side of E quation (44) reflects the total income of the household as the sum of wages earned off farm plus profits earned from agricultural production. Wages, w

PAGE 101

101 can positively or negatively contribute to income depending on whether the household is a net seller ( ) or net buyer ( ) of labor. Agricultural profits are calculated as farm revenue minus production costs. Farm revenue is calculated as the product of the price of item j , the quantity of marketable surplus, and which indicates participation in the sellers market for item j when sold in the market. Production costs are calculated based on the wage of hired labor, w and the cost of inp uts used for production, A household in autarky neither buys nor sells product j in the market, but consumes all units of the good produced at home; thus the price of an autarkic good is equal to the production costs. To optimize production and consumption decisions in this non separable agricultural household model, we maximize E quat ion (4 1) subject to constraints in Equations (4 2) through (4 4) using the following Lagrangian function: (4 10) We can further define the Langrangian by substitution, (4 11) The choice variables in this model include the optimal bundle of goods and services consumed, leisure, and the optimal allocation of productive resources including total labor input , n et quantity of labor sold, and inputs, used in production, Thus, the first order conditions are: (4 12)

PAGE 102

102 (4 13) (4 14) (4 15) (4 16) The demand equations for the consumption of goods and services produced at home or purchased from the market can be found by sim ultaneously solving E quations (411) to (4 15). The consumption of goods and services is a function of market participation, We are particularly interested in the demands for specific food items that are high in vita min A, protein, and iron content and how the demands for these items are impacted by production decisions and market participation. To analyze the relationship between agricultural production, nutrition and market participation, we can use the consumption of different food groups as a proxy for nutrient demand. According to the FAO Guidelines for Measuring Household and Individual Dietary Diversity, all food items can be categorized into one of fifteen food groups based on nutrient content (Kennedy et al. 2010). Table 41 shows the fifteen food group categories and examples of Bengali food in the context of our study. The sum of the number of food groups consumed by household members is known as the household dietary diversity score (HDDS). The HDDS measur es the variety of food items consumed by an individual over a given period of time, usually 24 hours (Kennedy et al. 2010). Following the above theoretical model, dietary diversity is derived from a series of decisions optimizing the production and consumption of a variety of

PAGE 103

103 nutrients. Our model suggests dietary diversity is a function of market participation and input decisions that ultimately lead to the household production of food. The following empirical analysis investigates how production decisions affect nutrient intake and the role of market participation in those decisions. Methodology The empirical methodology in this paper builds on the model used by Sibhatu et al. (2015) to investigate the role of market access in household nutrition. We first estimate thirteen logit models using producti on of the food groups in T able 4 1 as the dependent variable. Production of a food group is a binary variable equal to 1 if, in the last 12 months, the household produced any commodity best classified by that p articular food group. The production of sugar and spices is not prevalent in our study area, thus the two food groups are omitted from the analysis on production. We then estimate fifteen logit models using consumpti on of the food groups in T able 4 1 as th e dependent variables. Specifically, consumption is a binary variable equal to 1 if anyone in the household consumed that particular food group and 0 otherwise. We are particularly interested in the associations between agricultural production diversity m arket participation and the consumption of different food groups. To measure market participation we include a vector of binary variables indicating whether or not a household participates in markets for selling agricultural pr oducts (commercial agricultural products, fruits and vegetables, fish, meat, eggs, dairy) and markets for buying food (all food items). The general form of the model is: (4 17)

PAGE 104

104 W here is a binary variable indicating household production ( or consumption ) of food group j is a vector of market participation variables for household i and is a vector of farm and household characteristics. Specifically, includes a continuous variable for farm size, measured by the total decimals of cultivable land available and other household characteristics such as district, household size, religion of the household, age of the household head, gender of the household head, education of the household head, poverty score, and household food insecurity access score. District is a categorical variable indicating the district in which the household is located Household size is the count of all household members A household member is defined as anyone who habitually eats and sleeps in the home, including those who have been absent less than six months and have not established another residence. Religion of the household is an indicator variable equal to 1 for Muslim and 0 for non Muslim. I n Bangladesh, nonMuslim households are predominantly Hindu. Age of the head of household is a continuous variable. Education is a set of categorical variables taking a value of 1 for the highest level of education completed by the household head (none, primary, second ary, junior secondary, higher secondary, vocational, bachelor of science, post graduate, professional) where completion of no education is the omitted variable. Gender of the household head is an indicator variable equal to 1 if the household head is femal e and 0 otherwise. The poverty score is calculated by weighting responses for a set of simple poverty scorecard questions for Bangladesh, following Schreiner (2013). The poverty score, ranging 0 to 100, indicates the likelihood that household expenditures are below the national poverty line where households with a score of 0 are most likely to fall below the poverty line. The household food insecurity access score (HFIAS) is a continuous variable measuring the degree of food insecurity in the household in t he last 30 days. The score is calculated based on household responses to a number of questions about food vulnerability and

PAGE 105

105 responses to a lack of access to food in the last 30 days. A higher score [0,27] indicates that the household is more food insecure (Coates et al., 2007). Equation (417) models the production as well as the consumption of each food group The production and consumption models include market participation and all other characteristics defined above. However, to the consumption models w e also add farm diversity as an explanatory variable to investigate the relationship between production and consumption decisions. Farm diversity is measured as the count of food groups produced by the household in the last year [0,13]. In addition to the production and consumption of specific food groups, we are interested in the effect of market participation on farm diversity and dietary diversity. Th us, we also estimate E quation (417) using the farm diversity score and the household dietary diversity s core (HDDS) as our dependent variables. In this case, farm diversity score is the number of food groups [0,13] produced in the last year. HDDS is the number of food groups consumed in the last 24 hours. The diversity models are estimated using Poisson regr ession since the dependent variables are count variables. The HDDS model includes a binary variable indicating whether or not the individual is the primary person responsible for preparing food for the household. Otherwise, the covariates for the farm diversity and household dietary diversity models remain the same as the logit regressions described above. The data in our analysis comes from a clustered sample of communities where our partner agencies serve. T o account for this sampling strategy, the standa rd errors in each model are clustered at the community level. Data Collection The data for this analysis comes from a survey of 1,130 households in the Mymensingh and Borguna districts of Bangladesh. Mymensingh is a predominantly agricultural district located in northern Bangladesh, in the Dhaka division (Bangladesh Bureau of Statis tics, 2013).

PAGE 106

106 Agriculture accounts for 59.15 % of landholdings in Mymensingh (Bangladesh Bureau of Statistics, 2013). Rice, jute, and sugarcane are among the main cash crops and jackfruit, banana and pineapple are the main fruits produced in the Mymensingh d istrict (Bangladesh Bureau of Statistics, 2013). Borguna is a coastal district in southern Bangladesh in the Barisal division, bordering the Bay of Bengal. Of the total landholdings in Borguna district, 71.93 % include agricultural land (Bangladesh Bureau o f Statistics, 2013). Rice and pulses are among the main crops produced in Borguna, and the main fruits include coconut and banana (Bangladesh Bureau of Statistics, 2013). The survey of rural households was conducted from August to November 2016. This proj ect is part of a larger study funded by the USAID initiative Integrating Gender and Nutrition in Agricultural Extension Services (INGENAES) where we partnered with two institutions providing agricultural extension services: the Bangladesh Agricultural University Extension Center (BAUEC) and Shushilan, a local NGO. The survey participants were randomly selected from the membership lists of the two organizations in Mymensingh and Borguna, respectively. Shushilan is a national organization ; however for the purpose of this research we only engaged beneficiaries from the Managing Natural Resources by the Coastal Community (MaNaR) project BAUEC serves male and female farmers in 20 unions of the Mymensing sadar upazila1. The membership is 55% male and 45% fem ale. The MaNaR project primarily targets female beneficiaries, thus its membership is 94 % female and 5 % male. The Shushilan MaNar project area serves three unions in the Amtali upazila. Our survey sample included participants from 18 Shushilan communities and 35 BAUEC communities. 1 An upazila is equivalent to a county -level administrative area; a union is the sub -county level comprised of a cluster of villages

PAGE 107

107 The survey contained a household roster with demographics, simple poverty scorecard, agricultural production information (area planted and amount harvested, sold, kept for home consumption for each crop), household dietary divers ity (24 hour food frequency, 7 day food frequency, and source of food items), nutrition knowledge, market access, and household food insecurity access score questions. The sample includes 595 respondents from Mymensingh and 537 respondents from Borguna. I n Mymensingh, 94 % of the respondents live in male headed households and 6 % female headed. However, in Borguna, 75 % of the respondents reported femaleheaded households and 25% male headed. The gendered differences in the sample reflect the membership of th e partner organizations in the respective districts. Overall, we have a relatively large sample of female headed households, 15% compared to the national average of 12.5 % (World Bank, 2016b ). Table 4 2 presents the descriptive statistics for the farm and household characteristics of the total sample and by district. Within our sample, 3 6% of household heads have not completed any education and 33% of household heads have completed some primary school (class 1 to 6). The majority of households in Bangladesh identify Muslim as their religion (89.1 % ), followed by Hindu (Bangladesh Bureau of Statistics, 2013). Our sample has a slightly higher representation, 95 % of Muslim households. The average household size of our sample is 4.6 persons, consistent with the national average of 4.5 (Bangladesh Bureau of Statistics, 2013). According to the scorecard conversion table in Schreiner (2013), the mean poverty score of 4 8 indicates that 33.5 % of our sample is below the daily per capita expenditure poverty threshold o f US$1.25/day and there is a 19.6% likelihood that a household falls below the upper

PAGE 108

108 national poverty line of BDT2 52.64 per person per day. Indicators for agricultural production in this study include total access to cultivable land (in decimals) and the count of all crops and livestock. Cultivable land includes cropped land as well as fallow land, but excludes all dw elling and homestead land (Schreiner, 2013). The land may be owned, sharecropped, or rented. On average, our sample households have access to 73 decimals, or 0.73 acres, of cultivable land and produce three crops and animal source foods. To measure the le vel of participation in the buyers market, respondents were asked Did anyone in your household buy any food (from a market) to cook in the household in the last year? For a variety of products (commercial agricultural products, vegetables and fruits, fi sh, meat, eggs, and milk) respondents were also asked if they sell each of these products (at home, via a trader, or at the market). In our analysis, selling food is treated as a binary variable equal to 1 if the household sold any food p roducts and 0 othe rwise. Table 4 2 shows evidence of agricultural sales varying by district, which may suggest a difference in access to markets between districts. We further investigate the role of market participation in our logit models. The results are presented in the next section. Table 4 3 shows the mean and standard deviation of all dependent variables. The consumption of each food group is a binary variable equal to one if, in the last 24 hours, the respondent or anyone in the household consumed a food item that is categorized as that food group. As one would expect, cereal is the most frequently consumed food group. Ninety percent of households in our sample consumed cereals, namely rice, 24 hours prior to our survey. Vitamin A rich vegetables and fruits are among the lowest food groups consumed. Fish and pulses are the most frequently consumed sources of protein. 2 US $1 is approximately 78 BDT

PAGE 109

109 The production of each food group is a binary variable equal to one if, in the last 12 months, the household has produced a commodity with nutrients that most reflect that food group classification. Sixty three percent of our sample produces rice, whereas almost 90% of households in our sample produce vegetables (such as gourds), fruits (such as bananas), and dairy products (cow or goat milk). Regression Results The results from the logit regressions for household production and consumption of each of the food groups are presen ted as odds ratios in T ables A 8 and A 9, respectively. The results in T able A 8 show that households living in Mymensingh, in nort hern Bangladesh, are more likely to produce several food groups. Given the difference in soil fertility and climate it is not surprising that district affects the decision to produce food crops. The correlation between district and production diversity is perhaps due to the suitable climate for producing fruits and vegetables in Mymensingh. Borguna, on the other hand, is located near the Bay of Bengal and is one of the areas most affected by climate change (Bangladesh Bureau of Statistics, 2013). Landholdings also impact production decisions. Households with access to more cultivable land are more likely to produce all vegetable classifications, vitamin A rich fruit, pulses, dairy, and oilseeds. The results in T able A 8 indicate that the age and education o f the household head also impacts the decision making process for food production. Households headed by older individuals are more likely to produce white roots and tubers, other vegetables, fish, and pulses. If the household head passed the national SSC exam the household is twice as likely to produce vitamin A rich fruits, fish, and other vegetables compared to households whose heads have no education. Interestingly, larger households are more likely to produce animal source foods (eggs, fish, and dairy), which may reflect the labor intense nature of fish and livestock.

PAGE 110

110 Poverty and household food insecurity also appear to be correlated with the production of certain types of food H ouseholds with a higher poverty scorethose less likely to fall below th e poverty line are less likely to produce cereals. It is possible that more impoverished households are producing cereals for home consumption. We see a positive correlation between households with a higher poverty score and the likelihood of producing vit amin A rich fruit and fish. Households with higher food insecurity, indicated by the HFIAS, are less likely to produce several nutrient rich food groups. Overall, participation in the market for purchasing food does not have a statistically significant co rrelation with farm diversity. However, households who participate in buyers markets are less likely to produce oilseeds presumably because fats and oils require a great deal of processing and the items are readily available in the market. As one would ex pect, the likelihood that a household produces items in nearly all food groups is higher if the household participates in markets for selling food products. Similar to t he production results, T able A 9 shows that district is a statistically significant pr edictor of the likelihood that a food group is consumed within the household. Following intuition, the results of the logit regressions suggest a higher household food insecurity access score (HFIAS) reduces the likelihood of consuming many of the analyzed food groups, particularly those high in micronutrients such as vitamin A and protein. The HFIAS measures hunger as well as access to food. In several cases, farm diversity (the count of food groups produced by the household) increases the likelihood that a food group is consumed. For five of the fifteen food groups, buying food from the market significantly impacts the consumption of that food group. Households who buy food from the market are nearly 3 times more likely to consume cereals, namely rice, and vegetables ( e.g gourds, tomatoes, etc.) than households that

PAGE 111

111 do not buy food from the market. The likelihood of consuming eggs, dairy products, and sugar is also higher among households who purchase food from the market. On the contrary, households who s ell products at the market are less likely to consume leafy green vegetables, other vegetables, fruit, oilseeds, sugar, and spices. Tables 4 4 and 45 present the marginal effects from Poisson regressions on the farm diversity score and household dietary diversity score, respectively. Coefficient estimates from the Poisson regres sions can be found in Tables A 10 and A 11. The results for the pooled model in T able 4 4 show households in Mymensingh produce 1. 6 more food groups on average than households in B orguna. Households where the head passed the SSC exam produce 0.4 more food groups on average than households headed by an individual with no education. Disaggregating the data by district shows households headed by an individual who passed the SSC exam in Borguna are more specialized, producing 1.1 fewer food groups on average compared to households where the household head has no education. Household size and age of the household head are positively correlated with the average number of food groups produced in Mymensingh but not Borguna. Access to one additional decimal of cultivable land results in 0.004 more food groups produced on average. Contrary to economic theory, which suggests access to markets would promote specialization, our results suggest se lling food at the market is associated with higher farm diversity. A household who participates in markets for selling products produces 1.1 additional food groups on average compared to households that do not sell commodities. District is also a statisti cally significant predictor of household dietar y diversity, as shown in T able 4 5. Results from the pooled regression show that households in Mymensingh consume almost three additional food groups on average compared to Borguna. The findings

PAGE 112

112 reveal a posit ive correlation between farm diversity and household dietary diversity A household that produces one additional food group consumes 0.28 additional food groups on average. Conversely, households that sell agricultural products at the market consume 0.5 fe wer food groups on average compared to households who do not participate in markets for selling products. Purchasing food at the market has no statistically significant effect on household dietary diversity. Discussion To formalize the theoretical framewo rk linking agriculture and nutrition, this paper expands Barnum and Squires (1979) agricultural household model. Our framework contributes to the existing literature by including not only nutrition but also transaction costs associated with market partici pation in the model. Using household survey data from two districts of Bangladesh, we then empirically investigate the relationship s between market participation household production and consumption decisions for a variety of food groups. The results suggest that market participation indeed influences household decisions surrounding the production and consumption of different food groups. Some ambiguity surrounds the net result of market participation, however. Contrary to the theory of specialization, h ouseholds who engage in markets for selling agricultural products produce a larger variety of farm products. At the same time, participation in sellers markets is correlated with a lower household dietary diversity score on average. Participation in sellers markets may, therefore, incentivize households to produce and sell higher value, nutritious food items such as fruits and vegetables in order to purchase more staple crops such as rice. The analysis on the consumption of food groups, however, shows that participation in buyers markets is positively correlated with dietary diversity. Consistent with expectations, households with access to markets have a wider variety of food ite ms to choose from. These results warrant further analysis

PAGE 113

113 of the relationship between market participation, agricultural production and food group consumption. We recognize the potential endogeneity in our models, as market participation affects both produ ction and consumption. Further research should use an instrumental variable approach to analyzing the effect of market participation on production and consumption diversity. However, at this time, we do not have a strong instrument to use in the analysis. In addition to the role of market participation on production and consumption decisions, the results show consumption and production patterns largely vary by district in Bangladesh. Farm and dietary diversity are higher among households in northern Bangla desh (Mymensingh), where growing conditions are more favorable in terms of climate, soil fertility, soil salinity, and perhaps access to agricultural extension services through the Bangladesh Agricultural University Extension Center (BAUEC). Further resear ch on this topic should explore the extent to which geographical location impacts nutrition due to market access and price differentials in agricultural markets. B ecause our sample is not nationally representative, the variation in consumption and product ion patterns may also reflect the different project activities offered by our partner agencies in the respective districts. The partner organizations were selected due to their existing collaboration with our funding agency, under the criteria that their b eneficiaries had not received previous nutrition training. Thus, our results have some degree of bias from the sampling strategy. In particular, Shushilan, the partner agency in Borguna, focuses on climate resilient innovations for the extreme poor and landless. These households likely experience greater restrictions on market access due to financial constraints. Furthermore, Shushilan beneficiary households are predominately female headed. In Bangladesh, traditional gender norms limit

PAGE 114

114 womens access to mar kets. Future research should incorporate a measure of womens empowerment in addition to the gender of the head of household.

PAGE 115

115 Table 41. FAO food groups and examples Food Group Examples (this list is not comprehensive) Cereals Rice, bread, wheat, biscuits, puffed rice White roots and tubers White potatoes, turnips Vitamin A rich vegetables and tubers Pumpkin, carrots, squash, orange flesh sweet potato Dark leafy green vegetables Amaranth, spinach, taro leaf, pumpkin leaf, jute leaf Other vegetables Cucumber, radish, tomato, pepper, beans Vitamin A rich fruits Ripe papaya, ripe mango, orange, grapefruit Other fruits Banana, apple, jackfruit, watermelon, lychee Meat Chicken, beef, poultry, lamb, liver, kidney Eggs Eggs Fish and seafood Carp, hilsha fish, shrimp, prawn Legumes, nuts, and seeds Lentils, dal, black gram, kheshari, mung bean Milk and milk products Cow milk, goat milk, powdered milk, yogurt, curd, cheese Oils and fats Butter, ghee, mustard oil, soybean oil Sweets Sugar, honey, molasses, chocolate, ice cream, soda Spices, condiments, beverages Black pepper, salt, fish powder, chili sauce, cumin, coffee, tea

PAGE 116

116 Table 42. Descriptive statistics covariates, by district All Borguna Mymensingh Variable Mean St. dev. Mean St. dev. Mean St. dev. Mymensingh 0.52 0.50 0.00 0.00 1.00 0.00 Muslim 0.95 0.21 0.92 0.28 0.99 0.11 Household food insecurity access score 1.96 3.70 3.54 4.39 0.52 2.07 Male headed 0.85 0.36 0.75 0.43 0.94 0.23 Primary school 0.33 0.47 0.44 0.50 0.24 0.42 Junior secondary school 0.11 0.31 0.06 0.24 0.15 0.36 Secondary school 0.06 0.24 0.02 0.13 0.10 0.30 SSC pass 0.05 0.23 0.01 0.09 0.10 0.30 Post secondary education 0.07 0.25 0.00 0.06 0.12 0.33 Age of household head 45.14 12.69 44.53 12.48 45.69 12.86 Household size 4.62 1.68 4.13 1.44 5.07 1.76 Poverty 48.31 16.06 38.13 10.99 57.56 14.25 Total ag land (decimals) 72.88 104.06 49.39 97.96 94.23 104.92 Diversity of crops and livestock (count) 3.36 2.22 2.15 1.82 4.46 1.97 Buy food at market 0.40 0.49 0.61 0.49 0.21 0.40 Sell food at market 0.47 0.50 0.30 0.46 0.63 0.48

PAGE 117

117 Table 43. Mean household consumption and production of food groups All Borguna Mymensingh Dependent Variable Consumption Production Consumption Production Consumption Production Cereals 1.00 0.63 1.00 0.46 1.00 0.78 Vitamin A rich vegetables 0.18 0.09 0.17 0.01 0.20 0.17 White roots/tubers 0.77 0.07 0.66 0.07 0.86 0.07 Leafy green vegetables 0.53 0.17 0.49 0.05 0.56 0.27 Other vegetables 0.52 0.35 0.20 0.15 0.82 0.52 Vitamin A rich fruit 0.06 0.42 0.03 0.11 0.09 0.70 Other fruit 0.26 0.56 0.15 0.22 0.36 0.86 Meat 0.33 0.31 0.21 0.19 0.45 0.42 Eggs 0.24 0.58 0.18 0.47 0.28 0.67 Fish 0.75 0.27 0.58 0.12 0.90 0.41 Pulses 0.52 0.19 0.34 0.37 0.69 0.03 Dairy 0.22 0.29 0.08 0.26 0.34 0.31 Fats and oils 0.91 0.07 0.84 0.12 0.97 0.03 Sugar 0.24 -0.07 -0.39 -Spices 0.77 -0.72 -0.81 -Household Dietary Diversity Score 7.29 -5.72 -8.72 -(HDDS) <2.47> -<1.86> -<2.05> -Farm Diversity -3.36 -2.15 -4.46 (count of food groups produced) -<2.22> -<1.82> -<1.97>

PAGE 118

118 Table 44. Marginal effects from Poisson regression on farm diversity score (1) (2) (3) VARIABLES All Borguna Mymensingh District (Mymensingh = 1) 1.570*** --[0.290] --Religion (Muslim = 1) 0.461 0.506*** 0.684** [0.298] [0.142] [0.331] Household Food Insecurity Access Score 0.102*** 0.044 0.204*** [0.030] [0.030] [0.033] Male headed household 0.217 0.156 0.137 [0.155] [0.121] [0.324] Household head primary education 0.045 0.071 0.094 [0.172] [0.213] [0.205] Household head junior secondary education 0.243 0.096 0.381* [0.194] [0.414] [0.195] Household head secondary education 0.112 0.337 0.176 [0.183] [0.550] [0.251] Household head SSC pass 0.422* 1.128*** 0.753*** [0.220] [0.209] [0.278] Post secondary education 0.190 1.256*** 0.263 [0.277] [0.464] [0.371] Age of household head 0.008* 0.002 0.014** [0.004] [0.006] [0.006] Household size 0.084** 0.102 0.099** [0.033] [0.064] [0.041] Poverty score 0.003 0.007 0.001 [0.004] [0.009] [0.005] Total cultivable land (decimals) 0.004*** 0.003*** 0.004*** [0.001] [0.001] [0.001] Buy food at market 0.184 0.0881 0.349 [0.158] [0.236] [0.227] Sell food at market 1.101*** 1.025*** 1.160*** [0.128] [0.127] [0.203] Observations 1,130 535 595 Log Likelihood 2150 947.6 1186 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 119

119 Table 45. Marginal effects from Poisson regression on household dietary diversity score (1) (2) (3) VARIABLES All Borguna Mymensingh District (Mymensingh = 1) 2.383*** [0.169] Religion (Muslim = 1) 0.340 0.207 0.444 [0.218] [0.241] [0.328] Household Food Insecurity Access Score 0.093*** 0.082*** 0.085 [0.023] [0.015] [0.084] Male headed household 0.034 0.287 0.537** [0.229] [0.201] [0.256] Household head primary education 0.085 0.355*** 0.379* [0.146] [0.130] [0.213] Household head junior secondary education 0.200 0.039 0.475 [0.211] [0.371] [0.323] Household head secondary education 0.060 0.462 0.133 [0.201] [0.630] [0.261] Household head SSC pass 0.015 0.910* 0.170 [0.240] [0.476] [0.327] Post secondary education 0.254* 0.631** 0.529*** [0.149] [0.286] [0.203] Age of household head 1.74E 04 0.002 0.001 [0.004] [0.005] [0.006] Household size 0.017 0.076 0.003 [0.039] [0.067] [0.053] Poverty score 3.00E 03 0.012*** 0.008 [0.005] [0.005] [0.008] Total cultivable land (decimals) 0.001 0.002* 0.003*** [0.001] [0.001] [0.001] Farm diversity (count of food groups produced) 0.278*** 0.357*** 0.228*** [0.043] [0.051] [0.049] Buy food at market 0.218 0.071 0.434 [0.140] [0.158] [0.277] Sell food at market 0.474** 0.399* 0.402 [0.217] [0.219] [0.378] Respondent prepares food for the household 0.311** 0.328 0.470*** [0.137] [0.217] [0.172] Observations 1,130 535 595 Log Likelihood 2408 1084 1313 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 120

120 CHAPTER 5 CONCLUDING REMARKS Contributions While much of the existing literature focuses on supply side interventions for improving food security, this dissertation explore d pote ntial mechanisms for increas ing the demand for a diverse diet. In a randomized controlled trial, we analyze d the effectiveness of a behavioral economics nudge and a behavior change communication intervention designed to promote nutrition and dietary diversity among agricultural households in Bangladesh. Our research also investigate d the interdependence of nutrient rich food consumption and traditional constraints such as accessibility and affordability. We we re particularly interested in the potent ial for a FBDG icon to serve as a nutrition nudge In addition, we evaluated the effectiveness of nutrition education as a BCC tool Thus, o ur methods explore d the effectiveness of a behavioral economics nudge alone, behavior change communication alone, an d the combination of the two. While BCC is often used to promote health and nutrition in developing countries, to our knowledge no other study to date has produced empirical evidence on nudging nutrition decisions in a development context. W e evaluated th ese behavior change interventions in the short term, in a lab setting where food choices were observed as well as in the home environment using pre intervention and post intervention survey data to capture the long term effects. The lab setting remove d constraints that influence individuals food choices such as income and access to nutrient rich foods Measuring the outcomes in the home environment allowed us to make a more comprehensive evaluation of the interventions. First, we observe d whether the efficacy of either messaging mechanism change d in the home environment, where income and access constraints were restored Secondly, we measured the long term impact of repeated exposure to the nutrition nudge in the home

PAGE 121

121 environment. Even in the United St ates, where a growing body of literature investigates the use of behavioral economics in nutrition, few studies measure the effectiveness of food based dietary guidelines at nudging nutrition decisions in the home. Thus, our research informs literature on behavioral economics in nutrition for developed countries in addition to introducing novel methods for studying nutrition in the development context. The results suggest our behavior change interventions produce d weaker impacts in the home environment. Th us, this dissertation also investigated the role of income and access constraints on household nutrition decisions. We ma de a theoretical contribution to the economic framework linking agriculture and nutrition in developing countries. Specifically, we inc orporate d nutrition, market transaction costs, and the propensity to participate in markets into the agricultural household model. W e then empirically analyze d how market participation affects the variety of food groups produced and consumed by households in Bangladesh. Findings The experimental results indicate that short term exposure to a FBDG icon such as the BPP i s not an effective nudge by itself. However, nutrition education in a participatory workshop i s effective in the short run, and combining the icon with nutrition education increase s meal diversity The analysis of meal observation data found that nutrition education increased mean meal diversity by 2.1% The consumption of lentils and leafy green vegetables was higher among partici pants who received only nutrition education compared to the control. Similarly, nutrition and gender education led to a 0.017 increase on mean meal diversity compared to the control. There i s no statistically significant evidence that the BPP alone nudged participants to choose a greater variety of food. However, the BPP combined with nutrition education increased mean meal diversity by 2. 6% The BPP encouraged the consumption of lentils when the nudge was

PAGE 122

122 combined with nutrition education, but w e found no evidence that either nutrition intervention reduced rice consumption The long term analysis presented in C hapter 3 found mixed effects of the BPP and BCC interventions at home. Neither participatory training nor the nutrition nudge produced a change in 24hour individual dietary diversity. However, the mean food consumption score (a 7 day measure) increased by 3.5 points and 3.7 points among participants who were assigned to the BPP treatment and the BPP combined with nutrition education respectively F urthermore, nutrition education with the gender component increased the food consumption score in the home We suspect participants were better able to comply with the BP P and BCC messages over a 7 day period due to variation in consumption bundles based o n the availability and accessibility of food in the market. In summary, we found weak evidence that long term exposure to the BPP in the hom e environment increased dietary diversity. Given the discrepancy in treatment outcomes between the lab and home environments, and the dietary diversity measures, Chapter 4 explored the role of income and access constraints on household dietary diversity. A conceptual framework first e xtend ed the agricultural household model to include nutrition, demonstrat ing how market transaction costs affect both production and consumption decisions. Empirically, we then analyzed the correlation between market participation and decisions surrounding the production and consumption of nutrients. The results suggest that participat ion in sellers markets i s positively correlated with farm diversity as households that engage in sellers markets produce on average 1.1 additional food groups of 15 total possible food groups compared to subsistence households. Essentially, th ese finding s suggest that market orientation of the farm does not encourage specialization, but leads the farmer to produce a greater variety of food items. Our analysis revealed a negative relationship

PAGE 123

123 between participation in sellers markets and household dietary diversity. Households that s el l agricultural products at the market consume 0.5 fewer food groups on average compared to subsistence households As access to markets increase, there may be incentive for the household to sell more nutritious, high value foo d items in order to purchase more staple crops such as rice from the market. An obvious limitation of this analysis is the endogeneity of market participation, farm diversity, and dietary diversity. At this time, we do not have a strong instrument to corre ct for endogeneity. An analysis on price differentials would also strengthen this contribution ; however these data are difficult to obtain due to the abundance of price negotiation that occurs in the market. Policy Implications If, indeed, agricultural h ouseholds are selling nutritious food items and buying less nutritious food items for home consumption, then nutrition education has the potential to play a vital role in improving nutrition outcomes. Chapter s 2 and 3 s ought to inform agricultural extension agents and community health workers on the best practices for nutrition promotion. The results suggest a combined approach of behavior change communication with behavioral economics is most effective. Furthermore, long ter m exposure to a nutrition nudge in the home environment could create sustainable impacts, depending on the frequency of exposure and access to diverse food items. Among the participants who received the BPP, 11 % reported never using the BPP at home. In ma ny cases, individuals claimed to prefer the use of other plates. Additional consideration should be given to the design and practicality of nudges for the home environment, particularly in a developing country. Awareness of constraints such as literacy or gender norms surrounding decision making and access to markets is critical in the design and implementation of tools for behavior change.

PAGE 124

124 P recaution should be taken in the design of FBDG icons such as the BPP the complexity of the tool may impede the bene ficiarys ability to grasp messages in the short term. In the short term, our analysis found increase s in the consumption of lentils and leafy green vegetables drove the increase in meal diversity this may have be en due to the prominence of these food item s on the BPP. Furthermore, our short term results show ed that exposure to the BPP alone is not effective, but the combined interventions increased meal diversity. Combining methods of BCC with behavioral economics strategies is particularly important for uneducated beneficiaries who may be more reliant on verbal communication. Further research should investigate the extent to which these results translate to exposure to other messaging mechanisms such as posters or billboards. While our findings inform pra ctices for nutrition messaging, t he main implication of the theoretical model in Chapter 4 i s that nutrition sensitive agriculture initiatives must consider market accessibility. P romoting agricultural technology or nutrition awareness alone is not sufficient to improve dietary diversity Initiatives should take a comprehensive approach to nutrition sensitive agriculture, considering market accessibility as well as information and tools for behavior change. Future Research Chapter s 2 and 3 of this dissertat ion establish novel protocol for investigating the use of nudging to combat malnutrition in developing countries. The research methods can be extended to other developing countries to measure the effectiveness of local FBDG icons and other nutrition nudges as well as BCC strategies Further research should evaluate alternative mechanisms for changing behavior to improve nutrition among rural households in low income countries. The results from this dissertation suggest that traditional constraints such as i ncome

PAGE 125

125 and access to food markets may dampen the impact of nutrition interventions. Thus, incentive based program s such as conditional cash transfer s with nutrition nudge s should be explored. O ur theoretical contribution in Chapter 4 define d the role of nu trition in the utility maximization of agricultural households. However, in practice, households may not be aware of the importance of a diverse diet and hence may not consid er nutrient values of the food items they consume when maximizing utility The rel iance on culturally significant foods, such as rice in Bangladesh, is difficult to overcome. Thus, further studies should evaluate the impact of interventions that focus on the negative health imp acts of consuming too much rice. Alternatively, intervention s may emphasize the long term health benefits of consuming a greater variety of nutritious food items and the extent to which those benefits justify the additional cost of healthy food items.

PAGE 126

126 APPENDIX A SUPPLEMENTARY TABLES Table A 1. Covariate s ummary statistics by behavior change communication intervention (nutrition education) No education Nutrition education Nutrition and g ender education Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mymensingh district 0.52 0.50 0.53 0.50 0.53 0.50 Borguna district 0.48 0.50 0.47 0.50 0.47 0.50 Household (HH) monthly income (BDT) 10584.76 8717.20 11398.47 10921.08 10601.08 8123.49 Poverty score 48.42 16.29 47.90 15.63 48.67 16.04 Number of food groups produced 3.37 2.30 3.41 2.16 3.37 2.22 Household food insecurity access score (HFIAS) 1.91 3.63 1.97 3.73 2.03 3.81 Religion (Muslim = 1) 0.96 0.20 0.96 0.20 0.95 0.22 Male 0.28 0.45 0.30 0.46 0.28 0.45 Female 0.72 0.45 0.70 0.46 0.72 0.45 Age 38.43 12.80 38.35 12.85 37.17 11.58 Highest level of education None 0.35 0.48 0.38 0.49 0.35 0.48 Primary school 0.35 0.48 0.31 0.46 0.34 0.47 Junior secondary school 0.12 0.32 0.11 0.31 0.10 0.30 Secondary school 0.05 0.23 0.06 0.23 0.07 0.26 SSC pass 0.04 0.20 0.06 0.24 0.06 0.25 Postsecondary education 0.08 0.26 0.07 0.25 0.06 0.23 Spouse of household head 0.50 0.50 0.51 0.50 0.54 0.50 Other relationship to household head 0.13 0.33 0.14 0.35 0.10 0.30 Hunger at time of meal event 3.71 1.14 3.69 1.20 3.68 1.18 Baseline nutrition knowledge score 18.59 5.29 18.78 5.41 18.98 4.76

PAGE 127

127 Table A 2. Covariate summary statistics by behavioral economics intervention (Bengali Portion Plate) A (Regular, Regular) B (Regular, BPP) C (BPP, Regular) Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mymensingh district 0.56 0.50 0.52 0.50 0.51 0.50 Borguna district 0.44 0.50 0.48 0.50 0.49 0.50 Household (HH) monthly income (BDT) 11031.70 10131.77 10366.63 8564.72 11336.14 9520.77 Poverty score 49.03 16.81 47.88 15.48 48.13 15.69 Number of food groups produced 3.26 2.16 3.49 2.31 3.38 2.18 Household food insecurity access score (HFIAS) 1.79 3.41 2.17 4.00 1.92 3.67 Religion (Muslim = 1) 0.94 0.24 0.95 0.21 0.97 0.17 Male 0.18 0.38 0.42 0.49 0.25 0.43 Female 0.82 0.38 0.58 0.49 0.75 0.43 Age 36.84 12.02 39.11 12.61 37.76 12.52 Highest level of education None 0.31 0.46 0.38 0.49 0.39 0.49 Primary school 0.35 0.48 0.32 0.47 0.33 0.47 Junior secondary school 0.14 0.35 0.11 0.31 0.08 0.28 Secondary school 0.05 0.22 0.06 0.24 0.07 0.25 SSC pass 0.06 0.23 0.05 0.23 0.06 0.23 Postsecondary education 0.08 0.27 0.06 0.24 0.05 0.23 Spouse of household head 0.62 0.49 0.40 0.49 0.55 0.50 Other relationship to household head 0.13 0.34 0.09 0.29 0.15 0.36 Hunger at time of meal event 3.70 1.20 3.72 1.14 3.65 1.19 Baseline nutrition knowledge score 18.89 5.19 18.64 5.23 18.86 5.06

PAGE 128

128 Table A 3. Summary statistics by treatment groups defined in T able 21 Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 Group 8 Group 9 Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St. Dev. Mean St Dev. Mean St .Dev. Mean St .Dev. Mean St Dev. Mymensin gh district 0.58 0.49 0.59 0.49 0.52 0.50 0.51 0.50 0.51 0.50 0.54 0.50 0.48 0.50 0.51 0.50 0.52 0.50 Borguna district 0.42 0.49 0.41 0.49 0.48 0.50 0.49 0.50 0.49 0.50 0.46 0.50 0.52 0.50 0.49 0.50 0.48 0.50 Household (HH) monthly income (100 BDT) 103.39 72.37 122.65 127.96 104.66 93.93 97.56 83.06 110.48 98.92 102.63 71.59 119.99 104.69 110.09 101.54 111.30 77.84 Poverty score 49.42 17.40 49.49 16.68 48.22 16.41 47.00 15.78 47.18 14.98 49.54 15.63 49.26 15.62 47.28 15.32 48.11 16.16 Number of food groups produced 3.30 2.20 3.14 2.10 3.33 2.19 3.34 2.33 3.63 2.24 3.50 2.36 3.49 2.38 3.39 2.09 3.26 2.10 Household food insecurity access score (HFIAS) 1.67 3.24 1.68 3.19 2.00 3.77 2.27 4.20 2.42 4.12 1.81 3.65 1.70 3.14 1.74 3.70 2.31 4.03 Religion (Muslim = 1) 0.96 0.21 0.92 0.27 0.94 0.24 0.94 0.23 0.97 0.18 0.96 0.21 0.98 0.14 0.98 0.15 0.95 0.21 Male 0.13 0.33 0.20 0.40 0.20 0.40 0.44 0.50 0.43 0.50 0.39 0.49 0.24 0.43 0.26 0.44 0.24 0.43 Female 0.87 0.33 0.80 0.40 0.80 0.40 0.56 0.50 0.57 0.50 0.61 0.49 0.76 0.43 0.74 0.44 0.76 0.43 Age 38.51 13.50 34.68 10.81 37.39 11.39 38.67 12.14 40.86 13.54 37.69 11.86 38.00 12.95 38.85 13.04 36.34 11.44 Highest level of education None 0.30 0.46 0.33 0.47 0.28 0.45 0.40 0.49 0.38 0.49 0.37 0.48 0.32 0.47 0.43 0.50 0.41 0.49 Primary school 0.37 0.48 0.30 0.46 0.39 0.49 0.29 0.45 0.34 0.48 0.32 0.47 0.40 0.49 0.28 0.45 0.32 0.47 Junior secondary school 0.13 0.33 0.13 0.34 0.16 0.37 0.12 0.32 0.12 0.32 0.08 0.27 0.10 0.31 0.08 0.28 0.07 0.25 Secondary school 0.05 0.23 0.05 0.22 0.04 0.20 0.05 0.21 0.06 0.23 0.09 0.29 0.06 0.25 0.06 0.24 0.08 0.27 SSC pass 0.04 0.21 0.06 0.24 0.07 0.25 0.03 0.18 0.06 0.24 0.07 0.25 0.05 0.22 0.06 0.24 0.06 0.24 Postsecondary education 0.09 0.28 0.11 0.31 0.05 0.22 0.09 0.29 0.03 0.16 0.07 0.25 0.04 0.20 0.07 0.25 0.05 0.22 Spouse of household head 0.63 0.48 0.60 0.49 0.63 0.48 0.39 0.49 0.38 0.49 0.43 0.50 0.51 0.50 0.55 0.50 0.59 0.49 Other relationship to household head 0.13 0.33 0.15 0.36 0.12 0.32 0.11 0.31 0.10 0.30 0.07 0.26 0.16 0.37 0.18 0.38 0.12 0.32 Hunger at time of meal event 3.80 1.10 3.63 1.25 3.67 1.23 3.67 1.14 3.74 1.15 3.74 1.14 3.65 1.18 3.68 1.20 3.61 1.19 Baseline nutrition knowledge score 18.83 5.13 19.06 5.64 18.79 4.78 18.20 5.35 18.48 5.62 19.26 4.60 18.88 5.37 18.86 4.96 18.83 4.92

PAGE 129

129 Table A 4. Summary statistics by treatment groups defined in T able 22 Control N NG BPP BPP x N BPP x NG Mean Std.Dev. Mean Std.Dev. Mean Std.Dev. Mean Std.Dev. Mean Std.Dev. Mean Std.Dev.

PAGE 130

130 Table A 5. T test on mean m eal diversity score (MDS) by meal for BPP treatment groups Mean MDS Treatment (Meal 1, Meal 2) Meal 1 Meal 2 Difference T stat P value Treatment A (Regular, Regular) 0.77 0.77 0.00 0.29 0.772 Treatment B (Regular, BPP) 0.78 0.78 0.00 0.06 0.955 Treatment C (BPP, Regular) 0.77 0.78 0.01 2.92 0.004

PAGE 131

131 Table A 6. T test for order effects: mean difference in MDSBPP and MDSRegular by order of exposure to the BPP Treatment B (BPP in Meal 2) Treatment C (BPP in Meal 1) Difference (B vs. C) T stat P value P value lower P value upper Mean difference (MDSBPP MDSRegular) 0.0001 0.0144 0.0145 1.375 0.179 0.0896 0.9104 Std. Error 0.0073 0.0076 0.0106 Clusters 17 16 Intra cluster correlation = 0.061

PAGE 132

132 Table A 7. Coefficient estimates from Poisson regression of treatments on individual dietary diversity score (IDDS) (1) (2) VARIABLES IDDS IDDS BPP only 0.015 0.006 [0.073] [0.032] Nutrition education 0.018 0.004 [0.044] [0.039] Nutrition education with gender 0.009 0.029 [0.030] [0.026] BPP x nutrition education 0.005 0.002 [0.070] [0.031] BPP x nutrition education with gender 0.025 0.020 [0.076] [0.035] Post 0.089 0.094* [0.056] [0.054] BPP x post 0.011 0.006 [0.059] [0.058] Nutrition education x post 0.022 0.029 [0.064] [0.064] Nutrition education with gender x post 0.032 0.037 [0.061] [0.060] BPP x nutrition education x post 0.020 0.024 [0.061] [0.060] BPP x nutrition education with gender x post 0.001 0.006 [0.063] [0.061] Mymensingh district 0.407*** [0.018] Gender (female = 1) 0.018 [0.026] Age 0.001 [0.001] Primary school (highest level) 0.032** [0.016] Junior secondary school (highest level) 0.056** [0.026] Secondary school (highest level) 0.083*** [0.027] SSC pass (highest level) 0.068** [0.031] Postsecondary education 0.066* [0.034] Responsible for food preparation 0.037* [0.020]

PAGE 133

133 Table A 7. Continued. (1) (2) VARIABLES IDDS IDDS Total cultivable land (decimals) 1.20E 04*** [2.53E 05] Farm diversity (count of food groups produced) 0.017*** [0.005] Constant 1.981*** 1.608*** [0.056] [0.065] Number of observations 2,164 2,164 Log Likelihood 5153 4704 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 134

134 Table A 8. Results from logit regressions on food group production (odds ratios) VARIABLES Cereals Vitamin A rich veg White roots/tubers Leafy green veg Other vegetables Vitamin A rich fruit Fruit Meat Eggs Fish Pulses Dairy Oilseeds Mymensingh 3.420*** 0.496 29.558*** 5.175*** 3.979*** 11.516*** 14.223*** 2.781*** 1.706*** 2.018** 0.006*** 0.872 0.036*** [1.241] [0.228] [22.289] [2.005] [1.059] [2.836] [3.451] [0.633] [0.303] [0.683] [0.006] [0.198] [0.021] Muslim 0.206 0.350** 1.337 1.035 0.454* 0.757 0.722 0.973 1.714** 0.327*** 0.79 1.084 1.129 [0.221] [0.156] [0.586] [0.549] [0.188] [0.515] [0.275] [0.527] [0.365] [0.108] [0.397] [0.182] [0.668] Household food insecurity access score (HFIAS) 0.946* 0.938** 0.974 0.913*** 0.976 0.96 0.944** 0.986 0.963** 0.909*** 0.910*** 1.017 0.913 [0.029] [0.029] [0.054] [0.026] [0.024] [0.035] [0.024] [0.031] [0.017] [0.033] [0.030] [0.031] [0.076] Male headed household 3.046 0.588 2.583 0.765 0.702** 0.886 0.824 0.722 0.753* 1.329 0.864 0.985 1.538 [2.415] [0.269] [2.042] [0.244] [0.114] [0.221] [0.182] [0.160] [0.124] [0.401] [0.245] [0.180] [0.701] Household head primary education 1.298 0.575** 0.615* 0.994 1.082 1.131 0.685** 1.392* 1.063 1.409* 0.849 0.983 0.679 [0.638] [0.151] [0.171] [0.259] [0.270] [0.217] [0.130] [0.267] [0.155] [0.277] [0.205] [0.175] [0.274] Household head junior secondary education 1.696 0.93 0.467 1.078 1.742** 1.973*** 0.679 1.392* 1.003 2.069*** 0.933 0.905 0.697 [0.976] [0.372] [0.217] [0.260] [0.472] [0.434] [0.212] [0.243] [0.193] [0.534] [0.389] [0.274] [0.394] Household head secondary education 0.576 0.875 0.303*** 0.965 1.706 1.817 2.105** 0.964 0.782 1.618* 0.644 1.277 0.559 [0.440] [0.485] [0.138] [0.297] [0.564] [0.840] [0.800] [0.297] [0.231] [0.437] [0.559] [0.365] [0.441] Household head SSC pass 0.788 1.352 1.05 0.782 2.221** 2.441** 1.072 1.507 1.65 1.932* 0.773 1.024 0.384 [0.668] [0.632] [0.526] [0.248] [0.728] [0.866] [0.520] [0.506] [0.551] [0.662] [0.698] [0.298] [0.393] Household head postsecondary education 0.318* 1.09 0.715 0.993 0.985 1.874 2.641* 0.507* 0.991 2.576*** 0.304 0.962 2.233 [0.216] [0.703] [0.312] [0.344] [0.372] [0.779] [1.347] [0.184] [0.272] [0.877] [0.352] [0.361] [1.598] Age of household head 0.993 0.998 1.014* 0.995 1.013** 1.000 1.007 1.006 0.995 1.014** 1.012*** 1.009 0.998 [0.009] [0.009] [0.008] [0.007] [0.007] [0.008] [0.007] [0.006] [0.006] [0.006] [0.005] [0.006] [0.009] Household size 1.023 0.921 0.99 0.945 1.051 1.051 1.080* 1.051 1.133*** 1.085* 1.081 1.113* 1.152 [0.081] [0.065] [0.074] [0.051] [0.049] [0.059] [0.045] [0.041] [0.046] [0.052] [0.079] [0.062] [0.135] Poverty score 0.986* 1.005 0.998 0.998 0.998 1.014* 1.004 0.999 0.998 1.017*** 0.992 0.998 1 .000 [0.008] [0.011] [0.008] [0.007] [0.006] [0.007] [0.007] [0.006] [0.006] [0.006] [0.009] [0.006] [0.011] Total cultivable land (decimals) 1.11 1.002** 1.005*** 1.003*** 1.006*** 1.001** 1.002 1.001 1.000 1.000 1.012*** 1.004*** 1.005*** [0.085] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] [0.004] [0.001] [0.001] Buy food at market 1.101 0.67 0.817 1.237 1.019 1.325 1.267 1.111 0.969 0.888 1.071 1.168 0.472*** [0.406] [0.200] [0.290] [0.358] [0.187] [0.293] [0.327] [0.194] [0.143] [0.168] [0.263] [0.211] [0.131] Sell food at market 3.732*** 2.940*** 1.626 1.558* 2.287*** 1.167 1.773*** 1.820*** 1.493** 1.352* 3.163*** 2.027*** 3.950*** [1.529] [0.720] [0.524] [0.409] [0.421] [0.160] [0.302] [0.303] [0.239] [0.219] [0.689] [0.290] [0.988] Constant 0.133 0.391 0.001*** 0.098*** 0.120*** 0.063*** 0.176** 0.126*** 0.505* 0.057*** 0.254* 0.080*** 0.050*** [0.308] [0.325] [0.001] [0.069] [0.065] [0.053] [0.128] [0.083] [0.189] [0.029] [0.190] [0.033] [0.057] Observations 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 1,130 Pseudo R squared 0.675 0.0713 0.251 0.126 0.216 0.314 0.365 0.0792 0.0553 0.152 0.403 0.0739 0.223 Log Likelihood 241.6 270.7 261.9 445.9 575 528.1 491.8 641.7 727.5 561.2 327.7 627.9 228.5 Robust se eform in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 135

135 Table A 9. Results from logit regressions on food group consumption (odds ratios)

PAGE 136

136 Table A 10. Coefficient estimates from Poisson regression on farm diversity score VARIABLES All Borguna Mymensingh District (Mymensingh = 1) 0.465*** [0.094] Religion (Muslim = 1) 0.137 0.234*** 0.153** [0.090] [0.080] [0.072] Household Food Insecurity Access Score 0.030*** 0.020 0.046*** [0.008] [0.013] [0.007] Male headed household 0.064 0.072 0.031 [0.046] [0.056] [0.073] Household head primary education 0.013 0.033 0.021 [0.051] [0.099] [0.046] Household head junior secondary education 0.072 0.044 0.085** [0.058] [0.192] [0.043] Household head secondary education 0.033 0.156 0.039 [0.054] [0.250] [0.056] Household head SSC pass 0.125* 0.523*** 0.168*** [0.064] [0.076] [0.060] Post secondary education 0.056 0.582** 0.059 [0.082] [0.234] [0.082] Age of household head 0.002* 0.001 0.003** [0.001] [0.003] [0.001] Household size 0.025** 0.047 0.022** [0.010] [0.031] [0.009] Poverty score 0.001 0.003 1.57E 04 [0.001] [0.004] [0.001] Total cultivable land (decimals) 0.001*** 0.001*** 0.001*** [1.52E 04] [3.16E 04] [1.35E 04] Buy food at market 0.055 0.041 0.078 [0.047] [0.109] [0.051] Sell food at market 0.326*** 0.475*** 0.260*** [0.039] [0.065] [0.047] Constant 0.561*** 0.498 0.805*** [0.142] [0.341] [0.108] Observations 1,132 537 595 Log Likelihood 2150 947.6 1186 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 137

137 Table A 11. Coefficient estimates from Poisson regression on household dietary diversity score VARIABLES All Borguna Mymensingh District (Mymensingh = 1) 0.326*** [0.023] Religion (Muslim = 1) 0.046 0.036 0.051 [0.030] [0.042] [0.037] Household Food Insecurity Access Score 0.013*** 0.014*** 0.010 [0.003] [0.003] [0.010] Male headed household 0.005 0.050 0.062** [0.031] [0.035] [0.029] Household head primary education 0.012 0.062*** 0.043* [0.020] [0.022] [0.025] Household head junior secondary education 0.027 0.007 0.054 [0.029] [0.065] [0.037] Household head secondary education 0.008 0.081 0.015 [0.027] [0.110] [0.030] Household head SSC pass 0.002 0.159* 0.019 [0.033] [0.083] [0.038] Post secondary education 0.035* 0.110** 0.061** [0.020] [0.050] [0.024] Age of household head 2.38E 05 3.95E 04 8.34E 04 [0.001] [0.001] [0.001] Household size 0.002 0.013 3.52E 04 [0.005] [0.012] [0.006] Poverty score 3.98E 04 0.002*** 0.001 [0.001] [0.001] [0.001] Total cultivable land (decimals) 7.67E 05 3.12E 04** 2.92E 04*** [1.10E 04] [1.66E 04] [8.07E 05] Farm diversity (count of food groups produced) 0.038*** 0.062*** 0.026*** [0.006] [0.009] [0.006] Buy food at market 0.030 0.012 0.050 [0.019] [0.027] [0.032] Sell food at market 0.065** 0.070* 0.046 [0.030] [0.038] [0.043] Food preparer 0.043** 0.057 0.054*** [0.019] [0.038] [0.020] Constant 1.614*** 1.806*** 1.944*** [0.060] [0.111] [0.058] Observations 1,130 535 595 Log Likelihood 2408 1084 1313 Robust standard errors in brackets *** p<0.01, ** p<0.05, p<0.1

PAGE 138

138 APPENDIX B BASELINE QUESTIONNAIRE Q1 Hello, my name is _________________ and I am conducting research on behalf of the Bangladesh Agricultural University and the University of Florida. We are conducting research about household f ood availability and access. This research is for a doctoral dissertation at the University of Florida in the United States of America. First of all, I would like to thank you for taking the time to meet with me. We would like to ask you some questions about yourself, your household, and your food consumption and diet. This interview should take approximately 30 minutes to 1 hour, and again I thank you for your time. Your participation in this study will be confidential to the extent provided by law, an d your responses and comments will be anonymous. There are no direct benefits, risks, or compensation to you for participating in this study. Participation in this study is completely voluntary, and you may withdraw your consent to participate at any time without penalty. In addition to the interview, we invite you to attend a workshop and lunch we have scheduled for ________ (date on meal ticket). To participate in this research, you must attend the extension workshop. You will be given a meal as compensa tion for your time. Q2 Are you available and willing to participate in the workshop? Yes (1) No (2) Q3 Thank you for your participation. I will be asking questions about your household members and yourself, your household income, agricultural productio n, food availability and access. Your responses will be completely private and will only be connected with your name through an identification number. The list connecting your name with the identification number will be destroyed at the end of the study, s o after that, your responses will be completely anonymous. We understand that you may not have all of the precise information available. It is important for the information in this study to be as accurate as possible. Please try to answer as many questions as you can, but if you cannot remember some information, it is ok to answer I do not remember. Finally, you have the right to not answer any question we might ask. You may also choose to stop answering questions at any time. If you do, we will not collect any more information from you. However, we would keep and use the information we had already collected from you. Q4 Given this information, are you willing to participate in this survey and the workshop? Yes (1) No (2)

PAGE 139

139 Q5 Enumerator Name (Last name, First name) Afroza (2) Alamin (17) Ame (4) Anik (6) Anup (7) Ayesha (8) Azil (9) Diti (37) Eva (10) Himu (11) Jenny (12) Mahmud (14) Mumu (15) Nilima (39) Nosib (3) Pinkey (18) Prodip (19) Rabbi (20) Rubyeat (21) Sadia (Urvi) (24) Sajib (22) Shakir (25) Shaon (26) Shatabdi (36) Shirin (38) Shoaib (31) Shuvo (27) Tapoti (35) Tania (29) Wahidul (28) Q6 District Mymensingh (1) Borguna (2) Q7 Upazila Mymensingh Sadar (1) Amtoli (2) Q8 Union Chowra (1) Ho ldia (2) Answer If District Borguna Is Selected Q9 Sushilan Village Baitakata (1) Baitmore (2) Chalavanga (3) Chandra (4) Chandra Kapali (5) Ghatkhali (6) Gilatali (7) Holdia (8) Holodia (Paka) (9) Holodia (10) Kalibari (11) Kawnia (12) Kawnia Kapali (13) Loda (14) Patakata (15) Patakata midle (16) Uttar Tokta Bunia (17)

PAGE 140

140 Answer If District Mymensingh Is Selected Q316 BAU Village Boira moddopara sorkar bari (1) Boira mosque (2) Borobilarpar (3) Bot tola bazar (4) Bot tola hatem bapari (5) Chor nilokkhai ujanpara (6) Chorkalibari shomvuganj (7) Chornillokhia digolpara (8) Goneshampur (9) Gosta north (10) Jhaugau (11) Jogir ali chor nilokkhia (12) Kismat khagdohor (13) Maijbari (14) Mirjapur east (15) Mirjapur north (16) Mirj apur south (17) Muktijodda bazar chor (18) Mukhtijoddha bazar jogir algi (19) Pagla Bazaar (20) Pagla bazar kazi bari (21) Pagla bazar mojiborer bari (22) Ragobpur (23) Sathiapara (24) Sathiapra chor nillokkhia (25) Shailmari purrush (26) Suhila moddopara (27) Suhila nodir par (28) Suhila north (29) Sutiakhali middle (30) Sutiakhali palpara 1 (31) Sutiakhali palpara 2 (32) Vabokhali pwest (33) Vabukhali (34) Q317 BHH Code (Sushilan) or S.I. Number (BAUEC) Q10 Thank you for participating in this study. I would like to start by asking some information about you. Q11 Respondent Name First name (1) Last name (2) Q12 Do you have a mobile number where we can reach you? If so, what is your number? Q13 Gender Male (1) Female (2) Q14 Are you the head of household? Yes (1) No (2) Answer If Are you the head of household? No Is Selected Q15 What is your relationship to the head of household?

PAGE 141

141 Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law/Son in law (8) Brother/Sister (9) Father in law/Mother in law (10) Nephew/Niece (11) Grandfather/Grandmother (12) Grandson/Granddaughter (13) Sister in law/Brother in law (14) Brother's wife (15) Other (e.g. serva nt) (16) Q16 What is your marital status? Married (1) Single (never married) (2) Widowed (3) Divorced (5) Separated (4) Q17 What is your age?

PAGE 142

142 Q18 What is the highest level of education you have completed? None (1) Class 1 (2) Class 2 (3) Class 3 (4) Class 4 (5) Class 5 (6) Class 6 (7) Class 7 (8) Class 8 (9) Class 9 (10) Class 10 (17) SSC Pass (11) HSC Pass (12) Graduate (13) Post graduate (14) Medical (15) Vocational/Technical education (16) Q19 What i s your primary occupation? Farming (own land) (1) Poultry and livestock rearing (2) Sharecropper (3) Agricultural day labor/contract labor (4) Fishing (own boat) (5) Fishing labor (someone else's boat) (6) Fish farming (7) Boat operation (8) Rickshaw/van operator (9) Casual labor (10) Self employed in business/petty business Non agricultural day labor/contract labor (12) Regular salaried employment (13) Paid "volunteer" (14) Housework (child care/home care) (15) Servant/maid (16) Student (17) Beggar (18) Unemployed (19) Old/Disabled (20) Other (21) ____________________ N/A (22) Answer If Primary occupation Other Is Selected Q20 Specify your primary occupation.

PAGE 143

143 Q21 What is your secondary occupation? Farming (own land) (1) Poultry and livestock rearing (2) Sharecropper (3) Agricultural day labor/contract labor (4) Fishing (own boat) (5) Fishing labor (someone else's boat) (6) Fish farming (7) Boat operation (8) Rickshaw/van operator (9) Casual labor (10) Self empl oyed in business/petty business (11) Non agricultural day labor/contract labor (12) Regular salaried employment (13) Paid "volunteer" (14) Housework (child care/home care) (15) Servant/maid (16) Student (17) Beggar (18) Unemployed (19) Old/Disabled (20) O ther (21) ____________________ N/A (22) Answer If Secondary occupation Other Is Selected Q22 Specify your secondary occupation.

PAGE 144

144 Q23 Now I would like to ask some questions about your household members.How many individuals, NOT including yourself, live in your household? Please include the people who habitually eat and sleep in the home, including those who have been absent less than six months and have not established another residence. This includes maids or servants who live in the household. 0 (21) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) 8 (8) 9 (9) 10 (10) 11 (11) 12 (12) 13 (13) 14 (14) 15 (15) 16 (16) 17 (17) 18 (18) 19 (19) 20 (20) Q24 Household member name First name (1) Q25 Gender Male (1) Female (2) Q26 What is ${q://QID26/ChoiceTextEntryValue/1}'s age? If under 1 year, enter 0. Q27 Marital status Married (1) Single (never married) (2) Widowed (3) Divorced (4) Separated (5)

PAGE 145

145 Q28 Relationship to head of household Household head ( 1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law/Son in law (8) Brother/Sister (9) Father in law/Mother in law (10) Nephew/Niece (11) Grandfather/Grandmother (12) Grandson/Granddaughter (13) Sister in law/Brother in law (14) Brother's wife (15) Other (e.g. servant) (16) ____________________ Answer If Relationship to head of household (Any Loop) Other (e.g. servant) Is Selected Q29 Specify his/ her relationship to the head of household. Answer If Gender Female Is Selected And What is ${q://QID26/ChoiceTextEntryValue/2}'s age? If under 1 year, enter 0. Text Response Is Greater Than or Equal to 15 Q30 Is ${q://QID26/ChoiceTextEntryValue/1} currently pregnant? Yes ( 1) No (2) Unsure (3) Answer If Gender Female Is Selected And What is ${q://QID26/ChoiceTextEntryValue/2}'s age? If under 1 year, enter 0. Text Response Is Greater Than or Equal to 15 Q31 Is ${q://QID26/ChoiceTextEntryValue/1} currently breastfeeding? Yes (1) No (2) Unsure (3) Answer If Is ${q://QID26/ChoiceTextEntryValue/2} currently breastfeeding? Yes Is Selected Q32 Please tell me the age (months) of each child she is breastfeeding. First child (1) Second child (2) Third child (3)

PAGE 146

146 Q33 What is the highest level of education ${q://QID26/ChoiceTextEntryValue/1} completed? None (1) Class 1 (2) Class 2 (3) Class 3 (4) Class 4 (5) Class 5 (6) Class 6 (7) Class 7 (8) Class 8 (9) Class 9 (10) Class 10 (17) SSC Pass (11) HSC Pass (12) Graduate (13) Post graduate (14) Medical (15) Vocational/Technical education (16) Answer If What is ${q://QID26/ChoiceTextEntryValue/2}'s age? Text Response Is Greater Than or Equal to 10 Q34 How often does ${q://QID26/ChoiceTextEntryVal ue/1} go to the market to purchase food for the household? Always (11) Most of the time (12) About half the time (13) Sometimes (14) Never (15) Q35 Thank you for the information about your household members. Now I have a few more questions about you. An swer If Gender Female Is Selected Q36 Are you currently pregnant? Yes (1) No (2) Unsure (3) Answer If Gender Female Is Selected Q37 Are you currently breastfeeding? Yes (1) No (2) Answer If Are you currently breastfeeding? Yes Is Selected Q38 Please tell me the age (months) of each child you are breastfeeding. First child (1) Second child (2) Third child (3)

PAGE 147

147 Q39 What religion are you? Islam (1) Hindu (2) Buddhist (3) Christian (4) Other (5) ____________________ N/A (6) Answer If What religion are you? O ther Is Selected Q40 Specify religion. Q41 In the last year, did anyone in your household ever do any work for which he or she was paid on a daily basis? Yes (1) No (2) Q42 How many months were you employed in the last year? Q43 What was your monthly income (BDT)? Q44 What is your monthly household income (BDT)? Include income from all household members. Q45 What is your annual household income (BDT)? Include income from all household members. Q46 Are you the person who usually goes to the market t o purchase food for the household? Always (11) Most of the time (12) About half the time (13) Sometimes (14) Never (15) Q47 Next I will ask some questions about your household assets and agriculture production. Q48 What type of latrine does the household use? Open field (1) Kacha latrine (temporary or permanent) (2) Pacca (pit or water seal) (3) Sanitary (4) Q49 How many rooms does the household occupy (excluding rooms used for business)?

PAGE 148

148 Q50 What is the main construction material of the walls of the m ain room in the home? Tile/wood (1) Hemp/hay/bamboo (2) C.I. sheet (3) Cement (4) Q51 Does the household own a television? Yes (1) No (2) Q52 How many fans does the household own? Q53 How many mobile phones does the household own? Q54 Does the household own any bicycles, motorcycles/scooters, or motor cars, etc.? Yes (1) No (2) Q55 How much land does the household own? Specify for each land use. Cultivable land (currently cropped or fallow) (1) Dwelling house/homestead (2) Non cultivated land (not including homestead) (3) Land Owned (Decimals) (1) Q56 How much additional cultivable land is rented/sharecropped/mortgaged in or out by the household? (excluding uncultivable land and dwellinghouse/homestead land) Rented/sharecropped/morgaged IN (2) Rented/sharecropped/mortgaged OUT (4) Additional Cultivable Land (Decimals) (1) Q57 How much total land does the household have access to? (Decimals)Note: Verify that this equals the sum of the total land owned plus additional cultivable land.

PAGE 149

149 Q58 Did your household produce any cereal crops in the last year? (Mark all that apply) Rice (Aus, Aman, and/or Boro) (1) Maize (2) Wheat (3) Other (4) ____________________ Other (5) ____________________ No cereal (6) Q59 Count the number of cereal crops produced in the last year and confirm with the respondent. 0 (1) 1 (6) 2 (7) 3 (8) 4 (9) 5 (10) Q60 Type of cereal Q61 What was the total area planted for ${q://QID155/ChoiceTextEntryValue/1} in the last year? (Decima ls) Q62 How much ${q://QID155/ChoiceTextEntryValue/1} did you produce in the last year? Q63 Units of production Kilograms (2) Other (3) ____________________ Q64 How much ${q://QID155/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1)

PAGE 150

150 Q65 Did your household produce any vegetables in the last year? (Mark all that apply) Pumpkin (1) Bitter gourd (2) Bottle gourd (3) Eggplant (21) White potato (4) Sweet potato (orange) (5) Tomato (6) Cucumber (7) Red amaranth (8) Amaranth (9) Spinach (10) Cauliflower (11) Cabbage (12) Okra (13) Chili p epper (14) Onion (15) Turnips (16) Taro (17) Other (18) ____________________ Other (19) ____________________ No vegetables (20) Q66 Count the number of vegetable crops produced in the last year and confirm with the respondent 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) 10 (11) 11 (12) 12 (13) 13 (14) 14 (15) 15 (16) 16 (17) 17 (18) 18 (19) 19 (20) Q67 Type of vegetable Q68 Is this vegetable mixed crop with other vegetables? Yes (1) No (2) Answer If Is this vegetable mixed crop with other vegetables? Yes Is Selected Q69 In your mixed crop plot of land, is ${q://QID163/ChoiceTextEntryValue} more or less than half of the area planted? More than half (1) About half (2) Less than half (3)

PAGE 151

151 Q70 What was the total area planted for ${q://QID163/ChoiceTextEntryValue/2} in the last year? (Decimals)Note: If the vegetable is mixed crop, enter the approximate number of decimals allocated to this crop based on the previous question and the total land used for vegetab les. (more than half, about half, less than half). Q71 How much ${q://QID163/ChoiceTextEntryValue/2} did you produce in the last year? Q72 Units of production Kilograms (2) Pieces (3) Q73 How much ${q://QID163/ChoiceTextEntryValue/2} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q74 Did your household produce any pulses in the last year? (Mark all that apply) Lentils (1) Mung bean (2) Khesari (3) Other pulses (4) ____________________ Other pulses (5) ____________________ No pulses (6) Q75 Count the number of pulses produced in the last year and confirm with the re spondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) Q76 Type of pulse Q77 What was the total area planted for ${q://QID221/ChoiceTextEntryValue/2} in the last year? (Decimals) Q78 How much ${q://QID221/ChoiceTextEntryValue/2} did you produce in the last year ?

PAGE 152

152 Q79 Units of production Kilograms (2) Pieces (3) Q80 How much ${q://QID221/ChoiceTextEntryValue/2} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q81 Did your household produce any fruit in the last year? (Mark all that apply) Mango (1) Jujube (2) Jackfruit (3) Litchi (4) Guava (5) Papaya (6) Coconut (7) Banana (8) Other (9) ____________________ Other (10) ____________________ Other (11) ____________________ No fruits (12) Q82 Count the number of fruits produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) 10 (11) 11 (12) Q83 Type of fruit

PAGE 153

153 Q84 How many ${q://QID231/ChoiceTextEntryValue} trees produced fruit in the last year? Q85 How much ${q://QID231/ChoiceTextEntryValue} did you produce in the last year? Q86 Units of production Kilograms (2) Pie ces (3) Q87 How much ${q://QID231/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q88 Did your household produce any oilseed crops in the last year? (Mark all that apply) Mustard (1) Peanut (2) Soybean (3) Sesame (4) Other (5) ____________________ Other (6) ____________________ Other (7) ____________________ No oilseeds (8) Q89 Count the number of oilseed crops produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) Q90 Type of oilseed Q91 What was the total area planted for ${q://QID239/ChoiceTextEntryValue} in the last year? (Decimals)

PAGE 154

154 Q92 How much ${q://QID239/ChoiceTextEntryValue} did you produce in the last year? Q93 Units of production Kilograms (2) Q94 How much ${q://QID239/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q95 Did your household produce any fibrous crops in the last year? (Mark all that apply) Jute (1) Kenaf (8) Mesta (2) Other (5) ____________________ Other (6) ____________________ Other (7) ____________________ No fibrous crops (13) Q96 Count the number of fibrous crops produced in the last year and confir m with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) Q97 Type of fibrous crop Q98 What was the total area planted for ${q://QID247/ChoiceTextEntryValue} in the last year? (Decimals) Q99 How much ${q://QID247/ChoiceTextEntryValue} did you produce in the last year? Q100 Units of production Kilograms (2)

PAGE 155

155 Q101 How much ${q://QID247/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q102 Did your household raise any livestock for production in the last year? If yes, what animal products did you produce? (Mark all that apply) Cow milk (1) Cow meat (beef) (13) Chicken meat (5) Chicken eggs (14) Duck meat (15) Duck eggs (6) Goat milk (7) Goat meat (8) Pigeon meat (9) Other (10) ____________________ Other (4) ____________________ Other (11) ____________________ No livestock (12) Q103 Count the n umber of animal products the household produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) Q104 Type of animal product Note: Be specific about animal type. If milk, specify cow or goat. If eggs, specify chicken or duck. Animal product (1)

PAGE 156

156 Q105 How many animals did you have for producing ${q://QID169/ChoiceTextEntryValue/1} in the last year? Q106 How much ${q://QID169/ChoiceTextEntryValue/1} did you produce in the last year? Q107 Unit s of production Kilograms (1) Liters (2) Pieces (3) Q108 How much ${q://QID169/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Sold (1) Kept for home consumption (2) Product (1) Q109 Did your household cultivate any fish last year? (Mark all that apply) Rui (23) Katla (25) Carp (26) Tilapia (27) Koi (28) Mola (29) Shrimp (30) Prawn (31) Other (32) ____________________ Other (33) ____________________ Other (34) ____________________ No fish (24) Q110 Count the number of fish species cultivated in the last year and confirm with the respondent. Q111 Type of fish Species (1) Q112 How much ${q://QID187/ChoiceTextEntryValue/1} d id you produce in the last year? Q113 Units of production Kilograms (2) Pieces (3)

PAGE 157

157 Q114 How much ${q://QID187/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given in the previous question about total production. Sold (1) Kept for home consumption (2) Product (1) Q115 Are you the person who normally prepares food for the household? Yes (1) No (2) Answer If Are you the person who normally prepares food for the household? No Is Selected Q116 In your household, who normally prepares food? Note: if the respondent answers the name of the person, clarify that person's relationship to the head of household. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Br other's wife (16) Other (e.g. servant) (17) Answer If In your household, who normally prepares food?;Note: if the respondent answers the name of the person, clarify that person's relationship to the head of household. Other (e.g. servant) Is Selected Q117 Specify other who prepares the food for the household. Q118 Please describe the foods (meals and other snacks) that you ate or drank yesterday during the day and night, whether at home or outside the home. Start with the first food or drink in the morning. Breakfast (1) Snack (2) Lunch (3) Snack (4) Dinner (5) Q119 When the respondent recall of meals consumed in the last 24 hours is complete, fill in the food groups consumed by the respondent based on the information recorded above. Remembe r to ask whether that meal was consumed inside the home or outside of the home. For any food groups not mentioned in the meals and snacks, prompt the respondent by asking if he or she consumed a food item from this group. If the respondent has not consumed a food group, ask if anyone in the household consumed food items from that food group in the last 24 hours.

PAGE 158

158 Has anyone in the household consumed? Did the respondent consume? Yes inside the home (1) Yes outside the home (2) Inside and outside the home (3) Unsure (4) No (5) Yes inside the home (1) Yes outside the home (2) Inside and outside the home (3) No (4) Cereals (rice, bread, wheat, rice flakes, puffed rice, barley, biscuits, popcorn) White roots and tubers (white potatoes, turnips) Vitamin A rich vegetables and tubers (pumpkin, carrots, squash, orange flesh sweet potatoes) Dark leafy green vegetables (amaranth, spinach, taro leaf, pumpkin leaf, jute leaf) Other vegetables (cucumber, radish, pepper,

PAGE 159

159 cabbage, beans, cauliflower, onion) Vitamin A rich fruits (ripe papaya, ripe mangos, orange, grapefruit) Other fruits (banana, apple, jackfruit, watermelon, guava, plums, pineapple, melons, lemons, lychees) Meat (chicken, beef, poultry, lamb, liver, kidney) Eggs Fish and seafood Legumes, nuts, and seeds (lentils, dal, black gram, kheshari, mung beans) Milk and milk products (cow milk, goat milk,

PAGE 160

160 powdered milk, yogurt, curd, cheese) Oils and fats (butter, ghee, mustard oil, soybean oil) Sweets (sugar, honey, molasses, chocolate, ice cream, soda) Spices, condiments, beverages (black pepper, salt, fish powder, chili sauce, cumin, chili ) Q120 Now I would like to ask you about all of the different foods your household members have eaten in the last 7 days. Could you please tell me how many days in the past week your household has eaten the following foods? This can be inside the home or outside the home. (Enter a number 0 7) Please note your primary and secondary sour ce for each food item.

PAGE 161

161 Number of DAYS eaten in the past week. (0 7) Primary Source Secondary Source Cereals (rice, bread, wheat, rice flakes, puffed rice, barley, biscuits, popcorn) (1) White roots and tubers (white potatoes, turnips) (2) Vitamin A rich vegetables and tubers (pumpkin, carrots, squash, orange flesh sweet potatoes) (3) Dark leafy green vegetables (amaranth, spinach, taro leaf, pumpkin leaf, jute leaf) (4) Other vegetables (cucumber, radish, pepper, cabbage, beans, cau liflower, onion) (5) Vitamin A rich fruits (ripe papaya, ripe mango, orange, grapefruit) (6) Other fruits (banana, apple, jackfruit, watermelon, guava, plums, pineapple, melons, lemons, lychees) (7) Meat (chicken, beef, poultry, lamb, liver, kidney) (8) Eggs (9) Fish and seafood (10) Legumes, nuts, and seeds (lentils, dal, black gram, kheshari, mung beans) (11) Milk and milk products

PAGE 162

162 (cow milk, goat milk, powdered milk, yogurt, curd, cheese) (12) Oils and fats (butter, ghee, mustard oil, soybean oil) (13) Sweets (sugar, honey, molasses, chocolate, ice cream, soda) (14) Spices, condiments, beverages (black pepper, salt, fish powder, soy sauce, chili sauce, cumin, chili powder, cinnamon, coffee, tea) (15) Q122 I am going to ask you some questions about nutrition, vitamins and food. This is a survey, not a test. Your answers will help identify which dietary advice people find confusing. Please let me know if you need me to clarify any of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q123 Do you think that health experts recommend that people should be eating more, the same amount, or less of these foods? (Mark one box per food) More (1) Same (2) Less (3) I don't know (4) Vegetables (1) Sweets (2) Meat (3) Carbohydrates (4) Fatty foods (5) High fibre foods (6) Fruit (7) Salty foods (8) Q124 How many servings of fruits and vegetables per day do you think experts are advising people to eat? (One serving could be, for example, an apple or a spoonful of cucumbers). The response should be in number of times per day/servings, not grams.

PAGE 163

163 Q125 Experts classify food into groups. We are intereste d to see whether people are aware of what foods are in these groups. Q126 Do you think the following foods are high or low in vitamin A? High (1) Low (2) Unsure (3) Carrot (1) Banana (27) Rice (28) Wheat flour (47) Mola fish (49) Leafy greens (spinach, amaranth) (50) Q127 Do you think experts put these foods in the carbohydrates group? Yes (1) No (2) Unsure (3) Rice (1) Potatoes (3) Sweet potatoes (4) Chapati (5) Biscuits (6) Q128 Do you think these foods are high or low in protein? High (1) Low (2) Unsure (3) Chicken (1) Fish (2) Egg (3) Pulses (4) Milk (5) Fruit (6) Rice (7)

PAGE 164

164 Q129 Are you aware of any major health problems or diseases that are related to a low intake of fruits and vegetables? Yes (1) No (2) Unsure (3) Answer If Are you aware of any major health problems or diseases that are related to a low intake of fruits and vegetables? Yes Is Selected Q130 What diseases or health problems do you think are related to a low intake of fruits and vegetables? Q131 Are you aware of any major health problems or diseases that are related to how much sugar people eat? Yes (1) No (2) Unsure (3) Q132 What diseases or health problems do you think are related to sugar? Q133 Are you aware of any major health problems or diseases that are related to the amount of fat people eat? Yes (1) No (2) Unsure (3) Q134 What diseases or health problems do you think are related to fat? Q135 Now I am going to ask you some questions about nutrition for pregnant and lactating women. Please let me know if you need me to clarify any of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q136 How should a pregnant woman eat in comparison to a non pregnant woman to provide good nutrition to her baby to help him or her grow? Please list 4 practices she should do. Mark only the answers that the respondent names. Do not read these choices to the respondent. Eat more foo d / more energy (1) Eat more protein rich foods (2) Eat more iron rich foods (3) Use iodized salt when preparing meals (4) Other (5) ____________________ I don't know (6)

PAGE 165

165 Q137 Are there common supplements, or tablets, recommended to women during pregnancy ? If so, can you name them? Mark only the answers that the respondent names. Do not read these choices to the respondent. Iron supplements (1) Folic acid supplements (2) None (5) Other (3) I don't know (4) Q138 How should a lactating woman eat in comparis on with a nonlactating woman to be healthy and produce more breastmilk? Mark only the answers that the respondent names. Do not read these choices to the respondent. Eat more food / more energy (1) Eat more protein rich foods (2) Eat more iron rich foods (3) Use iodized salt when preparing meals (4) Other (5) I don't know (6) Q139 Now I have some questions about breastfeeding practices and infant and child nutrition. Please let me know if you need me to clarify any of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q140 How long should a baby receive nothing but breastmilk?Probe if necessary: Until what age is it recommended that a mother feeds nothing more than breastmilk? From birth to 6 months (1) From birth to 11 months (2) Other (3) I don't know (4) Q141 How long is it recommended that a woman breastfeeds her child?Probe if necessary: Until what age is it recommended that a mother continues breastfeeding? Six months or less (1) 611 months (2) 1224 months (3) More than 24 months (4) Other (5) I don't know (6) Q142 Now I have some questions for you about food purchases and agriculture markets.

PAGE 166

166 Q143 Are you the person who makes the final decision about food purchases a nd preparation? Yes (1) No (2) Answer If Are you the person who makes the final decision about food purchases and preparation? No Is Selected Q144 Who makes the final decision regarding food purchases and preparation? Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q145 If you had more income, would you spend more of your money on food? Yes (1) No (2) Q146 If you had more money available to spend on food would you consume t he same types of food? Yes (1) No (2) Q147 If you had more money available to spend on food, which of the following would you consume more of? (Mark all that apply) Cereals (1) Pulses (2) Vegetables (3) Fruits (4) Meat (5) Eggs (6) Fish (7) Milk or milk products (8) Sweets (9) Other foods (10) ____________________ I would not spend more money on food. (11)

PAGE 167

167 Q148 Did anyone in your household buy any food (from a market) to cook in the household in the last year? Yes (1) No (2) Answer If Did anyone in you r household buy any food (from a market) to cook in the household in the last year? Yes Is Selected Q149 How long does it take to walk to a place to buy food? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Answer If Did anyone in your household buy any food (from a market) to cook in the household in the last year? Yes Is Selected Q150 What mode of transportation does your household use to go to the market for buying food? By foot (1) By bicycle (2) By rickshaw/van (3) By car/truck (4) By motorcycle (5) By boat (6) Other (7) ____________________ Answer If Did anyone in your household buy any food (from a market) to cook in the household in the last ye... Yes Is Selected Q151 How much does this transport ation cost per trip to the market? (BDT) Q152 Does anyone in your household ever sell commercial agricultural products grown in your household (e.g. rice, maize)? Yes (1) No (2)

PAGE 168

168 Answer If Does anyone in your household ever sell commercial agricultural products grown in your household (e.g. rice, maize)? Yes Is Selected Q153 How long does it take to walk to the place to sell commercial agricultural products, for example to a market or buyer pick up location? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5) Answer If Does anyone in your household ever sell commercial agricultural products grown in your household? Yes Is Selected Q154 Who controls the income from sales of commercia l agricultural products? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q155 Does anyone in your household ever sell vegetables or fruits grown in your household? (e.g. pumpkins, cuc umbers, mangos) Yes (1) No (2) Answer If Does anyone in your household ever sell vegetables or fruits grown in your household? (e.g. pumpkins, cucumbers, mangos) Yes Is Selected Q156 How long does it take to walk to the place to sell vegetables or fruits? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 169

169 Answer If Does anyone in your household ever sell vegetables or fruits grown in your household? Yes Is Selected Q1 57 Who controls the income from the sales of vegetables/fruits? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q158 Does anyone in your household ever sell fish produced in your household? Yes (1) No (2) Answer If Does anyone in your household ever sell fish produced in your household?; Yes Is Selected Q159 How long does it take to walk to the place to sell fish? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5) Q160 Does anyone in your household ever sell animal protein (meat) produced in your household? Yes (1) No (2) Answer If Does anyone in your household ever sell animal protein (meat) produced in your house hold? Yes Is Selected Q161 How long does it take to walk to the place to sell animal protein (meat)? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 170

170 Answer If Does anyone in your household ever sell animal protein (meat) produced in your household? Yes Is Selected Q162 Who controls the income from the animal protein (meat)? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servan t) (17) Q163 Does anyone in your household ever sell eggs produced in your household? Yes (1) No (3) Answer If Does anyone in your household ever sell eggs produced in your household? Yes Is Selected Q164 How long does it take to walk to the place to sell eggs? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5) Answer If Does anyone in your household ever sell eggs produced in your household? Yes Is Selected Q165 Who controls th e income from the sale of eggs? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) ____________ Q166 Does anyone in your household ever sell milk or milk products pro duced in the household? Yes (1) No (2)

PAGE 171

171 Answer If Does anyone in your household ever sell milk or milk products produced in the household? Yes Is Selected Q167 How long does it take to walk to the place to sell milk or milk products? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5) Answer If Does anyone in your household ever sell milk or milk products produced in the household? Yes Is Selected Q168 Who controls the income from the sale of milk or milk products? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) ____________________ Answer If Yes Was Selected for Sale of Any Products. Q169 What mode of transportation does your household use to transport agricultural goods, fish, or vegetables/fruits to the market/selling points? By foot (1) By bicycle (2) By rickshaw/van (3) By car/truck (4) By motorcycle (5) By boat (6) Other (7) ____________________ Q170 Were there months, in the past 12 months, in which you did not have enough food to meet your family's needs? Yes (1) No (2)

PAGE 172

172 Answer If Were there months, in the past 12 months, in which you did not have enough food to meet your family's needs ? Yes Is Selected Q171 During which months did you not have enough food to meet your family's needs? (Check all that apply) January (1) February (2) March (3) April (4) May (5) June (6) July (7) August (8) September (9) October (10) November (11) December (12) Q172 In the past four weeks, did you worry that your household would not have enough food? Yes (1) No (2) Answer If In the past four weeks, did you worry that your household would not have enough food? Yes Is Selected Q173 How o ften did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q174 In the past four weeks were you or any household member not able to eat the kinds of food you preferred because of a lack of resources to get food? Yes (1) No (2) Answer If In the past four weeks were you or any household member not able to eat the kinds of food you preferred because of a lack of resources to get food? Yes Is Selected Q175 How often did this happen? Ra rely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q176 In the past four weeks, did you or any household member have to eat some foods that you really did not want to eat because of a lack of resources to obtain other types of food? Yes (1) No (2)

PAGE 173

173 Answer If In the past four weeks, did you or any household member have to eat some foods that you really did not want to eat because of a lack of resources to obtain other types of food? Yes Is Selec ted Q177 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q178 In the past four weeks, did you or any household member have to eat a smaller meal than you felt you need ed because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member have to eat a smaller meal than you felt you needed because there was not enough food ? Yes Is Selected Q179 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q180 In the past four weeks, did you or any household member have to eat fewer meals in a day because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member have to eat fewer meals in a day because there was not enough food? Yes Is Selected Q181 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q182 In the past four weeks, was there ever no food of any kind in your household because of a lack of resources to get food? Yes (1) No (2) Answer If In the past four weeks, was there ever no food of any kind in your h ousehold because of a lack of resources to get food? Yes Is Selected Q183 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3)

PAGE 174

174 Q184 In the past four weeks, did you or any h ousehold member go to sleep at night hungry because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member go to sleep at night hungry because there was not enough food? Yes Is Selected Q185 How often d id this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q186 In the past four weeks, did you or any household member go a whole day and night without eating anything because there was no t enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member go a whole day and night without eating anything because there was not enough food? Yes Is Selected Q187 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q188 In the past 12 months, how often did you or any of your family have to eat potato, wheat, or another grain although you wanted to eat rice (not including when you were sick)? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5) Q189 In the past 12 months, how often did you or any of your family skip entire meals due to scarcity of food? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5)

PAGE 175

175 Q190 In the pas t 12 months, how often did you personally eat less food in a meal due to scarcity of food? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5) Q191 In the past 12 months, how often did your family purchase food (rice, lentils, etc.) on credit or loan from a local shop? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) O ften (a few times in each month) (4) Regularly (every day or almost every day) (5) Q192 In the past 12 months, how often did your family have to borrow/take food from relatives or neighbors to make a meal? Never (1) Rarely (1 6 times in the last 12 months ) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5)

PAGE 176

176 APPENDIX C ENDLINE SURVEY Q466 Please make sure the GPS locator is ON for your tablet. Q1 Hello, my name is _________________ and I am conducting research on behalf of the Bangladesh Agricultural University and the University of Florida. We are contacting you to follow up on our research about household food availability and access. This research is for a doct oral dissertation at the University of Florida in the United States of America. First of all, I would like to thank you for taking the time to meet with me. We would like to ask you some questions about the ____________(BAUEC/Sushilan) event you attended and your food consumption and diet. This interview should take approximately 30 minutes to 1 hour. Your participation in this study will be confidential to the extent provided by law, and your responses and comments will be anonymous. There are no direct b enefits, risks, or compensation to you for participating in this study. Participation is completely voluntary, and you may withdraw your consent to participate at any time without penalty. Q3 I will be asking questions about your household members and y ourself, your household income, agricultural production, food availability and access. Your responses will be completely private and will only be connected with your name through an identification number. We understand that you may not have all of the prec ise information available. It is important for the information in this study to be as accurate as possible. Please try to answer as many questions as you can, but if you cannot remember some information, it is ok to answer I do not remember. Finally, you have the right to not answer any question we might ask. You may also choose to stop answering questions at any time. If you do, we will not collect any more information from you. However, we would keep and use the information we had already collected from you. Q4 Given this information, are you willing to participate in this survey? Yes (1) No (2) Q359 Did you attend the _________ (BAUEC/Sushilan) event and meal? Yes (1) No (2) Answer If Given this information, are you willing to participate in this survey? No Is Selected Q358 For our records, can you tell us why you do not want to participate? Note to interviewer: please remind the respondent about the privacy of their responses and that the individual does not have to know all of the answers. The st udy is anonymous. Will you reconsider? Yes (1) No (2)

PAGE 177

177 Q5 Enumerator Name (Last name, First name) Afroza (2) Alamin (17) Ame (4) Anik (6) Anup (7) Ayesha (8) Azil (9) Diti (37) Eva (10) Himu (11) Jenny (12) Mahmud (14) Mumu (15) Nilima (39) Nosib (3) Pinkey (18) Prodip (19) Rabbi (20) Rubyeat (21) Sadia (Urvi) (24) Sajib (22) Shakir (25) Shaon (26) Shatabdi (36) Shirin (38) Shoaib (31) Shuvo (27) Tapoti (35) Tania (29) Wahidul (28) Q467 GPS Coordinates Please enter the coordinates from the "Simple GPS Coordinate Display" app installed on your tablet. Latitude (1) Longitude (2) Q6 District Mymensingh (1) Borguna (2) Q7 Upazila Mymensingh Sadar (1) Amtoli (2) Q8 Union Chowra (1) Holdia (2)

PAGE 178

178 Answer If District Borguna Is Selected Q9 Sushilan Village Baitakata (1) Baitmore (2) Chalavanga (3) Chandra (4) Chandra Kapali (5) Ghatkhali (6) Gilatali (7) Holdia (8) Holodia (Paka) (9) Holodia (10) Kalibari (11) Kawnia (12) Kawnia Kapali (13) Loda (14) Patakata (15) Patakata midle (16) Uttar Tokta Bunia (17)

PAGE 179

179 Answer If District Mymensingh Is Selected Q316 BAU Village Boira moddopara sorkar bari (1) Boira mosque (2) Borobilarpar (3) Bot tola bazar (4) Bot tola hatem bapari (5) Chor nilokkhai ujanpara (6) Chork alibari shomvuganj (7) Chornillokhia digolpara (8) Goneshampur (9) Gosta north (10) Jhaugau (11) Jogir ali chor nilokkhia (12) Kismat khagdohor (13) Maijbari (14) Mirjapur east (15) Mirjapur north (16) Mirjapur south (17) Muktijodda bazar chor (18) Mukhtijoddha bazar jogir algi (19) Pagla Bazaar (20) Pagla bazar kazi bari (21) Pagla bazar mojiborer bari (22) Ragobpur (23) Sathiapara (24) Sathiapra chor nillokkhia (25) Shailmari purrush (26) Suhila moddopara (27) Suhila nodir par (28) Suhila north (29) Sutiakhali middle (30) Sutiakhali palpara 1 (31) Sutiakhali palpara 2 (32) Vabokhali pwest (33) Vabukhali (34) Q317 BHH Code (Sushilan) or S.I. Number (BAUEC) Q473 Participant meal ID number Q10 Thank you for participating in this study. I would like to start by asking some questions about you and your household members. Q11 Respondent Name First name (1) Last name (2) Father/Husband name (3) Q12 Do you have a mobile number where we can reach you? If so, what is your number? Q13 Gender Male (1) Female (2) Q14 Are you the head of household? Yes (1) No (2)

PAGE 180

180 Answer If Are you the head of household? No Is Selected Q15 What is your relationship to the head of household? Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law/Son in law (8) Brother/Sister (9) Father in law/Mother in law (10) Nephew/Niece (11) Grandfather/Grandmother (12) Grandson/Granddaughter (13) Sister in law/Brother in law (14) Brother's wife (15) Other (e.g. servant) (16) Q16 What is your marital status? Married (1) Single (never married) (2) Widowed (3) Divorced (5) Separated (4) Q17 What is your age? Q18 What is the highest level of education you have completed? None (1) Class 1 (2) Class 2 (3) Class 3 (4) Class 4 (5) Class 5 (6) Class 6 (7) Class 7 (8) Class 8 (9) Class 9 (10) Class 10 (17) SSC Pass (11) HSC Pass (12) Graduate (13) Post graduate (14) Medical (15) Vocational/Technical education (16)

PAGE 181

181 Q19 What is your primary occupation? Farming (own land) (1) Poultry and livestock rearing (2) Sharecropper (3) Agricultural day labor/contract labor (4) Fishing (own boat) (5) Fishing labor (someone else's boat) (6) Fish farming (7) Boat operation (8) Rickshaw/van operator (9) Casual labor (10) Self employed in business/petty business (11) Non agricultural day labor/contract labor (12) Regular salaried employment (13) Paid "volunteer" (14) Housework (child care/home care) (15) Servant/ma id (16) Student (17) Beggar (18) Unemployed (19) Old/Disabled (20) Other (21) ____________________ N/A (22) Answer If Primary occupation Other Is Selected Q20 Specify your primary occupation. Q21 What is your secondary occupation?Far ming (own land) (1) Poultry and livestock rearing (2) Sharecropper (3) Agricultural day labor/contract labor (4) Fishing (own boat) (5) Fishing labor (someone else's boat) (6) Fish farming (7) Boat operation (8) Rickshaw/van operator (9) Casual labor (10) Self employed in business/petty business (11) Non agricultural day labor/contract labor (12) Regular salaried employment (13) Paid "volunteer" (14) Housework (child care/home care) (15) Servant/maid (16) Student (17) Beggar (18) Unemployed (19) Old/Disabl ed (20) Other (21) ____________________ N/A (22) Answer If Secondary occupation Other Is Selected Q22 Specify your secondary occupation.

PAGE 182

182 Q23 Now I would like to ask some questions about your household members.How many individ uals, NOT including yourself, live in your household? Please include the people who habitually eat and sleep in the home, including those who have been absent less than six months and have not established another residence. This includes maids or servants who live in the household. 0 (21) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) 8 (8) 9 (9) 10 (10) 11 (11) 12 (12) 13 (13) 14 (14) 15 (15) 16 (16) 17 (17) 18 (18) 19 (19) 20 (20) Q360 In the last four months, has anyone in your household migrated/moved away? Yes (1) No (2) Answer If In the last four months, has anyone in your household migrated/moved away? Yes Is Selected Q362 How many people moved away? Q405 In the last four months, has any new member moved into your household? Yes (1) No (2) Answer If In the last four months, has any new member moved into your household? Yes Is Selected Q406 How many people moved into your household? Q381 Who moved away? First Name (1)

PAGE 183

183 Q361 What is ${q://QID406/ChoiceTextEntryValue/1}'s relationship to the head of household Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q444 What is ${q://QID406/ChoiceTextEntryValue/1}'s gender? Male (1) Female (2) Q445 What is ${q://QID406/ChoiceTextEntryValue/1}'s age? Q432 What is the name of the individual who moved into your household in the last four months? First Name (1)

PAGE 184

184 Q433 What is ${q://QID432/ChoiceTextEntryValue/1}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q447 What is ${q://QID432/ChoiceTextEntryValue/1}'s gend er? Male (1) Female (2) Q448 What is ${q://QID432/ChoiceTextEntryValue/1}'s age? Q374 In the last four months, has any household member's level of education changed? Yes (1) No (2) Q375 How many household members' level of education has changed? Q380 Who completed a higher level of education?

PAGE 185

185 Q378 What is ${q://QID405/ChoiceTextEntryValue}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mot her (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q449 What is ${q://QID405/ChoiceTextEntryValue}'s gender? Male (1) Female (2) Q450 What is ${q://QID405/ChoiceTextEntryValue}'s age? Q379 What is the highest level of education ${q://QID405/ChoiceTextEntryV alue} has completed? None (1) Class 1 (2) Class 2 (3) Class 3 (4) Class 4 (5) Class 5 (6) Class 6 (7) Class 7 (8) Class 8 (9) Class 9 (10) Class 10 (11) SSC Pass (13) HSC Pass (14) Graduate (15) Post graduate (16) Medical (17) Vocational/Technical Education (18) Q365 In the last four months, has anyone in your household gotten married? Yes (1) No (2) Q382 How many household members got married in the last four months?

PAGE 186

186 Q385 In the last four months, has anyone in your household gotten divorced or separated? Divorced (1) Separated (4) No (3) Q386 How many household members got divorced or separated? Q383 Who got married? First Name (1) Q367 What is ${q://QID408/ChoiceTextEntryValue/1}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Gran dfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q452 What is ${q://QID408/ChoiceTextEntryValue/1}'s gender? Male (1) Female (2) Q451 What is ${q://QID408/ChoiceTextEntryValue/1}'s age? Q384 Does ${q://QID408/ChoiceTextEntryValue/1} still live in your household? Yes (1) No (2) Q387 Who got ${q://QID410/ChoiceGroup/SelectedChoices}? First Name (1)

PAGE 187

187 Q388 What is ${q://QID412/Choic eTextEntryValue/1}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q453 What is ${q://QID412/ ChoiceTextEntryValue/1}'s gender? Male (1) Female (2) Q454 What is ${q://QID412/ChoiceTextEntryValue/1}'s age? Q389 Does ${q://QID412/ChoiceTextEntryValue/1} still live in your household? Yes (1) No (2) Q368 In the last four months, has anyone in your household had a baby? Yes (1) No (2) Answer If In the last four months, has anyone in your household had a baby? Yes Is Selected Q371 How many household members had a baby in the last four months? 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) Q370 Who had a baby? First name (1)

PAGE 188

188 Q390 What is ${q://QID395/ChoiceTextEntryValue/1}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) ( 21) Q455 What is ${q://QID395/ChoiceTextEntryValue/1}'s age? Q372 How old is the baby now? Q373 Is the mother breastfeeding the baby? Yes (1) No (2) Q391 In the last four months, has any household member become pregnant? Yes (1) No (2) Q392 How many household members became pregnant in the last four months? 0 (0) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) Q393 Who became pregnant? First name (1)

PAGE 189

189 Q394 What is ${q://QID418/ChoiceTextEntryValue/1}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Gran dfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brother's wife (20) Other (e.g. servant) (21) Q456 What is ${q://QID418/ChoiceTextEntryValue/1}'s age? Q395 In the last four months, have any of your household members died? Yes (1) No (2) Q396 How many of your household members have died in the last four months? 0 (0) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) 8 (8) 9 (9) 10 (10) Q397 What is the name of the household member who died? Q39 8 How old was ${q://QID422/ChoiceTextEntryValue}?

PAGE 190

190 Q400 What is ${q://QID422/ChoiceTextEntryValue}'s relationship to the head of household? Household head (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Father in law (8) Mother in law (9) Daughter in law (22) Son in law (23) Brother (10) Sister (11) Nephew (12) Niece (13) Grandfather (14) Grandmother (15) Grandson (16) Granddaughter (17) Sister in law (18) Brother in law (19) Brothe r's wife (20) Other (e.g. servant) (21) Q457 What is ${q://QID422/ChoiceTextEntryValue}'s gender? Male (1) Female (2) Q35 Thank you for the information about your household members. Now I have a few more questions about you. Answer If Gender Female Is Selected Q36 Are you currently pregnant? Yes (1) No (2) Unsure (3) Answer If Gender Female Is Selected Q37 Are you currently breastfeeding? Yes (1) No (2) Answer If Are you currently breastfeeding? Yes Is Selected Q38 Please tell me the age (months) of each child you are breastfeeding. First child (1) Second child (2) Third child (3) Q403 How many children ages 6 to 12 years old currently live in your household? Q404 How many children ages 6 to 12 years old currently attend a school or educational institution?

PAGE 191

191 Q39 What religion are you? Islam (1) Hindu (2) Buddhist (3) Christian (4) Other (5) ____________________ N/A (6) Answer If What religion are you? Other Is Selected Q40 Specify religion. Q41 In the last year, did anyone in your h ousehold ever do any work for which he or she was paid on a daily basis? Yes (1) No (2) Q42 How many months were you employed in the last year? Q43 What was your monthly income (BDT)? Q44 What is your monthly household income (BDT)? Include income fro m all household members. Q45 What is your annual household income (BDT)? Include income from all household members. Q46 Are you the person who usually goes to the market to purchase food for the household? Always (11) Most of the time (12) About half the time (13) Sometimes (14) Never (15)

PAGE 192

192 Q409 Do any other household members usually go to the market to purchase food for the household? Member 1 Member 2 Member 3 Member 4 Head of household (1) Wife of household head (2) Husband of household head (3) Son (4)

PAGE 193

193 Q47 Next I will ask some questions about your household assets and landholding. Q48 What type of latrine does the household use? Open field (1) Kacha latrine (temporary or permanent) (2) Pacca (pit or water seal) (3) Sanitary (4) Q49 How many rooms does the household occupy (excluding rooms used for business)? Q50 What is the main construction material of the walls of the main room in the home? Tile/wood (1) Hemp/hay/bamboo (2) C.I. sheet (3) Cement (4) Q51 Does the household own a television? Yes (1) No (2) Q52 How many fans does the household own? Q53 How many mobile phones does the household own? Q54 Does the household own any bicycles, motorcycles/scooters, or motor cars, etc.? Yes (1) No (2) Q55 How much land does the household own? Specify for each land use. Cultivable land (currently cropped or fallow) (1) Dwelling house/homestead (2) Non cultivated land (not including homestead) ( 3) Land Owned (Decimals) (1) Q56 How much additional cultivable land is rented/sharecropped/mortgaged in or out by the household? (excluding uncultivable land and dwellinghouse/homestead land) Rented/sharecropped/morgaged IN (2) Rented/sharecropped/mortgaged OUT (4) Additional Cultivable Land (Decimals) (1)

PAGE 194

194 Q57 How much total land does the household have access to? (Decimals)Note: Verify that this equals the sum of the total land owned plus additional cultivable land. Q115 Are you the person who normally prepares food for the household? Yes (1) No (2) Answer If Are you the person who normally prepares food for the household? No Is Selected Q116 In your household, who normally prepares food? Note: if the respondent answers the name of the person, clarify that person's relationship to the head of household. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Broth er (10) Sister (11) Father in law (12) Mother in law (18) Nephew (13) Niece (19) Grandfather (14) Grandmother (20) Sister in law (15) Brother in law (21) Brother's wife (16) Other (e.g. servant) (17) Answer If In your household, who normally prepares food? Note: if the respondent answers the name of the person, clarify that person's relationship to the head of household. Other (e.g. servant) Is Selected Q117 Specify other who prepares the food for the household. Q118 Please descr ibe the foods (meals and other snacks) that you ate or drank yesterday during the day and night, whether at home or outside the home. Start with the first food or drink in the morning. Breakfast (1) Snack (2) Lunch (3) Snack (4) Dinner (5) Q119 When the respondent recall of meals consumed in the last 24 hours is complete, fill in the food groups consumed by the respondent based on the information recorded above. Remember to ask whether that meal was consumed inside the home or outside of the home. For any food groups not mentioned in the meals and snacks, prompt the respondent by asking if he or she consumed a food item from this group. If the respondent has not consumed a food group, ask if anyone in the household consumed food items from that food group in the last 24 hours.

PAGE 195

195 Has anyone in the household consumed? Did the respondent consume? Yes inside the home (1) Yes outside the home (2) Inside and outside the home (3) Unsure (4) No (5) Yes inside the home (1) Yes outside the home (2) Inside and outside the home (3) No (4) Cereals (rice, bread, wheat, rice flakes, puffed rice, barley, biscuits, popcorn) White roots and tubers (white potatoes, turnips) Vitamin A rich vegetables and tubers (pumpkin, carrots, squash, orange flesh sweet potatoes) Dark leafy green vegetables (amaranth, spinach, taro leaf, pumpkin leaf, jute leaf) Other vegetables (cucumber, radish, pepper,

PAGE 196

196 cabbage, beans, cauliflower, onion) Vitamin A rich fruits (ripe papaya, ripe mangos, orange, grapefruit) Other fruits (banana, apple, jackfruit, watermelon, guava, plums, pineapple, melons, lemons, lychees) Meat (chicken, beef, poultry, lamb, liver, kidney) Eggs Fish and seafood Legumes, nuts, and seeds (lentils, dal, black gram, kheshari, mung beans) Milk and milk products (cow milk, goat milk,

PAGE 197

197 powdered milk, yogurt, curd, cheese) Oils and fats (butter, ghee, mustard oil, soybean oil) Sweets (sugar, honey, molasses, chocolate, ice cream, soda) Spices, condiments, beverages (black pepper, salt, fish powder, soy sauce, chili sauce, cumin, chili powder, cinnamon, coffee, tea) Q120 Now I would like to ask you about all of the different foods your household members have eaten in the last 7 days. Could you please tell me how many days in the past week your household has eaten the following foods? This can be inside the home or outside the home. (Enter a number 0 7)

PAGE 198

198 Number of DAYS eaten in the past week. (0 7) Primary Source Secondary Source Cereals (rice, bread, wheat, rice flakes, puffed rice, barley, biscuits, popcorn) (1) White roots and tubers (white potatoes, turnips) (2) Vitamin A rich vegetables and tubers (pumpkin, carrots, squash, orangeflesh sweet potatoes) (3) Dark leafy green vegetables (amaranth, spinach, taro leaf, pumpkin leaf, jute leaf) (4) Other vegetables (cucumber, radish, pepper, cabbage, beans, cauliflower, onion) (5) Vitamin A rich fruits (ripe papaya, ripe mango, orange, grapefruit) (6) Other fruits (banana, apple, jackfruit, watermelon, guava, plums, pineapple, melons, lemons, lychees) (7) Meat (chicken, beef, poultry, lamb, liver, kidney) (8) Eggs (9) Fish and seafood (10) Legumes, nuts, and seeds (lentils, dal, black gram, kheshari, mung beans) (11) Milk and milk products (cow milk, goat milk, powdered milk, yogurt,

PAGE 199

199 curd, cheese) (12) Oils and fats (butter, ghee, mustard oil, soybean oil) (13) Sweets (sugar, honey, molasses, chocolate, ice cream, soda) (14) Spices, condiments, beverages (black pepper, salt, fish powder, soy sauce, chili sauce, cumin, chili powder, cinnamon, coffee, tea) (15) Q122 I am going to ask you some questions about nutrition, vitamins and food. This is a survey, not a test. Y our answers will help identify which dietary advice people find confusing. Please let me know if you need me to clarify any of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q123 Do you think that health experts recommend that people should be eating more, the same amount, or less of these foods? (Mark one box per food) More (1) Same (2) Less (3) I don't know (4) Vegetables (1) Sweets (2) Meat (3) Carbohydrates (4) Fatty foods (5) High fibre foods (6) Fruit (7) Salty foods (8) Q124 How many servings of fruits and vegetables per day do you think experts are advising people to eat? (One serving could be, for example, an apple or a spoonful of cucumbers). The response should be in number of times per day/servings, not grams. Q125 Experts classify food into groups. We are interested to see whether people are aware of what foods are in these groups.

PAGE 200

200 Q126 Do you think the following foods are high or low in vitamin A? High (1) Low (2) Unsure (3) Carrot (1) Banana (27) Rice (28) Wheat flour (47) Mola fish (49) Leafy greens (spinach, amaranth) (50) Q127 Do you think experts put these foods in the carbohydrates group? Yes (1) No (2) Unsure (3) Rice (1) Potatoes (3) Sweet potatoes (4) Chapati (5) Biscuits (6) Q128 Do you think these foods are high or low in protein? High (1) Low (2) Unsure (3) Chicken (1) Fish (2) Egg (3) Pulses (4) Milk (5) Fruit (6) Rice (7) Q129 Are you aware of any major health problems or diseases that are related to a low intake of fruits and vegetables? Yes (1) No (2) Unsure (3)

PAGE 201

201 Answer If Are you aware of any major health problems or diseases that are related to a low intake of fruits and vegetables? Yes Is Selected Q130 What diseases or health problems do you think are related to a low intake of fruits and vegetables? Q131 Are you aware of any major health problems or diseases that are related to how much sugar people eat? Yes (1) No (2) Unsure (3) Q132 What diseases or health problems do you think are related to sugar? Q133 Are you aware of any major health problems or diseases that are related to the amount of f at people eat? Yes (1) No (2) Unsure (3) Q134 What diseases or health problems do you think are related to fat? Q135 Now I am going to ask you some questions about nutrition for pregnant and lactating women. Please let me know if you need me to clarify a ny of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q136 How should a pregnant woman eat in comparison to a non pregnant woman to provide good nutrition to her baby to help him or her grow? Please list 4 practices she should do. Mark only the answers that the respondent names. Do not read these choices to the respondent. Eat more food / more energy (1) Eat more protein rich foods (2) Eat more iron rich foods (3) Use iodized salt when preparing meals (4) Other (5) ____________________ I don't know (6) Q137 Are there common supplements, or tablets, recommended to women during pregnancy? If so, can you name them? Mark only the answers that the respondent names. Do not read thes e choices to the respondent. Iron supplements (1) Folic acid supplements (2) None (5) Other (3) I don't know (4)

PAGE 202

202 Q138 How should a lactating woman eat in comparison with a non lactating woman to be healthy and produce more breastmilk? Mark only the answer s that the respondent names. Do not read these choices to the respondent. Eat more food / more energy (1) Eat more protein rich foods (2) Eat more iron rich foods (3) Use iodized salt when preparing meals (4) Other (5) I don't know (6) Q139 Now I have som e questions about breastfeeding practices and infant and child nutrition. Please let me know if you need me to clarify any of my questions. Feel free to ask me any questions you may have. If you do not know the answer, it is ok to respond "I don't know." Q140 How long should a baby receive nothing but breastmilk?Probe if necessary: Until what age is it recommended that a mother feeds nothing more than breastmilk? From birth to 6 months (1) From birth to 11 months (2) Other (3) I don't know (4) Q141 How long is it recommended that a woman breastfeeds her child?Probe if necessary: Until what age is it recommended that a mother continues breastfeeding? Six months or less (1) 611 months (2) 1224 months (3) More than 24 months (4) Other (5) I don't know (6)

PAGE 203

203 Q326 The following questions refer to the BAUEC/Sushilan event you attended and the nutrition training session. Your answers will be kept confidential and will not affect your participation in future events or programs. Q327 How would you rate the information presented in the nutrition training "What goes on the plate"? (1 is not helpful at all, 5 is very helpful) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q328 How likely are you to share the information from the nutrition training "What goes on the plat e" with other individuals (family members, friends, etc.)? (1 is not at all likely, 5 is very likely) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q410 With whom are you most likely to share the nutrition training "What goes on the plate" information? Spouse (1) Parents or in laws (2) Children (3) Other family members (5) Friends female (6) Friends male (7) Other (8) ____________________ No one (9) I did not receive nutrition training. (10)

PAGE 204

204 Q329 Please rank the facilitators of the nutrition training. (1 i s poor, 5 is excellent) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q411 How would you rate the information presented in the gender training "Who gets what to eat"? (1 is not helpful at all, 5 is very helpful) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q412 How likely are you to share the information from the gender training "Who gets what to eat" with other individuals (family members, friends, etc.)? (1 is not at all likely, 5 is very likely) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q413 With whom are you most likely to share the information from the gender training "Who gets what to eat" information? Spouse (1) Parents or in laws (2) Children (3) Other family members (5) Friends female (6) Friends male (7) Other (8) ____________________ No one (9) I did not receive nutrition training. (10)

PAGE 205

205 Q414 Please rank the facilitators of the gender training. (1 is poor, 5 is excellent) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) N/A (6) Q330 Please rank the quality of the lunch buffet meal at the event. (1 is poor, 5 is excellent) 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) Q331 Did you like the food at the event? Yes (1) No (2) Q332 Show the plate to the participant and say "Now I would like to ask you some questions about this FHI360 SHIKHA food plate." Q415 At the BAUEC/Sushilan event, did you receive the FHI360 SHIKHA food plate to take home? Yes (1) No (2) Answer If At the BAUEC/Sushilan event, did you receive this plate to take home? Yes Is Selected Q333 Did you receive enough partitioned plates for each member of your household? Yes (1) No (2) Q425 Do you still have all of the FHI360 SHIKHA food plates you were given? Yes (1) I have some but I gave one or more away to someone else (2) I have some but I disposed of one or more plate (3) I no longer have any SHIKHA plates because I gave them all away (6) I no longer have any SHIKHA plates because I disposed of them (7) I did not receive any SHIKHA plates (8)

PAGE 206

206 Q468 How often do you use the FHI360 SHIKHA food plate to eat your meals? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) I did not receive any SHIKHA plates (5) Q335 How often do you refer to the plate to make decisions on the type of food to prepare? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) I did not receive any SHIKHA plates (5) Q437 How often do you refer to the plate to make decisions on the type of food to consume? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More th an 10 times in last 4 weeks) (3) I did not receive any SHIKHA plates (5) Q336 How often do you refer to the plate to make decisions on the amount of food to serve others? Never (1) Rarely (4) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) I did not receive any SHIKHA plates (5) Q432 How often do you refer to the plate to make your own decisions about the amount of food to consume? Never (1) Rarely (4) Sometimes (3 4 times in the last 4 weeks) (2) Often (More t han 10 times in last 4 weeks) (3) I did not receive any SHIKHA plates (5) Q334 In the last 7 days, have you used this FHI360 SHIKHA food plate during a meal eaten during the day or at night? Yes (1) No (2)

PAGE 207

207 Answer If In the last 7 days, have you used this FHI360 SHIKHA food plate during a meal eaten during the day or at night ? Yes Is Selected Q424 How many times have you used the plate during a meal in the last 7 days? Q469 Why do you not use/refer to the FHI360 SHIKHA food plate? I prefer other to us e plates during my meal (1) I do not have access to the food items displayed on the plate (2) I do not understand the plate message (3) I cannot afford the food items displayed on the plate (4) Other (5) N/A (8) Answer If Why do you not use the FHI360 SHI KHA food plate? Other Is Selected Q470 Describe "other" reason why you do not use the FHI360 SHIKHA food plate. Answer If Why do you not use the FHI360 SHIKHA food plate? I do not have access to the food items displayed on the plate Is Selected Q471 What type of food items on the FHI360 SHIKHA food plate do you not have access to? Meat (1) Fish (2) Vegetables (3) Rice (4) Dal (5) Answer If Why do you not use the FHI360 SHIKHA food plate? I cannot afford the food items displayed on the plate Is Selected Q472 What type of food items on the FHI360 SHIKHA food plate are not affordable for you? Meat (1) Fish (2) Vegetables (3) Rice (4) Dal (5) Q339 In the last 7 days, how many people in your household, other than yourself, have used the FHI360 SHIKHA food pl ate? Q356 How many people did you talk to about the BAUEC/Sushilan event, the nutrition or gender training, or about the FHI360 SHIKHA food plate?

PAGE 208

208 Q343 I would like to ask more information about the individuals in your household who are using the plate Answer If In the last 7 days, has anyone used the partitioned plate during a meal eaten during the day or at night ? Yes Is Selected Q341 Who has used the plate? First name (1) Q427 Gender of ${q://QID366/ChoiceTextEntryValue/1} Male (1) Female (2) Q428 What is ${q://QID366/ChoiceTextEntryValue/1}'s age? Q426 What is ${q://QID366/ChoiceTextEntryValue/1}'s relationship to the head of household? Head of household (19) Wife of head of household (41) Husband of head of household (40) Mother in law (20) Father in law (21) Cousin Female (3) Cousin Male (4) Aunt (5) Uncle (6) Grandfather (7) Grandmother (8) Father (9) Mother (10) Daughter (11) Son (12) Granddaughter (13) Grandson (14) Sister (15) Brother (16) Niece (17) Nephew (18) Q416 How many times in the last 7 days has ${q://QID366/ChoiceTextEntryValue/1} used the FHI360 SHIKHA food plate during a meal? Q429 How did ${q://QID366/ChoiceTextEntryValue/1} use the FHI360 SHIKHA food plate during the meal(s) in the last 7 days? (Select all that apply) Ate a meal using the plate (2) Referred to the SHIKHA plate to make decisions about what food to prepare (1) Referred to the SHIKHA plate to make decisions about what type of food to consume (7) Referred to the SHIKHA p late to make decisions about how much food to serve others (3) Referred to the SHIKHA plate to make own decisions about how much food to consume (4) Other (5) Unsure (6)

PAGE 209

209 Answer If How has {q://QID366/ChoiceTextEntryValue/1} used the FHI360 SHIKHA food plat e? Other Is Selected Q430 Specify how this person has used the plate if "other" is selected. Q433 How often does ${q://QID366/ChoiceTextEntryValue/1} refer to the plate to make decisions on the type of food to prepare? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) Unsure (6) Q436 How often does ${q://QID366/ChoiceTextEntryValue/1} refer to the plate to make decisions on the type of food to consume? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) Unsure (7) Q434 How often does ${q://QID366/ChoiceTextEntryValue/1} refer to the plate to make decisions on the amount of food to serve others? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) Unsure (6) Q435 How often does ${q://QID366/ChoiceTextEntryValue/1} refer to the plate to make his/her own decisions about how much food to consume? Never (1) Rarely (4) Sometimes (3 4 times in last 4 weeks) (2) Often (More than 10 times in last 4 weeks) (3) Unsure (7) Q438 Now I would like to ask you some questions about the people you talked to about the training or food plate. Q347 What is the name of the person who you talked to about t he FHI360 SHIKHA food plate? Father/Husband's Name (1) Given/First Name (2)

PAGE 210

210 Q417 Gender of the person you talked to Male (1) Female (2) Q351 What district does he/she live in? Mymensingh (1) Borguna (2) Other (3) Answer If What district does he/she live in? Other Is Selected Q421 Specify district and village this person lives in. District (1) Village (2) Answer If What district does he/she live in? Borguna Is Selected Q352 Sushilan Village Baitakata (1) Baitmore (2) Chalavanga (3) Chandra (4) C handra Kapali (5) Ghathkhali (6) Gilatali (7) Holdia (8) Holodia (Paka) (9) Holodia (10) Kalibari (11) Kawnia (12) Kawnia Kapali (13) Loda (14) Patakata (15) Patakata midle (16) Uttar Tokta Bunia (17) Other (18) Answer If Sushilan Village Other Is Selected Q418 Specify name of village if "other" is selected

PAGE 211

211 Answer If What district does he/she live in? Mymensingh Is Selected Q353 BAU Village Boira moddapara sorkar bari (1) Boira Mosque (2) borobilarpar (3) Bot tola bazar (4 ) Bot tola hatem bapari (5) Chor nilokkhai ujanpara (6) Chorkalibari shomvuganj (7) Chornillokhia digolpara (8) Goneshampur (9) Gosta north (10) Jhaugau (11) Jogir ali chor nilokkhia (12) Kismat khagdohor (13) Maijbari (14) Mirjapur east (15) Mirjapur north (16) Mirjapur south (17) Muktijodda bazar chor (18) Mukhtijoddha bazar jogir algi (19) Pagla Bazaar (20) Pagla bazar kazi bari (21) Pagla bazar mojiborer bari (22) Ragobpur (23) Sathiapara (24) Sathiapra chor nillokkhia (25) Shailmari purrush (26) S uhila moddopara (27) Suhila nodir par (28) Suhila north (29) Sutiakhali middle (30) Sutiakhali palpara 1 (31) Sutiakhali palpara 2 (32) Vabokhali west (33) Vabukhali (34) Other (35) Answer If BAU Village Other Is Selected Q419 Specify name of village if "other" is selected.

PAGE 212

212 Q355 What is this person's relationship to you: Friend Female (1) Friend Male (2) Coworker Female (38) Coworker Male (39) Acquaintance Female (40) Acquaintance Male (41) Spouse (19) Mother in law (20) Father in law (21) Cousin Female (3) Cousin Male (4) Aunt (5) Uncle (6) Grandfather (7) Grandmother (8) Father (9) Mother (10) Daughter (11) Son (12) Granddaughter (13) Grandson (14) Sister (15) Brother (16) Niece (17) Nephew (18) Q357 What did you tell this person? I told them I went to a training (1) I told them about the training and shared what I learned with them (2) I showed them the plate (3) I told them about the training, I shared what I learned, and I showed them the plate (4 ) Other (5) Answer If What did you tell this person? Other Is Selected Q420 Specify what you told this person if "other" is selected. Q337 Prior to the BAUEC/Sushilan event had you or anyone in your household seen the FHI360 SHIKHA project food plate? Ye s (1) No (2) Answer If Prior to the event had you or anyone in your household seen the FHI360 Shika project food plate ? Yes Is Selected Q338 Where did you or your family member first see the FHI360 SHIKHA food plate? (Name of NGO or government institution and briefly describe the setting)

PAGE 213

213 Q58 Did your household produce any cereal crops in the last year? (Mark all that apply) Rice (Aus, Aman, and/or Boro) (1) Maize (2) Wheat (3) Other (4) ____________________ Other (5) ____________________ No cereal (6) Q59 Count the number of cereal crops produced in the last year and confirm with the respondent. 0 (1) 1 (6) 2 (7) 3 (8) 4 (9) 5 (10) Q60 Type of cereal Q61 What was the total area planted for ${q://QID155/ChoiceTextEntryValue/1} in the last year? (Decim als) Q62 How much ${q://QID155/ChoiceTextEntryValue/1} did you produce in the last year? Q63 Units of production Kilograms (2) Other (3) ____________________ Q64 How much ${q://QID155/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q458 Price received for sale of ${q://QID155/ChoiceTextEntryValue/1} (BDT per unit)

PAGE 214

214 Q65 Did your household produce any vegetables in the last year? (Mark all that apply) Pumpkin (1) Bitter gourd (2) Bottle gourd (3) Eggplant (21) White potato (4) Sweet pot ato (orange) (5) Tomato (6) Cucumber (7) Red amaranth (8) Amaranth (9) Spinach (10) Cauliflower (11) Cabbage (12) Okra (13) Chili pepper (14) Onion (15) Turnips (16) Taro (17) Other (18) ____________________ Other (19) ____________________ No vegetables (20) Q66 Count the number of vegetable crops produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) 10 (11) 11 (12) 12 (13) 13 (14) 14 (15) 15 (16) 16 (17) 17 (18) 18 ( 19) 19 (20) Q67 Type of vegetable Q68 Is this vegetable mixed crop with other vegetables? Yes (1) No (2) Answer If Is this vegetable mixed crop with other vegetables? Yes Is Selected Q69 In your mixed crop plot of land, is ${q://QID163/ChoiceTextEntryValue} more or less than half of the area planted? More than half (1) About half (2) Less than half (3)

PAGE 215

215 Q70 What was the total area planted for ${q://QID163/ChoiceTextEntryValue/2} in the last year? (Decimals)Note: If the vegetable is mixed crop, calculate and enter the approximate number of decimals allocated to this crop based on the previous question and the total land used for vegetables. Q71 How much ${q://QID163/ChoiceTextEntryValue/2} did you produce in the last year ? Q72 Units of production Kilograms (2) Pieces (3) Q73 How much ${q://QID163/ChoiceTextEntryValue/2} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q459 Price received for sale of ${q://QID163/ChoiceTextEntryValue/2} (BDT per unit) Q74 Did your household produce any pulses in the last year? (Mark all that apply) Lentils (1) Mung bean (2) Khesari (3) Other pulses (4) ____________________ Other pulses (5) ____________________ No pulses (6) Q75 Count the number of pulses produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) Q76 Ty pe of pulse

PAGE 216

216 Q77 What was the total area planted for ${q://QID221/ChoiceTextEntryValue/2} in the last year? (Decimals) Q78 How much ${q://QID221/ChoiceTextEntryValue/2} did you produce in the last year? Q79 Units of production Kilograms (2) Pieces (3) Q80 How much ${q://QID221/ChoiceTextEntryValue/2} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consu mption (2) Product (1) Q460 Price received for sale of ${q://QID221/ChoiceTextEntryValue/2} (BDT per unit) Q81 Did your household produce any fruit in the last year? (Mark all that apply) Mango (1) Jujube (2) Jackfruit (3) Litchi (4) Guava (5) Papaya (6) Coconut (7) Banana (8) Other (9) ____________________ Other (10) ____________________ Other (11) ____________________ No fruits (12)

PAGE 217

217 Q82 Count the number of fruits produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) 10 (11) 11 (12) Q83 Type of fruit Q84 How many ${q://QID231/ChoiceTextEntryValue} trees produced fruit in the last year? Q85 How much ${q://QID231/ChoiceTextEntryValue} did you produce in the last year? Q86 Un its of production Kilograms (2) Pieces (3) Q87 How much ${q://QID231/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to lan d owner (3) Sold (1) Kept for home consumption (2) Product (1) Q461 Price received for sale of ${q://QID231/ChoiceTextEntryValue} (BDT per unit) Q88 Did your household produce any oilseed crops in the last year? (Mark all that apply) Mustard (1) Peanut (2) Soybean (3) Sesame (4) Other (5) ____________________ Other (6) ____________________ Other (7) ____________________ No oilseeds (8)

PAGE 218

218 Q89 Count the number of oilseed crops produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) Q90 Type of oilseed Q91 What was the total area planted for ${q://QID239/ChoiceTextEntryValue} in the last year? (Decimals) Q92 How much ${q://QID239/ChoiceTextEntryValue} did you produce in the last year? Q93 Units of production Kilograms (2) Q94 How much ${q://QID239/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept for home consumption (2) Product (1) Q462 Price received for sale of ${q://QID239/ChoiceTextEntryValue} (BDT per unit) Q95 Did your household produce any fibrous crops in the last year? (Mark all that apply) Jute (1) Kenaf (8) Mesta (2) Other (5) ____________________ Other (6) ____________________ Other (7) ____________________ No fibrous crops (13)

PAGE 219

219 Q96 Count the number of fibrous crops produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 ( 6) 6 (7) Q97 Type of fibrous crop Q98 What was the total area planted for ${q://QID247/ChoiceTextEntryValue} in the last year? (Decimals) Q99 How much ${q://QID247/ChoiceTextEntryValue} did you produce in the last year? Q100 Units of production Kilograms (2) Q101 How much ${q://QID247/ChoiceTextEntryValue} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Paid to land owner (3) Sold (1) Kept fo r home consumption (2) Product (1) Q463 Price received for sale of ${q://QID247/ChoiceTextEntryValue} (BDT per unit)

PAGE 220

220 Q102 Did your household raise any livestock for production in the last year? If yes, what animal products did you produce? (Mark all that apply) Cow milk (1) Cow meat (beef) (13) Chicken meat (5) Chicken eggs (14) Duck meat (15) Duck eggs (6) Goat milk (7) Goat meat (8) Pigeon meat (9) Other (10) ____________________ Other (4) ____________________ Other (11) ____________________ No livestock (12) Q103 Count the number of animal products the household produced in the last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) Q104 Type of animal product Note: Be specific about animal type. If milk, specify cow or goat. If eggs, specify chicken or duck. Animal product (1) Q105 How many animals did you have for producing ${q://QID169/ChoiceTextEntryValue/1} in the last year? Q106 How much ${q://QID169/ChoiceTextEntryValue/1} did you produce in the last year? Q107 Units of production Kilograms (1) Liters (2) Pieces (3)

PAGE 221

221 Q108 How much ${q://QID169/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given above for total production. Sold (1) Kept for home consumption (2) Product (1) Q464 Price received for sale of ${q://QID169/ChoiceTextEntryValue/1}(BDT per unit) Q109 Did your household cultivate any fish last year? (Mark all that apply) Rui (23) Katla (25) Carp (26) Tilapia (27) Koi (28) Mola (29) Shrimp (30) Prawn (31) Other (32) ____________________ Other (33) ____________________ Other (34) ____________________ No fish (24) Q110 Count the number of fish species cultivated in th e last year and confirm with the respondent. 0 (1) 1 (2) 2 (3) 3 (4) 4 (5) 5 (6) 6 (7) 7 (8) 8 (9) 9 (10) 10 (11) 11 (12) Q111 Type of fish Species (1)

PAGE 222

222 Q441 How many fish ponds do you use to cultivate ${q://QID187/ChoiceTextEntryValue/1}? Q443 What is the total land area of the fish ponds used to cultivate ${q://QID187/ChoiceTextEntryValue/1}? Q442 Is this pond used for polyculture? (Many species in one pond) Yes (1) No (2) Q112 How much ${q://QID187/ChoiceTextEntryValue/1} did you produce in the la st year? Q113 Units of production Kilograms (2) Pieces (3) Q114 How much ${q://QID187/ChoiceTextEntryValue/1} did you sell and how much did you keep for home consumption? Note: Verify that the total does not exceed the answer given in the previous question about total production. Sold (1) Kept for home consumption (2) Product (1) Q465 Price received for sale of ${q://QID187/ChoiceTextEntryValue/1}(BDT per unit) Q142 Now I have some questions for you about food purchases and agriculture markets. Q143 Are you the person who makes the final decision about food purchases and preparation? Yes (1) No (2)

PAGE 223

223 Answer If Are you the person who makes the final decision about food purchases and preparation? No Is Selected Q144 Who makes the final decision regarding food purchases and preparation? Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q145 If you had more income, would you spend more of your money on food? Yes (1) N o (2) Q146 If you had more money available to spend on food would you consume the same types of food? Yes (1) No (2) Q147 If you had more money available to spend on food, which of the following would you consume more of? (Mark all that apply) Cereals ( 1) Pulses (2) Vegetables (3) Fruits (4) Meat (5) Eggs (6) Fish (7) Milk or milk products (8) Sweets (9) Other foods (10) ____________________ I would not spend more money on food. (11) Q148 Did anyone in your household buy any food (from a market) to cook in the household in the last year? Yes (1) No (2)

PAGE 224

224 Answer If Did anyone in your household buy any food (from a market) to cook in the household in the last year? Yes Is Selected Q149 How long does it take to walk to a place to buy food? Less than 30 min utes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Answer If Did anyone in your household buy any food (from a market) to cook in the household in the last year? Yes Is Selected Q150 What mode of transportation does your household us e to go to the market for buying food? By foot (1) By bicycle (2) By rickshaw/van (3) By car/truck (4) By motorcycle (5) By boat (6) Other (7) ____________________ Answer If Did anyone in your household buy any food (from a market) to cook in the household in the last ye... Yes Is Selected Q151 How much does this transportation cost per trip to the market? (BDT) Q152 Does anyone in your household ever sell commercial agricultural products grown in your household (e.g. rice, maize)? Yes (1) No (2 ) Answer If Does anyone in your household ever sell commercial agricultural products grown in your household (e.g. rice, maize)? Yes Is Selected Q153 How long does it take to walk to the place to sell commercial agricultural products, for example to a mar ket or buyer pick up location? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 225

225 Answer If Does anyone in your household ever sell commercial agricultural products grown in your ho usehold... Yes Is Selected Q154 Who controls the income from sales of commercial agricultural products? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q155 Does anyone in yo ur household ever sell vegetables or fruits grown in your household? (e.g. pumpkins, cucumbers, mangos) Yes (1) No (2) Answer If Does anyone in your household ever sell vegetables or fruits grown in your household? (e.g. pumpkins, cucumbers, mangos) Yes I s Selected Q156 How long does it take to walk to the place to sell vegetables or fruits? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 226

226 Answer If Does anyone in your household ever sell vegetables or fruits grown in your household? (e.g. pumpk... Yes Is Selected Q157 Who controls the income from the sales of vegetables/fruits? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. s ervant) (17) Q158 Does anyone in your household ever sell fish produced in your household? Yes (1) No (2) Answer If Does anyone in your household ever sell fish produced in your household? Yes Is Selected Q159 How long does it take to walk to the place to sell fish? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 227

227 Answer If Does anyone in your household ever sell fish produced in your household? Yes Is Selected Q440 Who controls the income from the sale of fish? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q160 Does anyone in your household ever sell animal protein (meat) produced in your household? Yes (1) No (2) Answer If Does anyone in your household ever sell animal protein (meat) produced in your household? Yes Is Selected Q161 How long does it take to walk to the place to sell animal protein (meat)? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 228

228 Answer If Does anyone in your household ever sell animal protein (meat) produced in your household? Yes Is Selected Q162 Who controls the incom e from the animal protein (meat)? Mark all that apply. Head of household (1) Wife of household head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) Q163 Does anyone in your household ever sell eggs produced in your household? Yes ( 1) No (3) Answer If Does anyone in your household ever sell eggs produced in your household? Yes Is Selected Q164 How long does it take to walk to the place to sell eggs? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 229

229 Answer If Does anyone in your household ever sell eggs produced in your household? Yes Is Selected Q165 Who controls the income from the sale of eggs? Mark all that apply. Head of household (1) Wife of h ousehold head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) ____________________ Q166 Does anyone in your household ever sell milk or milk products produced in the household? Yes (1) No (2) Answer If Does anyone in your household ever sell milk or milk products produced in the household? Yes Is Selected Q167 How long does it take to walk to the place to sell milk or milk products? Less than 30 minutes (1) 30 minutes to 1 hour (2) 1 to 2 hours (3) More than 2 hours (4) Sell at the household (5)

PAGE 230

230 Answer If Does anyone in your household ever sell milk or milk products produced in the household? Yes Is Selected Q168 Who controls the income from the sale of milk or milk products? Mark all that apply. Head of household (1) Wife of ho usehold head (2) Husband of household head (3) Son (4) Daughter (5) Father (6) Mother (7) Daughter in law (8) Son in law (9) Brother (10) Sister (11) Father in law/Mother in law (12) Nephew/Niece (13) Grandfather/Grandmother (14) Sister in law/Brother in law (15) Brother's wife (16) Other (e.g. servant) (17) ____________________ Answer If Does anyone in your household ever sell commercial agricultural products grown in your household... Yes Is Selected Or Does anyone i n your household ever sell vegetables or fruits grown in your household? (e.g. pumpk... Yes Is Selected Or Does anyone in your household ever sell fish produced in your household? Yes Is Selected Or Does anyone in your household ever sell animal protein ( meat) produced in your household? Yes Is Selected Or Does anyone in your household ever sell milk or milk products produced in the household? Yes Is Selected Or Does anyone in your household ever sell eggs produced in your household? Yes Is Selected Q169 W hat mode of transportation does your household use to transport agricultural goods, fish, or vegetables/fruits to the market/selling points? By foot (1) By bicycle (2) By rickshaw/van (3) By car/truck (4) By motorcycle (5) By boat (6) Other (7) ____________________ Q439 How much is the transportation cost per trip to the market to sell products? (BDT) Q170 Were there months, in the past 12 months, in which you did not have enough food to meet your family's needs? Yes (1) No (2)

PAGE 231

231 Answer If Were there months, in the past 12 months, in which you did not have enough food to meet your family's needs? Yes Is Selected Q171 During which months did you not have enough food to meet your family's needs? (Check all that apply) January (1) February (2) March (3) April (4) May (5) June (6) July (7) August (8) September (9) October (10) November (11) December (12) Q172 In the past four weeks, did you worry that your household would not have enough food? Yes (1) No (2) Answer If In the past four weeks, did you worry that your household would not have enough food? Yes Is Selected Q173 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q174 In the past four weeks were you or any household member not able to eat the kinds of food you preferred because of a lack of resources to get food? Yes (1) No (2) Answer If In the past four weeks were you or any household member not able to eat the kinds of food you preferred because of a lack of resources to get food? Yes Is Selected Q175 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q176 In the past four weeks, did you or any household member have to eat some foods that you really did not want to eat because of a lack of resources to obtain other types of food? Yes (1) No (2)

PAGE 232

232 Answer If In the past four weeks, did you or any household member have to eat som e foods that you really did not want to eat because of a lack of resources to obtain other types of food? Yes Is Selected Q177 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 week s) (3) Q178 In the past four weeks, did you or any household member have to eat a smaller meal than you felt you needed because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member have to eat a smaller meal than you felt you needed because there was not enough food? Yes Is Selected Q179 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q180 In the past four wee ks, did you or any household member have to eat fewer meals in a day because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member have to eat fewer meals in a day because there was not enough food? Y es Is Selected Q181 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q182 In the past four weeks, was there ever no food of any kind in your household because of a lack of resources to get food? Yes (1) No (2) Answer If In the past four weeks, was there ever no food of any kind in your household because of a lack of resources to get food? Yes Is Selected Q183 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3)

PAGE 233

233 Q184 In the past four weeks, did you or any household member go to sleep at night hungry because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member go to sleep at night hungry because there was not enough food? Yes Is Selected Q185 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q186 In the past four weeks, did you or any household member go a whole day and night without eating anything because there was not enough food? Yes (1) No (2) Answer If In the past four weeks, did you or any household member go a whole day and night without eat ing anything because there was not enough food? Yes Is Selected Q187 How often did this happen? Rarely (1) Sometimes (3 4 times in the last 4 weeks) (2) Often (More than 10 times in the last 4 weeks) (3) Q188 In the past 12 months, how often did you or a ny of your family have to eat potato, wheat, or another grain although you wanted to eat rice (not including when you were sick)? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5) Q189 In the past 12 months, how often did you or any of your family skip entire meals due to scarcity of food? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in t he last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5)

PAGE 234

234 Q190 In the past 12 months, how often did you personally eat less food in a meal due to scarcity of food? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5) Q191 In the past 12 months, how often did your family purchase food (rice, lentils, etc.) on credit or loan f rom a local shop? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every day) (5) Q192 In the past 12 months, how often did your family have to borrow/take food from relatives or neighbors to make a meal? Never (1) Rarely (1 6 times in the last 12 months) (2) Sometimes (7 12 times in the last 12 months) (3) Often (a few times in each month) (4) Regularly (every day or almost every d ay) (5)

PAGE 235

235 LIST OF REFERENCES Ahmed, A.U., Ahmad, K., Chou, V., Hernandez, R., Menon, P., Naeem, F., Naher, F., Quabili, W., Sraboni, E., and Yu B 2013. The Status of Food Security in the Feed the Future Zone and Other Regions of Bangladesh: Resul ts from the 20112012 Bangladesh Integrated Household Survey. International Food Policy Research Institute (IFPRI), Bangladesh Policy Research and Strategy Support Program. http://ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/127518. [ Accessed Nov. 1, 2016] Arimond, M and Ruel M. 2004. Dietary Diversity Is Associated with Child Nutritional Status: Evidence from 11 Demographic and Health Surveys. Journal of Nutrition, 134: 25792585. Alderman, H., 2007. Improving nutrition through community growth promotion: longitudinal study of the nutrition and early child development program in Uganda. World development 35(8) : 13761389. Azzarri, C. Zezza, A., Haile, B., and E. Cross. 2015. Does livestock ownership affect animal source foods consumption and child nutritional status? Evidence from rural Uganda. The Journal of Development Studies 51(8), pp.1034 1059. Ball, K., Crawford, D. and Mishra, G 2006. Socio economic inequalities in women's fruit and vegetable intakes: a multilevel study of individual, social and environmental mediators. Public health nutrition, 9(05) : 623630. Banerjee, Abhijit V. and Duflo E 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty N ew York: Public Affairs Bangladesh Bureau of Statistics (BBS). 2013. District Statistics 2011 Mymensingh District. BBS Statistics and Information Division (SID), Ministry of Planning, Government of the Peoples Republic of Bangladesh. Barnum, H.N. and Squ ire, L. 1979. An econometric application of the theory of the farm household. Journal of Development Economics 6(1) : 79102. Barrett, C.B. 2008. Smallholder market participation: Concepts and evidence from eastern and southern Africa. Food policy 33(4) : 299317. Becker, G.S. 1965. A Theory of the Allocation of Time. The Economic Journal 75(299): 493 517. Behrman, J.R. and Deolalikar, A.B. 1988. Health and nutrition. Handbook of Development Economics 1, 631711. Belli, P.C., Bustreo, F., and Preker A 2 005. Investing in childrens health: what are the economic benefits? Bulletin of the World Health Organization, 83(10): 777784.

PAGE 236

236 Brown, O.N., OConnor, L.E. and Savaiano D 2014. Mobile MyPlate: A Pilot Study Using Text Messaging to Provide Nutrition Education and Promote Better Dietary Choices in College Students. Journal of American College Health, 62(5): 320327. Buzby, J.C. and Guthrie, J.F. 2002. Plate Waste in School Nutrition Programs: Final Report to Congress. Washington, D.C.: Economic Researc h Service. United States Department of Agriculture, March 2002. Report No. ERS E FAN 02009. Cameron, A.C. and Miller D 2014. A Practitioners Guide to Cluster Robust Inference. The Journal of Human Resources, 50(2) : 317372. Cameron, A.C., Gelbach, J., and Miller D. 2008. Bootstrapbased improvements for inference with clustered errors. Review of Economics and Statistics 90(3) : 414427. Campbell, K.J., Abbott, G., Spence, A.C., Crawford, D.A., McNaughton, S.A., and Ball K 2013. Home food availabilit y mediates associations between mothers nutrition knowledge and child diet. Appetite 71 : 16. Carletto G., Ruel, M., Winters, P., and Zezza, A. 2015. Farm Level Pathways to Improved Nutritional Status: Introduction to the Special Issue. The Journal of D evelopment Studies 51(8) : 954 957. Chege, C.G., Andersson, C.I. and Qaim, M., 2015. Impacts of supermarkets on farm household nutrition in Kenya. World Development 72: 394407. Coates, J., Swindale, A., and Bilinsky P 2007. Household Food Insecurity Access Scale (HFIAS) for Measurement of Food Access: Indicator Guide (v.3). Washington, D.C.: FHI360/FANTA. Comstock, E.M., St. Pierre, R.G., and Mackiernan, Y.D. 1981. Measuring individual plate waste in school lunches. Journal of the American Dietetic Association, 79(3): 290 296. Chung, K. 2012. An Introduction to Nutrition Agriculture Linkages. MINAG/DE Research Report 72E. Maputo, Mozambique: Directorate of Economics, Ministry of Agriculture. Dillon, A., K. McGee, and Oseni G 2015. Agricultural Production, Dietary Diversity, and Climate Variability. The Journal of Development Studies 51(8) : 976995. Duflo, E., Glennerster, R., and Kremer M. 2008. Using Randomization in Development Economics Research: A Toolkit . T. Schultz and John Strauss, eds., Handbook of Development Economics Vol. 4. Amsterdam and New York: North Holland, 4. FAO, 2016. Food based dietary guidelines. http://www.fao.org/nutrition/education/fooddietary guidelines/home/en/ [Accessed June 1, 2016] FAO and WFP. 2012. Household Dietary Diversity Score and Food Consumption Score: A Joint Statement of FAO and WFP. http://www.f ao.org/docrep/meeting/024/mc147e.pdf [Accessed June 16, 2016]

PAGE 237

237 FAO, IFAD, and WFP. 2015. The State of Food Insecurity in the World 2015. Meeting the 2015 international hunger targets: taking stock of uneven progress. Rome, FAO. FHI360/ USAID 2015. Designi ng a Food Plate for Dietary Counseling of Pregnant Women. FHI360/USAID SHIKHA. https://www.fhi360.org/resource/designingfood plate dietary counselingpregnant women [Accessed October 24, 2016]. FHI360/ USAID 2016. Ministry of Health and Family Welfare En dorses Food Plate as a Counseling Tool for Pregnant and Lactating Women. FHI360/USAID SHIKHA. https://www.fhi360.org/resource/ministry health and family welfare endorses food plate counseling tool pregnant and [Accessed October 24, 2016]. Fitzsimons, E., Malde B. Mesnard A. and VeraHernandez M. 2016. Nutrition, information, and household behavior: Experimental evidence from Malawi. Journal of Development Economics 122(2016): 113126. Friedman, B.J. and R.D. Hurd Crixell. 1999. Nutrient intake of children eating school breakfast. Journal of the American Dietetic Association, 99(2): 219221. Haider, R., Ashworth, A., Kabir, L., Huttly, S. 2000. Effect of community based peer counselors o n exclusive breastfeeding practices in Dhaka, Bangladesh: a randomised controlled trial. The Lancet 356(9242): 16431647. Hanks, A. A., Wansink, B., and Just, D.R 2014. Reliability and Accuracy of Real Time Visualization Techniques for Measuring School C afeteria Tray Waste: Validating the Quarter Waste Method. Journal of the Academy of Nutrition and Dietetics 114(3): 470474. Hanks, A. S., Just, D.R. Smith, L.E. and Wansink B. 2012. Healthy convenience: nudging students toward healthier choices in the lunchroom. Journal of Public Health, 34(3): 370376. Hatloy, A., Torheim L.E. and Oshaug A. 1998. Food variety a good indicator of nutritional adequacy of the diet? A case study from an urban area in Mali, West Africa. European Journal of Clinical Nutrition 52, 891898. Henderson, J. 2016. Integrating Gender and Nutrition within Agricultural Extension Services: A Facilitators Guide. USAID: Feed the Future, INGENAES. http://ingenaes.illinois.edu/wp content/uploads/Gender and Nutrition Facilitators Guide.pdf [Accessed July 29, 2016]. Hossain, M., Naher, F. and Shahabuddin, Q. 2005. Food security and nutrition in Bangladesh: progress and determinants. Electronic Journal of Agricultural and Development Economics 2(2 ): 103132. Iannotti, L., Cunningha m, K., and Ruel, M. 2009. Improving Dietary Quality and Micronutrient Nutrition: Homestead Food Production in Bangladesh. Washington, D.C., IFPRI Discussion Paper No. 00928.

PAGE 238

238 IFPRI (International Food Policy Research Institute). 2011. Leveraging Agriculture for Improving Nutrition and Health: Highlights from an International Conference. Washington, DC. Jones, A.D., Shrinivas, A. and Bezner Kerr R 2014. Farm production diversity is associated with greater household dietary diversity in Malawi: Findings fr om nationally representative data. Food Policy 46(2014) : 112. Just, D. R., Mancino, L., and Wansink, B. 2007. Could Behavioral Economics Help Improve Diet Quality for Nutrition Assistance Program Participants? In: United States Department of Agriculture, Economic Research Service Report Number 43 (eds). Keding, G.B., Msuya, J.M., Maass, B.L. and Krawinkel, M.B. 2012. Relating dietary diversity and food variety scores to vegetable production and socio economic status of women in rural Tanzania. Food Securi ty 4(1) : 129 140. Kennedy, G., Ballard, T. and Dop M 2010. Guidelines for Measuring Household and Individual Dietary Diversity. Nutrition and Consumer Protection Division, Food and Agriculture Organization of the United Nations. Rome, Italy. Kongsbak, I., Skov, L.R., Nielsen, B.K., Ahlmann, F.K., Schaldemose, H., Atkinson, L., Wic hmann, M. and PrezCueto, F.J. 2016. Increasing fruit and vegetable intake among male university students in an ad libitum buffet setting: A choice architectural nudge intervention. Food Quality and Preference 49: 183 188. Kropp J.D., Abarca Orozco, S.J., Israel, G.D., Diehl, D.C., G alindo Gonzalez, S., Headrick, L.B., and Shelnutt, K.P. 2017. A Plate Waste Evaluation of the Farm to School Program. Journal of Nutrition Education and Behavior: article in press. Kumar, N., Harris, J., and Rawat R 2015. If They Grow It, Will They Eat a nd Grow? Evidence from Zambia on Agricultural Diversity and Child Undernutrition. The Journal of Development Studies, 51(8) : 10601077. Leak, T.M., Swenson, A., Vickers, Z., Mann, T., Mykerezi, E., Redden, J.P., Rendahl, A. and Reicks, M. 2015. Testing t he effectiveness of in home behavioral economics strategies to increase vegetable intake, liking, and variety among children residing in households that receive food assistance. Journal of nutrition education and behavior 47(2) : e1 e9. Leak, T.M., Swenson A., Rendahl, A., Vickers, Z., Mykerezi, E., Redden, J.P., Mann, T., and Reicks, M. Examining the feasibility of implementing behavioural economics strategies that encourage home dinner vegetable intake among low income children. Public Health and Nutriti on, 20(8): 13881392. Linnemayr, S. and Alderman H 2011. Almost random: Evaluating a large scale randomized nutrition program in the presence of crossover. Journal of Development Economics 96(2011) : 106 114.

PAGE 239

239 Mackinnon, J.G. and Webb, M.D. 2017. Wild B ootstrap Inference for Wildly Different Cluster Sizes. Journal of Applied Econometrics, 32(2) : 233 254. Micronutrient Initiative/World Bank/UNICEF. 2009. Investing in the Future: A United Call to Action on Vitamin and Mineral Deficiencies: Global Health Report. Micronutrient Initiative, Toronto, ON, Canada. http://www.unitedcalltoaction.org/documents/Investing_in_the_future.pdf [ Accessed May 10, 2017] Miller, G.F., Gupta, S., Kropp, J.D., Grogan, K.A. and Mathews, A. 2016. The effects of pre ordering an d behavioral nudges on National School Lunch Program participants food item selection. Journal of Economic Psychology 55(2016): 4 16. Muller, C. 2009. Do agricultural outputs of partly autarkic peasants affect their health and nutrition? Evidence from Rw anda. Food Policy 34(2009): 166 175. Parmenter, K. and Wardle, J. 1999. Development of a general nutrition knowledge questionnaire for adults. European Journal of Clinical Nutrition, 53(4) : 298308. Rabbani, A. 2014. Household Food Security in Bangladesh : Going beyond Poverty Measures. Bangladesh Development Studies Vol. XXXVII (1,2) : 103 125. Rah, J.H., Akhter, N., Semba, R.D., de Pee, S., Bloem, M.W., Campbell, A.A., Moench Pfanner, R., Sun, K., Badham, J., and Kraemer K 2010. Low dietary diversity i s a predictor of child stunting in rural Bangladesh. European Journal of Clinical Nutrition, 64: 13931398. Rashid, D.A., Smith, L.C., and Rahman, T. 2011. Determinants of dietary quality: evidence from Bangladesh. World Development 39(12): 22212231. Raw lins, R., Pimkina, S., Barrett, C.B., Pedersen, S., and Wydick, B. 2014. Got milk? The impact of Heifer Internationals livestock donation programs in Rwanda on nutritional outcomes. Food Policy 44(2014): 202213. Richter, S.L., Vandervet, L.M., Macaskill, L.A., Salvadori, M.I, Seabrook, J.A. and Dworatzek, P.D.N. 2012. Journal of the Academy of Nutrition and Dietetics 112(10): 16031607. Roy, S.K., Jolly, S.P., Shafique, S., Fuchs, G.J., Mahmud, Z., Chakraborty, B., and Roy, S. 2007. Prevention of malnutrition among young children in rural Bangladesh by a food health care educational intervention: a randomized, controlled trial. Food and nutrition bulletin, 28(4) : 375 383. Ruel, M., Alderman H ., and t he Maternal an d Child Nutrition Study Group. 2013. Nutrition sensitive interventions and programmes: how can they help to accelerate progress in improving maternal and child nutrition? The Lancet 382: 536551. Schreiner, M. 2013 Simple Poverty ScorecardTM: Bangladesh. microfinance. com/English/Papers/Scoring_Poverty_Bangladesh_2010_EN. pdf, retrieved, 4.

PAGE 240

240 Shatenstein, B., Claveau, D., and Ferland, G. 2002. Visual observation is a valid means of assessing dietary consumption among older adults with cognitive deficits in long term care settings. Journal of the American Dietetic Association 102(2): 250252. Sibhatu, K.T., Krishna, V.V. and Qaim M. 2015. Production diversity and dietary diversity in smallholder farm households. PNAS 112(34), 1065710662. Singh, I., Squir e, L. and Strauss, J., 1986. Agricultural household models: Extensions, applications, and policy The World Bank. Thaler, R.H. and Sunstein, C.R. 2009. Nudge: Improving Decisions about Health, Wealth, and Happiness Penguin Books. Thompson, B. and Amoroso L 2011. FAOs Approach to NutritionSensitive Agricultural Development ICN2 Second International Conference on Nutrition. FAO and WHO. Thorndike, A .N., Riis, J., Sonnenberg, L.M., and Levy, D. E. 2014. Traffic light labels and choice architecture: pr omoting healthy food choices. American Journal of Preventive Medicine 46(2) : 143149. Torheim, L.E. Ouattara, F., Diarra, M.M., Thiam, F.D., Barikmo, I., Hatloy, A., and Oshaug, A. 2004. Nutrient adequacy and dietary diversity in rural Mali: association and determinants. European Journal of Clinical Nutrition, 54: 594604. Townsend, Robert. 2015. Ending poverty and hunger by 2030: an agenda for the global food system Washington, D.C. : World Bank Group. http://documents.worldbank.org/curated/en/2015/04/24367067/endingpoverty hunger 2030agendaglobalfood system [ Accessed April 1 2016] United States Department of Agriculture (USDA). 2016. ChooseMyPlate.gov. http://www.choosemyplate.gov van Ansem, W.J., Schrijvers, C.T., Rodenburg, G., and van de Mheen, D. 2014. Maternal educational level and childrens healthy eating behaviour: role of the home food environment (cross sectional results from the INPACT study). International Journal of Behavioral Nutrition and Physical Activity 11(1) : 1. Wooldridge, J.M. 2004. Cluster sample methods in applied econometrics. American Economic Review 93(2) : 133138. Wong, H.L., Shi Y. Luo, R., Zhang, L., and Rozelle, S. 2014. Improving the Health and Education of Elementary Schoolchildren in Rural China: Iron Supplementation Versus Nutritional Training for Parents. The Journal of Development Studies 50(4) : 509519. World Bank. 2006. Repositioning nutrition as central to development: A strategy for large scale action. Washington DC.

PAGE 241

241 World Bank. 2011. South Asia Food and Nutrition Security Initiative. http://go.worldbank.org/WHNECFL7D0 [ Accessed January 4, 2016] World Bank. 2013. Bangladesh Nutrition at a glance Nutrition at a glance ; Bangladesh. Washington DC; World Bank. http://documents.worldbank.org/curated/en/2013/04/17695190/bangladesh nutrition glance [Accessed January 5 2016] World Bank. 2016a In Madagascar, safet y net programs promote nutrition, early childhood development, support productive activities of the poor. The World Bank Group. http://www.worldbank.org/en/results/2016/09/19/in madagascar safety net programs promote nutrition early childhood development s upports productive activities of the poor [Accessed September 1, 2016]. World Bank. 2016b World DataBank Health and Nutrition Population Statistics. The World Bank Group. http://databank.worldbank.org/data/reports.aspx?source=health nutrition and population statistics [ Accessed January 4, 2016] World Food Programme (WFP) 2008. Food Consumption Analysis: Calculation and Use of the Food Con sumption Score in Food Security Analysis. United Nations World Food Programme, Vulnerability Analysis and Mapping Branch (ODAV). Rome, Italy.

PAGE 242

242 BIOGRAPHICAL SKETCH Kelly A. Davidson grew up on a grain and cattle farm in Nevada, Ohio and moved to Stamping Ground, Kentucky during high school. She completed her Bachelor of Science degree at the University of Kentucky in 2008, majoring in a gricultural e conomics and f oreign l angua ge i nt er national e conomics, and minoring in Fr ench As an undergraduate, Kelly studied abroad in France She also conducted international fieldwork in the Republic of Georgia, collaborating on a project for higher education in agriculture. Furthermore, s he served as a short term consultant for the World Bank Commodity Risk Management Group and a research associate at GlobalAgRisk, Inc. Kelly received a Master of Science degree in a gricultural economics at the University of Kentucky in 2009. Kelly worked as a f ishe ries e conomist for the NOAA National Marine Fisheries Service, Pacific Islands Fisheries Science Center from 2009 to 2011 She developed an interest in aquaculture economics and served as faculty for the University of Hawaii Aquaculture Training and Online Learning course on business and marketing. Inspired to pursue further teaching experience, Kelly returned to the mainland as a lecturer in agribusiness at the Un iversity of Tennessee at Martin. She taught a variety of courses in agribusiness, economics, and policy. Her enduring passion for agricultural economics and international development motivated her to return to graduate school. Kelly began her doctoral studies in f ood and r esource e conomics at the University of Florida in 2013. With support from t he USAID Feed the Future initiative Integrating Gender and Nutrition within Agricultural Extension Services (INGENAES), Kelly traveled to Bangladesh to pilot the INGENAES technology assessment toolkit. Conversations with fish farmers and development agen cies spurred her interest in nutrition initiatives for rural households. Specifically, she decided to measure the effectiveness of mechanisms that were

PAGE 243

243 being used to promote nutrition. Inspired by the previous work of her advisor, Dr. Jaclyn Kropp, Kelly p ursued this dissertation work, graciously funded by INGENAES. After graduation, Kelly plans to continue her research on behavioral economics in food and agricultural policy. She is specifically interested in analyzing polic ies that impact low income households, both domestically and internationally. She also hopes to inspire the next generation of agricultural economists through future research and teaching endeavors