Consumer product class knowledge

MISSING IMAGE

Material Information

Title:
Consumer product class knowledge a typology and impact on information search and evaluation
Physical Description:
x, 307 leaves : ill. ; 28 cm.
Language:
English
Creator:
Assar, Amardeep, 1952-
Publisher:
s.n.
Publication Date:

Subjects

Subjects / Keywords:
Consumers' preferences   ( lcsh )
Consumers -- Attitudes   ( lcsh )
Marketing thesis Ph. D
Dissertations, Academic -- Marketing -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1987.
Bibliography:
Bibliography: leaves 298-306.
Statement of Responsibility:
by Amardeep Assar.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001031510
oclc - 18157814
notis - AFB3684
System ID:
AA00002151:00001

Full Text














(XONSUtIER POIUT CLASS KNOWLEDGE :
A TYPOLOGY AND IMPACT ON
INFtRMATION SEARCH
AND EVALUATION






By


AMARDEEP


ASSAR


A DISSERTATION PRESENTED TO THE (GRADUATE SCHOOL
OF THE UNIVERSITY OF FLRIDA IN
PARTIAL FULFILLMENT OF THE R~fEQIEMh1'S
lR THE DEGREE OF DOCLEO OF PHILOSHY



UNIVERSITY OF FLORIDA

1987






















































Copyright


1987


Amardeep Assar



















To Nandini,


Vijith and Viraj

















wish


thank


several


people


their


help


in doing


this


research.


Very


special


thanks


to Dr


. Dipankar Chakravarti


for all


effort during


this project


He helped me


immeasurably to


think


in a


more


disciplined


manner,


to write with clarity


His guidance and


comittnment


excellence


teacher


researcher


were


inspiration.


Thanks


are


also


to Dr


John Lynch,


Gordon Bechtel and Dr


Horowitz


their


insights.


Finally,


would like


to thank Dr


Manoj


Agarwal


of S.U.N.Y.


Binghamton


for his assistance


in writing the


computer program for the study


of correlation


information.


ACNOWLEDEMEMS
















TABLE OF CONmEITS

PAGE




AIBSTRACT. ................... ........ iX


Knowledge Effects in the Consumer
Knowledge Research in Marketing..
Knowledge Research in Psychology.
Knowledge Research in Economics..
Secondary Effects of Knowledge...
Overview of the Dissertation.....


Approaches


Environment..................... 1
. . . . . . . . .3
. . . . . . . . .9
. . . . . . .. . . .9
1 $@@@ 60@0 @@@ @@@@ @@@
1)()1(( 1)))))()11 @ @@@@$@@@@@ @@@@@@@@@


to Studying Consumer Knowledge


Prior Definitions and Measures.........
Aggregate Product Familiarity: Effects
Evaluation and Choice...............
Aggregate Knowledge: Effects on Search,
and Choice..........................
Aggregate Product Familiarity Studies:


Disaggregate Studi
Evaluation and
Disaggregate Studi
Moderating Effects
Moderating Effects


Overview of


..... Search
on Search


..... .......o.............19
Evaluation
..........Secondary ............. .25
Secondary Effects.........28


es of Knowledge: Effects on Search,
Choice.........................................32
es of Knowledge: Secondary Effects.............38
of Knowledge Structure........................39
of Other Background Factors...................41


a Typology of Product Class Knowledge..


Components Of the Product Class
Terminology.................
Brand-Attribute Values......
Attribute Range Knowledge...
A++vr41-u, n ('nncl m+4- n


Knowledge Typology...............51
................ . . . ...51
. .. *.. ... ....... .. .. .. ..52


. o......o.. ............ 15


.................17


LCT DUCTION...................................................... 1


LITERATURE REVIEW................................................14


CONCEPTUAL FRAMEMKKE..........................


..............45













Distinctions from Other Knowledge
Functions of Knowledge in Consimer

KN(WLEXIE OF ATIRIBUE RANGE IN A
ON EVALUATION AND SEARCH.......


Typologies..............
SInformation Processing.


PROrUCT C


S.......61
.......62


ATBOCRY: EFFECTS
....... . . .....65


Social Judgement Research on Range Ef
Economics Research on Range Effects..
Marketing Research on Range Effects..
Range Knowledge in Sequential Search.
Attribute Range Knowledge: Effects on
Issues for Research..................


fects.


.a....
Evalu


S g . . . . .66
. . . . .. .69
. .. . . . ...71
.. .. ... .. ..... .. 73
nation and Search......74
......................80


AN EMPIRICAL INVESTIGATION OF A RIBTERANGE N
EFFECTS: CONCEPIUAL HYPOHES AND STUDY DESIN...............84


Range Effects on Provisional Evaluations: A Proposed
The Evaluation Mechanism........................
Search Patterns Contingent on Overall Evaluation.


L Mechanism.


..85


Conceptual Hypothese
Effects on Evalua
Effects on Search
Secondary Effects
Special Cases: At
Expectations..
Experimental Method.
Subjects.........
Stimuli..........
OvYY tf^HTW


The range know
The brand-attr
The search tasi
The evaluation
Task evaluation


sti.
tions....


on Confidence
tribute Values


in Evaluations.
Violating Range


S.
..
..
S*
gt


. ...... .... .85
.. ... ... .....88
.. .... ......91
.............91
.............94
.............97


S.. .. .. .. .. . .. .. . .. ..... .. .. .98
...................... ......................101
. . .... .. .. ... ......... 102
............. . . . . . . ....102

ledge manipulation.........................103
ibute value manipulation....................104
k. ..........................................104
task and confidence judgements.............105
ns and manipulation checks................. 106


AN PMPIRICAL INVESTIGATION OF ATTRIBUTE RANGE KNOWLEDGE
EFFEITS: FINDINGS AND DISCUESION............................. 109


Preliminary Validity Checks.........
Task Perception Measures........


Career
Tests of
Tests of
Tests of
Tfent of


a


Ownership and Knowledge of Relevant Attributes


" the Evaluation Hypotheses....
the Search Hypotheses........
the Confidence Hypotheses....
SHnvmthtapna Roaardlint vi Rvnwta


Rlnnda V inla lti e


........110
........110
........111
........113
........121
....... .132


JL &


..... ... @.... .













KNOWLEDGE OF INTER-ATrRIBUlE (XORELATIONS IN A PRODUCT
CATEGORY: EFFECTS ON SEARCH AND EVALUATION........ ........148


The Price-Quality Relationship...............................
ConsLmer Perceptions of the Price-Quality Relationship....
Objective Evidence Regarding the Price-Quality Relationshi
Correlational Knowledge and Inferences.......................
Correlational Knowledge and Infornmation Search...............
Correlational Knowledge in Consumer Decisions:
A Proposed Mechanism ........................* .............
The Role of Background Factors............................
S m ary. ..... ............. ... ... ............ ...........


...1
p..1

...I


...161
...165
...166


AN EMPIRICAL INVESTIGATION OF THE EFFECTS OF INTER-AfIRIBUTE
CORRELATION KNOWLEDGE: CONCEPTUAL HYPOHESS AND S1IUDY DESIGN. 167


Inter-Attribute Correlations:


in a Consumer


Decision Scenario......................
The Role of Background Factors.......
Inter-Attribute Correlational Knowledge
Product Evaluations.................
Inter-Attribute Correlational Knowledge


Brand Choice................
Secondary Effects on Confidence
Conceptual Hypotheses.............
Effects on Search..............
Effects on Evaluations.........
Effects on Choices.............
Secondary Effects on Confidence
Design of the Proposed Study......
The Study Setting..............
The Stimuli....................
Key Manipulations..............
Correlational knowledge.....
The involvement manipulation
The time pressure manipulatiu
Subjects... .................


Experimental
Operational
Expected
Tests of
Tests of
Tests of
Tests of


Procedures..........
Hypotheses and Depende
Modal Search Patterns a
the Search Hypotheses.
the Evaluation Hypothel
the Choice Hypotheses.
Secondary Confidence ET


a. i .......1 -


and

and


S.. ............. ..... ... ...172
.. .............. .. ....... .173
S. ........ ............ ....174
..... ....... .. ... ......... .174
..... ......... ...... ....... 175
..... .........................177
............................. .178
...... ....................... 179
..... ........ .. . . . 179
..............................180
.. .... ..... . .... 184
..............................184
..............................185
on............................186
.. .. .. .... .. .. ... .. .. .... .188
................... ...........188
. . . . . . 188
nt Measures...................190
and Choices...................191
S...... ........................195
ses...........................196
........... .................. 196
effects Hypotheses............. 197


.................168
. ... ........ .. 170

............... .171














Review of


the Dissertation....


y of hnpirical Results..


Contributions of


the Research.


Future Research Directions....


..................................199
. .. .. .... .. ......... .... ..202
..................................203
..................................205


APPENDICES


WIDE RANGE C(XDITION

NARROW RANGE (XONDITIO

NO-RANGE CONDITION IN


INSTRUCTIONS AND QUESTIONNAIRE.............209

N INSTlRUTIONS AND QESTIONNAIRE...........217

STIRUIJIONS AND QUESTIONNAIRE...............225


RANGE STUDY:


SEARCH TASK CODING...................


...232


TASK INSTIRUC-IONS


AND MANIFULATIONS FR HORREATION STUDY


INFORMED CONSENT FORM FORCE RELATIONS STUDY.....................240


ORIENTING


INSTRUCTIIONS


FOR CORRELATION STUDY


.242


TASK QUESTIONNAIRE FOR CORRELATION STUDY.


10 SEARCH TASK SOFTWARE AND TEXT DISPLAY FOR (fRRELATION SIUDY.....275

REF -E ES..........................................................298

BIO(RAWHICAL SKETCHF.................................................307


.....199


CONCLSIONl.... .. ... ... .......... .... ..................


SAMPLE RUN (XRCORRELATION STUDY................................245
















Abstract of Dissertation Presented


to the Graduate School


of the University


of Florida in Partial


Fulfillment


of the


Requirements


the Degree of Doctor of Philosophy


ONXSUMER PRODUCT CLASS KNOWLEDGE:


A TYPOLOGY


IN~RIATION SEARCH
AND EVALUATION

By

Amardeep Assar


IMPACT ON


December


Chairman:


1987


Dipankar Chakravarti


Cochairman:


Major


John G.


Department


Lynch,


Marketing


Consumers'


product


class


knowledge


believed


to significantly


influence


search,


evaluation


choice.


Past


research


typically


examined


role


aggregate


terms


(e.g.


high


versus


familiarity


The current


research


takes a disaggregate


view,


examining


specific


Secondary


functions


effects


use


different


types


confidence are also examined.


knowledge.


Further,


knowledge


strncture


and factors


like


involvement and


time pressure may affect


functions.


taxonomy


consumer


product


class


knowledge


is proposed,


consisting


terminology


brand-attribute


values,


- .I -


#m II -q *-









Attribute


range


knowledge


serve


context


forming


preliminary


evaluations


brand-attribute


information and


these may


mediate


sequential


search


patterns.


Secondary effects on confidence are


proposed.


an experiment


involving choice of


an SLR camera,


three


levels of


range

with


knowledge


seven


(wide,


narrow


brand-attribute


and no range


locations


Subject


information )

s lacking ra


were crossed


nge knowledge


avoided


extreme


evaluations,


searched


attribute


lower


confidence


evaluations.


width


range


did not affect


evaluations.


When range was known,


brand-based search


was more


frequent


favorable


brand-attribute


values.


Next,


consumers


use


inter-attribute


correlations


make


inferences


about


a product's attribute


values


and attain


parsimony in


information


search.


Only


one of


a pair of


correlated attributes may be


searched,


particularly


under


high


time


pressure or


involvement.


Such


reduced


search


lead


different evaluations


, choices and


lowered confidence


proposed


study


manipulates


knowledge


inter-attribute


correlation,


time


pressure


task


involvement


in choosing a stereo


receiver


Subjects


can


sequentially search a computerized


information


base,


obtaining


brand-attribute


information


flexibly


. The


values are


configured


that


subjects


using


correlational knowledge


to search


parsimoniously


are


likely


choose


inferior


brands and have


lower


confidence.
















HAP-ER


IDBODUCTIGN


generally


well-a aepted


that


the knowledge


that a


possesses


about


product class


likely to guide search,


evaluation


choice


behavior


many


significant


ways.


This


rather simple


proposition


can


illustrated


using sam examples of how a onmaer


influenced


different


types


knowledge


that


ay be


available


envi


moreover ,


formal


review


literatures


marketing,


oonsuer research,


puyvhology and econ cs


shaws


mlny


significant


investigations


ting that prior product


knowledge


)rtant moderating


influences n earcr,


evaluation and


choice.


purpose of


this dissertation


to further


investigate


nature


these influences.


In particular,


this dissertation postulates


empirically


explores


specific


types


of product


knowledge


ccns amr


information


processing


choice.


This


introductory


chapter


will


motivate


topic ard give an overview of


the dissertation.


Knawl


e Effects


the Cas-er Ewvir -oe nt


begin by developing soae examples of how know


evaluation and choice behavior.


e might guide a


Consider a


conaser'as


seBc,











encounter


terms


like


shutter speed,


lens aperture and motor drive.


consumer


can


understand


this


terminology based on prior knowledge,


more


information


likely


that


that


available


about


will


able


these features of


to interpret


an S camera.


Hence,


the consumer may search more


for this


type of


information.


Another


photographer


person


friend)


know


that


(perhaps


based on a conversation with a


"Nikon cameras are very good.


" Such knowledge


overall


evaluation of


a brand may be seen as


an adequate basis


making a choice.


Other brands may not be examined,


and even


if more


detailed


brand-attribute


information


available


about Nikon,


this


information may not be searched.


Over


guide


beyond


evaluations.


these effects on search,


Thus


prior knledge my also


a previously-experienced camera user may rate


Nikon


crais


generally


very good.


This overall


evaluation my be


retrieved


applied


knowledge


when


earlier


evaluating


brand


evaluation


new Nikon


can


mode


affect


Thus,


a subsequent


evaluation.


another


example


knowledge


affect


evaluations,


consider


a consumer who knows


that camra warranties range frm 6


to 18


months.


Given


this


range


attribute


values,


1 having an


month


warranty may be evaluated as


the best available on that attribute


searched


further


on other attributes.


1 that offers a


6-month


warranty


receive


imfavorable


evaluation


and not be


Iode


mode


Another










conaumr


examines the advertisements of


amil-order firm and finds


that


these


firn


sell


the aaa merchandise cheaper than regular camera


stores.


Based


this perceived correlation between


type of


store and


price,


may examine alternative


in a camera store,


yet buy


through


mail-order,


even


though


there may be


risks associated with not


having


inspected


the specific camera


that


is prctased.


another


example


knowledge


certain


kinds


inter-attribute


associates

an expensive


correlation


guide choice,


high prices with better


brand even


consider a camera-buyer


lens quality.


This coarser may


though he or she has no way of


testing the


perfarrx


rcpara lenses.


Thus,


the choice of


a specific and my


guided


general


associative rule or a piece of


knowledge retrieved from memory.


Knowlede Resmearch


in Marketing


While


the preceding examples are anecdotal,


any research papers


marketing


literature


have


developed


general


thie


that


knowledge about a product class affects search,


maot of


evaluation and choice.


the traditional models of buyer behavior have


postulated


relationships


between


knowledge


aearh .


Thus,


it bas


been


that a cnmwier uho is familiar with a produt class ay


search


extensively for


information,


first


identifying the attrihrtes


search


identifying criterion levels


to me


in evaluating brands on


those


attrihftes


(Bett an


1979).


For familiar poduct classes,


there


argued











known


(Bettan


1979).


depending


on the


level


familiarity,


both


the nature and extent of


search may differ.


Knowledge


choices


also known


product category.


to affect can'ers'


Of the brads knom


evaluations and brad


to a omsfer,


for.


evoked


set,


and are actually considered before king a


choice.


r ay be predisposed tads sze brands,


i.e.


SU-


favored


over others.


According to Hoard and Sheth


(1969)


-a-rwnrv
1-^OKJE3J0'


higher


class experience


ir.re-0e


inexperienced


cansaer y examine different brands ad


evaluate


them.


With a moderate degree of


familiarity,


fewer


brands


be actively evaluated.


Wsn failiarity is high,


cansi


favor


brands


only


these my be


considered.


Bmeaer


whicd


brands


evaluated


may also differ with


familiarity.


Inexperi enced


oensamers may try to construct choice rules


brard(s)


are encamtered.


With experience,


these choice rules


may be well-learned andI recalled a used subsequenly.


Over


h, ax


foar


attitudes


towards


c-fl


encnomtered


brands.


traditional


marketing


literature provides a


rich


base


research evidence


that


these attitudes play a key role


choices


(Fishbein and AJzen


1975)


In sa


, therefore,


the marketing


literature


traditionally


emtasized


that


the ay cancer search


information,


evaluate


brands


mike


choices


on their


Hence,


depends


predi posed


product


when











Knnal


s Remeard


in Pm-olox1-


Many


studies


in the psychology


literature exaimne how knowledge of


cdain my affect


information


processing about


that &ain.


level


dahin


knowledge


found to produce qualitative differences in


way information


is processed.


proposition


that eem


to be well-acexpted is


that a person dho


relatively


little


about a specific cdmain may find it difficult


encode new information about


that dcain.


.e, in a study of


chess


players,


Chase and Simn


(1973)


find


that asters are better than


novices


recalling


ame


arrangements.


authors


proposed


that


greater


knowledge


chess


enable


maters


to see specific gae


patterns


the arrangmnts of


the pieces and lead


to an advantage in


eroding new


information.


Prior


knowledge


enable


person


determine


appropriate


courses


of action when placed in different situations.


relevant attributes of


a product ay vary across different usage


situations.


who has greater knowledge of


these differences


wre


likely


attributes


selectively


aiate for a situation.


tion


Thus,


SB


specific


is to be used


take


action


pictures of


a sports event,


ay went a motor


drive


that


winds


film


rapidly.


be umed to take


portraits,


conster may want a camera


that uses


larger sized film,


get better detail


. Thus,


knowledge of


desirable benefits


spec


ific


knows


seek










a coceptual


paper


Larkin et al.


(1980)


*trategies


that


aight


oyed


by experts and novices in solving physics problems.


They


suggest


that


the expert ay work


thrag whole sequenc of


steps


that


are


known to be relevant.


In contrast,


the novice may perform ne


step


tiK,


and check each step before going on


to the next.


They


carclude


that


more


knowledgeable


perns


able


to process


information


Given


more


problem


efficiently and faster than a less knoll


(that requires search and evaluation),


able one.


the former may


towards


solution


problem


while


latter


incremental


checking and re eckin.


, the


literature in psychology has developed and tested


several


propositions about how a person's level of


affect


rich


din knowledge may


he or she subsequently processes related


information.


sepni


fairly


clear


that


these propositions have


great


deal


relevance


examining


role


class


knowledge


in c*nsm ers'


search,


evaluation aend choice.


Knowledge Research in ENc mics


Stigler


(1961)


devel


hoped


one of


the best-hkwn propositions about


Can star


search


in the enoiucs


literature.


He argues that


it pays


search


information


as long as the gargiral gains of


doing


so exceed the marginal


costs.


The

r aseks the


lowest


price


S^----


to rely on knowledge of


the distribution of


prices and


quality,


and the costs of


search


in order to search optically.


oCopare


product











lesser


degree


search


behavior


consumers


in more


conventional


product choice


situations.


conceptual


paper


Wilde


1980a)


argues


that search


behavior


should


vary


different


underlying


distributions


prices


quality


update


empirical


their


apartments


expectations


during


sequential


study


Meyer


perceptions


information


1982)


about


search .


models how consumers


a candidate set of


probability


inspecting


alternative


increased


monotonically


with


its riskless


value


the variability associated with attributes


Hence


, according


this


literature


, consumers


rely on


their knowledge of


variability of


attributes during


information search.


economics


literature,


there


implicit


theory


knowledge,


although consumers are assumed


to rely


on their knowledge of


things--prices,


quality


search


costs.


In fact


Nelson's


1976)


classification


search"


"experience"


goods


is based on


implicit


assumptions of homogeneity


in consumer knowledge about certain


product


categories.


However,


such


views are


perhaps somewhat


limited


ignore


rich


variability


in consumer knowledge across


product


categories


between


individuals.


Acknowledging


this


Dohan


(1976)


points


that


prices


are only one of


many types of


information


buyer


may wish


to obtain from


the seller


Wilde


(1980b)


recognizes


that


models are restrictive,


ignoring other types


of information sought,


types


search


how factors


like


time pressure may affect search











Secondary


Effects of Knowledge


The

evaluation


research

n and


undertaken


choice.


here


Working


will


through


focus


on prepuchase search,


these mechanisms


one can see


that


knowledge may alter


effects


variables


information


such


processing and


recall


thus have secondary


recognition


confidence with which an evaluation


is held.


Secondary


marketing


effects of knowledge have


psychology


been


example,


found in


subjects


irical


familiar


work


with


stereos


are


better able


to recall


complex


ideas presented


in a stimulus


stereo


(Alba


1983


Johnson and Russo


(1984)


find


that


amount


information


recalled


about


automobiles


shows


task


familiarity


interaction.


When subjects


perform a judgment


task


recall


increases


with


product


familiarity


. But


when


they perform a choice


task,


amount


information


recalled


inverted-U shaped


function


product


familiarity


familiarity


high,


Less


while more


information

information


is recalled when

is recalled when


product


familiarity


Several


other


is at a moderate


studies


level.


also establish similar secondary effects of


knowledge


recall


and recognition


.Thus


, subjects


with a great deal


knowledge


baseball


better


recall


and recognize


information


that


they


have


previously


learnt


about


the game


(Chiesi,


Spilich and Vosa


1979)

know


another


more about


study


political


(Fiske,


Kinder and Larter


systems are


better able


1982)


to recall


subjects who

information










ade


choices


aOng


brands .


They


find that


the subjects who are wore


familiar with microwave ovens have greater confidence


their choices.


even


preliminary


review of


literature reveals


that


product


class


knowledge


may have significant


ifnu-t o


search,


evaluation


choice


product category.


Moreover,


evidence


also


suggests


that


these effects,


knowledge may also have


contingent


ipactao


the recall and recognition of product


information,


confidence


associated


with


the specific choices


that are


HeB .


basic purpose of


this dissertation is to develop a conceptual


frmeork


within


thich


these


effects


may


be analyzed and


pirically examined.


The next section overview the dissertation.


Overview of


the Dissertation


Chapter


2 presents a literature review that form


basis


for the


framework


in this dissertation.


It is argued that


even


though


well-accepted that consumer knowledge of


aproduct


class


can


affect


information


evaluation


choice,


past


research


in this area is


limited.


This


because the role of knowledge


typically


been studied only at an aggregate level,


making only the


coarsest


discrimination such as high


versus low knowledge.


In fact,


knowledge


consists of any different


types


that can play very different


roles


informaticm processing.


Since past research generally ignores


this


very


basic


point,


a central


proposition of


this dissertation is


that


product class knowledge should be viewed in disaggregate


coceCeptual


proposed


search,


oonsiier










conceptual


fraework


in Chapter 3 therefore


takes a


disagregate


taxon y


view


this


ooasuBer


knowledge


product class knowledge.


explicit


It develop, a


links between


different


types of knowledge


identified in


this


taxcncy eand their role


cOImsters'


searh


evaluation and choice processes.


Also,


while


central


feature of


thee


irical


parts of


the research,


frmework


recognizes that


the structure or organization of knowledge


moderate


certain information processing factions are


performed.


example


C-orIer


whose


knowledge


camera brands is


structured


evaluative


categories


msh as good and bad cyeras my


serarh


the good brands selectively buhat not consider the


"bad" brands at


all,


even


when


wderlying


attribute


information base may not be


cop


lete or even accurate.


second


example,


consumer


may know that for a particular


usage


situation,


oae


brand


better than another


(e.g.


"For action


pictures,


Nikon


better


than


Yashica")


When


situational


appropriateness


organized


brand,


only the favored brad may be


searhed.


persacn


knows


that


a particular mage


situation,


one


attribute


aore


rrtant


than another


(e.g.


"For


portraits


situational


is preferable


appropriateness


to use caeras that use


organized


larger film").


term


ten


attributes,


different brands may be searched on


the relevant attribute.


propoeed


proposes


suppoee










which


are advanced.


Consider a person considering purchae of


a stereo.


knows only basic


terminology,


he or ashe my only check


stereo


system is,


and how


A it costs.


Salespersron


asked

knows


advice and recedatians.


On the other hand,


consumr


the aore advanced


terminology,


that person may search on


attributes


like


wow


flutter,


Dolby


noise


reduction


and total


barsxic


distortion.


This person may be c

compar


ing technial


exarinig


reports


audio


-gaines.


Thus,


evaluation


activities


perfo


rled


differently,


depending on hob


knowledge


is structured.


functions


knowledge


also


moderated


other


bwk-Higin


factors


such


search coets,


coarser involvement levels


time


preure.


these


factors appear


in the studies that


will


reviewed in


the next chapter.


example,


consider


trying to choose among a large


Lube


barnds.


knows


that


attributes are


correlated,


rely


this


correlation knowledge and


only


one attribute aut of the correlated pair.


This effect my


ocaeatuated


search


coats


are


high


there


is time


pressure.


Thus,


situation


where


aoosfr


is awre of the


existence


correlation


betMeen,


say,


wattage and fidelity


ratings


stereos,


be or she my obtain information only about wattage


and thus acut da


the ammnt of


search effort expended.


powerful


search


data,


Some


examine


when


dhen










emonrioe.


This


chapter also


identifies potential .eccmdary effects on


recall,

effects


recognition


knowledge


confidence.


structure


here apriate,


other


raerm


factors


ting

are


effects


types


of knowledge


in the above taxaonoy are


explored


this


dissertation.


first


these is


consumers


knowledge


the range of


values available on an attribute


"attribute


knXledge.


" apter 4 review.


literature and


develops


scenario


that


used


explore the


likely effects of


attribute range knowledge on search and evaluation.


Chapter


presents


pirical study of


the effects of


attribute


range


knowledge.


ItBannir


parposed


whioh attribute range


knowledge


affects


preliminary


evaluation


given


piece of


brad-attribute


information.


This provisional


evaluation


then mediates


further


search.


CQape


tinug


explanations of


these effects are


tested


context


camera choice


task.


effects on confidence


are also


inveLtifatsd.


OCapter


presents


findings


the empirical study of


range


knowledge.


then


discusses


gains


-ae


fra the disaggregate


conceptualization


irical


investigation of


range knowledge


that


formed


basis


this


. Theoretical,


managerial


public


policy issues are disoassed in


the light of


findings.


backgromi


proposed.


empirically


range


Secondary


rifnnr


research










correlations


conmize


effort


expended in information Mearh.


This


affect


what


information is acquired,


turn ay affect


evaluation


choice.


Formal hypotheses are


presented


this regard.


Chapter


hypotheses


describes


are


design


tested


exper


context


which


a stereo choice


these


task.


actual


conduct of


this study is outside the scope of


this dissertation.


However,


the study design specifically points


to the establi


wae


t of


future


research program.


Chapter 9 reviews


the dissertation,


s War


ime


findings


discumse


contribution


this


work and the


expected directions


for future research.
















CHAPIAI 2
LTfRAlf RnVIW


central


idea poeed in


the previous chapter


that conm


ana


have


different


types


knowledge


that


4nrrr^t


on inf


stion


evaluation


choice.


addition,


those


effects


on how


information


processed iy have further secondary effects an recall,


recognition and confidence


in evaluatioms.


This


chapter


will


review


literature


relating


cons


knowledge,


framework


doing


presented


the stage for the


in Chapter 3.


forml conceptual


We begin by emphasizing that


canoster


knowledge


studied


with a focus on its


leaning and


storage


-Dry,


or its activatima and retrieval


frm mry,


or hb


used


specific


cansuser tasks.


We first establish


that


this


dissertation


last


ientianed


focus.


next


discuss


definitions


coaBmSar


knowledge


familiarity


available


in the


literature.


This


followed by a detailed review of


research which has


studied


knowledge


aggregate


term.


To aspects are discussed


effects


search


evaluation


choice context,


secondary effects on recall,


recognition and confidence.


Following


this discussion of


aggregate views of


consm er knowledge,


turn


work


that


taken


a disaggregate


view.


Some specific


meaC h











effects


search


evaluation


choice,


then


looking at


secondary


effects on


recall


, recognition and confidence.


Approaches


to Studying Consuner Knowl


have


mentioned


earlier


, consumer knowledge may be studied


with


focus


on its learning and storage


in mory


on activation and


retrieval


from


memory


or on how it


is used


in specific consumer


tasks.


first of


these approaches


is important as


indicates how consumer


cognitions


develop.


this


domain


one


examine


issues such as


covariation


assessment--to


what extent can a consumer detect and learn


presence


a relationship between


two attributes


of a brand in a


product


category


(e.g.


Bettman,


John and Scott


1986


John


, Scott and


Bettman


1986).


While


studies


this


type are being conducted in


marketing


literature


, this dissertation does not


focus aon


these


issues.


Studies


consumer knowledge can also


focus on


the activation and


retrieval


product


category


information


memory


. These studies


typically


focus


encoding and retrieval


factors operate during


consumer


judgment and choice


tasks.


fairly


well


psychology


literature


that


encouraging


improves


(Craik


attention


recall.


Lockhart


stimuli


For example


1972)


elaboration


, if processing of


or a person


during


information is


is required


encoding


"deeper"


to think about an


object


relation


to himself


(Tulving and Pearlstone


1966)


recall


likely to be higher.


known










helpful


recall


if they are


the saame cue as those available during


eKcding4.


researchers


have


study


accessibility


product


decision


tanks


well.


Johnson and Russo


(1978)


have


studied


memory


product


infortmtion.


They find that


overall


evaluations of products are remembered better than intermediate


stages


choice


promesses


end brand attribte information.


Alba and


Chattopadhyay


brands


(1985)


recall


have studied the


iqpect of presenting a subset of


reaining brands


in a product category.


authors


find


that


brands


are undifferentiated,


giving subjects


sauch


subset


inhibits


recall


other


brands.


brands


perceived


being


differentiated,


part-category


cues


inhibit


recall,


else


srvec


cume


other


sutcategories and thereby


enhance recall


of other brands.


Biehal and Chakravarti


(1982,


1983,


1986)


find that


the fort and


goals


wuer


which


information


learned affects how it is retrieved


used


-sbent


choice


tasks .


Brand


informticm learned in


fashion is easily retrieved,


dmreas brMd int


tion learned


choice


relatively


less accessible


in mry.


These


differences


access


ibility


product information impact on


ubequet evaluation and choice processes.


While


study


activation


retrieval


aoonser knowledge,


issues are critical aspects of


the research


in this dissertation focuses


are


extended


When


directed


incidental










learned


retrieved


from


Wery,'


foctB


how these type of


knowledge


are used in


the conmsuer decision process,


asmirsg that


thia


knowledge


learned


retrieved or


is otherwise irde available in


the course of


a c


r task.


Prior Definitions and Measures


research,


product class knowledge has


been


thought of


many


with


different 'as.


HMot of


total


theme coaeptualiations are


knowledge,


and do not distinguish between


different


types of knowledge.


Also,


the tern familiarity and


knowledge


have


been used interchangeably.


A review of


these different definitions


literature


sggests


about exactly hw


product class knowledge might be ca~ceptualized.


Johnson


Rhsso


(1984)


view


"product familiarity"


ad knowledge


product


class


s8mTBOUs


ideas.


They


interested


exaining


whether different aaont of knowledge affect new information


acquisition


choice


judent


tasks.


tWile


they


acknowledge


that


knowledge


can <(


Sfroi different sources,


including


direct


experi


enoe


(purcasin and uing a product)


or indirectly


(from


advertising


or "ord-of-mouth),


they do not distinguish


between


types of


knowledge.


Knowledge


levels are measured by subjects'


self-afseents


familiarity with


the product category.


Olson


(1981)


define


familiarity as


cognitive


starutures


a product


(preutmbly caMued by various prodct-related


concerned


oonaensi


Unt


lack


arksa










different


type


knowledge,


acquired frm different source,


these


authors


ea U


self-reports


familiarity


product


sage


overall


experience,


level,


frequency of


obtaining


ume0


typology of knowledge


ia not explored


further.


individual' a


subjective


perceptions,


i.e.


h a person


thinks


used by Biehal


(1983)


as a definition of


class


knowledge.


Similar


ideas


are


devel


defining


subjective


knowledge


(Bruct


1985)


and sujective failiarity


(Park and


Lesaig


1981)


These authors explore propositions suggesting that a lack


confidence


one's knowledge ay motivate a person to search are


information.


These


studies


also


uIe


aelf-ae ~enta


knowledge/faailiarity as the primary


t appro h.


Jaju


Reilly


(1979)


view familiarity in term of


"frequency of


overall


familiarity


knowledge


select


beat


brand.


" Their


definition


covers usage experience,


total aunt of


knowledge


and a specific


type of knowledge


(i.e.


how to elect


the best


brand).


their


enpirical


uBm


self-asaesaents of


faiiliarity/knowledge as aeamses.


above


review points up the


lack of


a clear con-sens- regarding


knowledge


best operatioalized and also shows a clear


attention


the devel


t of


a taxcnmy of knowledge.


only


paper that has gone to


trouble of


developing a


taxonomy is onem


Brucks


(1985),


lays about alone classification acme


for different


o sd


leaDSue-aal


OlJU1"


know


product


use,


Once


agin,


work


product


ladk











rheeate Prodact Familiarity:


Effects on


Evaluation and Choie


predmicant


view


iliarity or product class knowledge


existing


literature


is at an aggregate


level.


The entire knowledge


that


individual has about


the product class


leads


to a feeling


"being familiar"


with


the productclass.


cosrmer's sense of "being familiar"


is typically assessed by


obtaining


content


self-ratings


what


individuals


indicate


level


amsxm


they


familiarity.


considered.


have a high degree of


specific


Thus,


familiarity with


product


class of


cameras.


they may differ considerably in what


they


actually


knIw ard how that knowledge was acquired.


Sperson ay


great deal about


the tehnial


features of


erat,


having used


-my


brands


for taking different kinds of


pictures


Another person ay


about


prInces,


warranty term and overall assessents of different


brands,


having


sought


information


stores,


fra


inependent


surges


distinguish


between


fron


what


friends.


these


Clearly,


individuals


rtant


because these


differences


in the type of knowledge that


they pjoesnill aost likely


the to bu amaeras


in very different wmys.


section


will


review


body


the knowledge


literature


wherein


these


distinctions


are not de and product class


knowledge


viewed


aggregate.


While


ae objective of


this


fl- l I- ..~ A t. t a S- --


A* a -


Q@i,


knws


know


know


about


know


lead


next


1HI_ __*





~ 1 II


I


i k











available


literature


suggests


that


level


of knowled e has


effects


maUat


information sought,


as well as an specific


patterns


information


seeking.


Also


there


isa om evidence


for a


link


between


knowledge


choice


behavior.


Thus,


Bettan and Park


(1980)


cansiaers'


prior knowledge and experience and the


choice affect


types of


information used,


kinds


subjects


choice


with


processing


high


that


familiarity


Occ.r


rely


They hypothesize that


wore heavily on prior


information,


instead


processing


currently


available


information.


Those


since


with


they


iliarity may not be able


to process new information


lack the basic knowledge necessary to mke sease of


the ew


trial.


This


Snre


likely


the product class is relatively


c-Hp


lex.


High


familiarity


subjects


are


motivated


new


information,


being able


to rely on what


they already know.


By contrast,


majtets


with


moderate


familiarity


are


more


likely


process


information,


since


they


are


able


UBe


their


prior knowledge


ass milate


information.


They


are also ore motivated to seek new


information.


authors


also expect


that


the degree of


attribute processing is


negatively


related


degree


prior


knoIledge,


i.e.,


attribute


processing


decreases


brand


processing


knowledge


ireses.


Sine ost of


a conser's c on duct-related experience


exa ine


ThaB


dhen


process










authors


self


ratings


iuage and ownerahip of


micrmwve


ovems


screen


subjects


into


low,


od.


rate


high


familiarity


gruaps.


Sujets


are given a brand x attribute atrix and


perform


tasks .


First,


an elimination task,


they select


those


that


are


acceptable


thei.


Then


, subjects coo the brand


they most prefer from the set of


acceptable brands.


Reliance


prior


information


highest


high


familiarity


new


subjects,


information.


suggesting


amxout of


that


these subjects search less for


attribute-wiae processing


lower for


experienced


oosmers These


results


show


that


which


cmnsuters search and evaluate alternatives -ay


depId


on f


iliarity.


Johmnson


ability

study


(1984)


examine whether familiarity affects the


learn new information in choice and judgment contexts.


uses autoobiles as the product class.


This


Subjects are given a brand


attribute


mtrix


and asked to rate each alternative


(judent task)


choose one frm the set


(choice task)


After


eting the assi


gned


task,


they


provide


self-


their previous knowledge of


automobiles,


their


image


experience


with


product


class .


Subject


are


asked to recall


information they were given and any


other


judgments or


cmpar


iasns they a during the task.


The aamt of


information recalled is used as a measure of


researchers propose that


searcd.


the ore familiar subjects my have an


advantage


during


judgment


task.


They are able to encode


search,


brands


Russo


?


then


encoding


COIPl










familiar


subjects


not acquire redwuiant


information about existing


product.


IThey


selective


attention


relevant


information and


ignore


irrelevant


inform tion.


In the choice task,


they expect subject


ignore


Sa


information,


thus


reduce


earh as


liarity


i r-e -


amount


information


recalled


is tapubu


to be an


inverted-U


function of


familiarity.


Wen familiarity is


low,


choice


task,


recall


expected to


increase,


consistent with an


"emri


dhen


Bmdiani-B


(i.e.


increasing


familiarity


facilitates


learning)


rather


than


with


"selectivity" mechanism.


In any case,


authors


find


that


amoamt


information recalled showa an


inverted-U shaped fiction of


familiarity in


the choice


task.


Using


only


aggregate


conceptualization


knowledge,


it is


difficult


nail dohn


the explicit causes of


the findings.


Consider a


prospective


camera buyer tho


is familiar with


the meaning of


terns


like


aperture


size,


motor


drive


etc.


This


may be able


to mke


better


aense of


information available about


these attributes


(i.e.


have


encoding


advantage)


may search more brands on


that attribute,


to a oonsi


ho does not know the meaning of


these technical


terr.


conmser ~ho knoz s


that


picture quality depends Dainly on the


lens


quality


cera


look for ratings of


lenses


in photo


magazines


and not pay


hA attention to brands Aichd have features


like


microchipe


that


tell


the amer en it is okay to


take a picture.


Knmwing


attribute


importance enables


this er to be more selective


shaped










Thus,


even


is plausible that certain


types of know


sedte


give


cwastaer an aenoding advantage or


leed


to selectivity in


information


effects


processing,


directly


possible


in aggregate studies ch as J


verify any of


hnaron and Rhuso


theM


(1984)


Conceptually,


US


types


knowledge


incresse search,


dheres


others


attesuate


In any case,


the point remains


that


there is


-uch


gained


linking


frO, adopting a disaggregate view of knowl


identified type of


product class knowledge


e and


to its function


in consaer tasks.


Several


knowledge


other


effects.


studies


Moore


conduct


L-eMlnm


similar


(1980)


aggregate


study


analyses


pattern of


information


for a non-durable product


For dch of


eds,


subjects acquired information on,


dhoee and cons


oe of


five


brands


bred.


They were questioned about familiarity with


types of


bread


attributes,


similarities


different


types of


breads


etc.,


order


evaluate


knowledge or ubat


loosely


tered "

different


experience.


types


" The


effects


variables,


external


(e.g.


situational


msech of


variables,


rmber of

potential


payoff/product


>rtanoe,


knowledge


exp


other


individual


differences)


are


studied.


the 1nwlexge/zexperience-related


variables


mact


more


the varianre


in search


than any of


other types.


also


rake


distinction betwa


expen


arcquired


-ii -I--


though


eeah


seeking


(breed)


tiportan-e


authors










xtern l


SMrch


during the


exper


inmt


is pots ed to be


negatively


related


experi


ence


prior to


the experiment.


Informatimn acquisition


found


related


prior experience,


self


reported


nutrition


knowledge


shopping behavior,


but unrelated to


experience


with


bakery.


during


experiment


.easmd


ntmer


previ pase


experiment.


External search is expected to be netatively related to the


of previous purchases,


and this is supported by the data.


Thkea


together,


findings


this


study


suggest


that


knowledge/experience


an important


predictor of


external


search


that


More


experience


leads


lower


information


acquisition


nnx-durable


products.


Yet,


despite


this


important general statent,


study is unable to pin do&w


the specific effects of


different


types


of khrledge of


a product category.


another study of


information seeking,


Biehal


(1983)


eamines


iaWrct


prior


experaene


individual differences on search


auto


repair


Services.


If previ~s experi


is high,


the co-smmr ay


search


information available internliy,


i.e.


there


less


need


search for external


information.


The data are


from


of hoeholds


that had incurred at least one major auto


repair expose


in the preceding


12 months.


independent


variable


interest is the ccnaser's


perception


extent of prior knwledge sabht auto repairs.


To Measure mmy


Siosures


Experience


term


survey










visits


repair


ccpanrnes


number of


friends from na


information


sought.


Prior


information


faind


to be positively


correlated


with


internal


search and negatively correlated with external


search.


effects of


price and brand familiarity on brand preferences are


discussed


Monroe


(1976)


Subjects categorized their own experiences


with


three


products,


coffee,


fabric


softener


spray


cologne.


Subjects


were placed in one of


three cognitive


levels


familiarity),


taking


into


ccoxxnt


and usage experience.


Subjects


crmpa


brands,


pucase


as the price of


experience


one was varied.

, that experience


When subjects had previous

Itial information dominated


price as a factor


in choice.


summary,


previous


constuner


research


has established that


aggregate


level


knowledge


product


class


related


information


search


evaluation


choice


context.


However,


relatively


little


been


documented


about


the specific


impacts of


different


types


knowledge


a product category


on the different


dependent


variables analyzed


in these studies.


Aggregate Knowledge:


Effects on Search,


Evaluation and Choice


studies


psychology explicitly


link the aggregate amount of


knowledge


some


functions


which


knowledge


might serve.


first


links


a higher


level


of knowledge


to an increased ability to encode new


information.


second suggests


that experts ay be more efficient


purchase










OCase


Simon


(1973)


sought


to isolate and define


into


which


information


is supposedly ended


in chess


perception


tasks .


perceptual


task,


subjects


have


c essbor d


view


reconstruct


positions


onto another board.


Chess masters and novices do


differ


performance


this


task .


second


task,


arra~ngmet


pieces


is observed briefly and


then reconstructed frn


memory


Recall


masters


is no different


frEm that


for novices when


pieces


are arranged randoly


But masters have superior recall


geme


arranemets.


authors


propose


that chess masters may have an


advantage


encoding game arrangements,


enabling the. to recall


those


better


than


novices


While


encoding


advantage


is one plausible


explanation,


also


possible


that


masters


recall


arrangements


they


observe


experiment,


only reconstruct


previously


learned


arrafgeents


after


categorizing


presented


arrangement


encoding


heuristic


that


labels


it as a previously


learned


gam


pattern.


this were true


,the coding


"advantage"


indeed


turn


disadvantage


whereby


criterially


similar game


arrangement


labelled


errmneomly


ensuing reconstruction -y


well


retrieving


differences


detail


between


presented and


"ercxxec1"


arrangemut.


Larkin


. (1980)


propose


that


task performance of


experts may


differ


from


that


novices,


with


the experts using more efficient


problem-solving


strategies.


example,


experts may have available a


"chunks"


e m










take.


This may


enable experts


to work more efficiently and with less


effort expended when solving probl


in that doain.


greater


domain


knowledge helps subjects encode new


work more efficiently,


information about a


constaers having greater knowledge of


product


class


behave


differently


from thoLe who have


less.


consumer


translate


knowledge


a usage situation


to assess what


kind


product


performance


attribute


specifications


will


appropriate.


This


reduce


effort


in searching for


information.


For example,


an audiophile may examine high quality stereo


equi


use


only in a situation where acoustics are


good.


He or


advise


rnovice


to consider only cheaper compact stereos


use


in a small


apartment.


above


knowledge


studies


my serve


explicitly


in information


consider


processing.


functions


In another study,


that


Miyake


Norman


(1979


examine


whether


knowledge


is a pre-requisite for


asking


questions


about


a ew topic.


The subjects


in this


study differ


their


level


knowledge


a text-editor,


arl are


learning more


about


that domain.


Trained subjects ask more questions when material


hard,


while novices ask more questions


The extent of

information search


when it


question asking can be


behavior.


is easy.


thought of


In the conxmter context


as an expression of

t, people with more


knowledge


search


more


when


the externally available


information


matches


their


current


level


knowledge about a


product class.


expel


xded











Aggregate


Product


Familiarity Studies:


Secondary Effects


knowledge


affects


search


, evaluation and choice,


it affects how


both


and new


have


secondary


information


effects


is processed.


recall,


As a consequence


recognition


, knowledge


confidence


derived


from


evaluations.


These


mediated


effects


are


often


considerable


interest


in consumer choice situations.


studies


marketing


literature suggest


that consumers'


level


of knowledge may be


positively related


to the accuracy and amount


information


recalled


, following search and evaluation.


In the


first


these


studies,


Moore


Lehmann


1980)


find


that knowledge


related


search


behavior


choice.


While


they do not explicitly


hypothesize


this,


authors


suggest


that search and choice may be


linked


aided


recall.


Correct


recall


of attributes


is positively


related


to the amount of


external


search


on that attribute


, and also


stated


importance of


an attribute.


For a high-knowledge consumer


there


ess


need


to search


for external


information


However,


important attribute


is searched


, recall


is likely


to be higher.


Moore


Lehmann


also


find


that


brand


recall


accuracy is more


highly


related


purchases


a brand within


the experiment


than


acquisition of


information about


brand


. This


is understandable


considers


that


be particularly useful


to retain


brand-specific


information


for a


brand


that


is merely examined but not


purchased.


other


hand


if the brand


chosen


, additional


, as


one











studies


that


find


that


chosen


brand


information


is remembered better


than


information about


rejected brands


(Biehal and Chakravarti


1982)


seoroi


study of


this


issue


Alba


1983


has examined whether


extent


Subjects'


knowledge


self-assessments


related


knowledge


product


about


information


stereos


are


recall.


used


separate


them


into


high


vs.


knowledge


groups


finds


that


subjects


are


familiar


with


the product are better able


to recall


complex


ideas


presented


relatively


technical


stimulus


advertisement.


These


subjects also evaluate


the advertisement as


being


more


useful.


These


findings


suggest


that


relative


unfamiliar


subjects,


familiar


subjects


better


cop


rehernd


ads with


canp


information


. Thus


they may find such ads more useful.


Alsb


(1983)


study


also


shows


that


the product


is rated more


favorably


subjects with lower knowledge.


More familiar subjects my


able


understand


implications


of the attribute


levels and


evaluate


brand


accordingly


Less


familiar


subjects


understand


attributes


and may


give


brand a higher rating than


they


results


are able


suggest


tnderstarding


to understand the attribute


likelihood of


uncertainty


, both


levels


evaluation being


types of


Even


though other


lowered by


results point


lack of


to secondary


effects of knowledge.


the psychology


literature,


a large body of


work examines how the


level


knowledge


affect


subjects


recall


recognition.











Chiesi,

knowledge


Spilich


about


VOaS


baseball


have


1979)


better


find that subjects with higher

recall and recognition of new


information


given


to them about


the game.


Morris,


Grneberg,


Sykes and


Merrick


(1981)


find a high correlation


between subjects'


perforuanoe on


questionnaire


of knowledge about soccer,


their recall


scores,


immediately


following


broadcast.


contrast,


subjects


in a control


group


show


correlation between general knowledge of


soccer and


number


correct


predictions


nade.


authors


suggest


that


experimental


subjects


could


have


guessed correctly.


Rather,


knowledgeable


subjects


have


greater


interest


in the scores,


and encode


the new information more effectively.


domain


political


knowledge,


Fiske


, Kinder and Larter


(1983)


examine knowledge-based strategies


in processing


information,


function


levels


expertise.


knowledge about


Expert

political


subjects

systems)


(i.e.


those who had higher


are better able


to recall


information


inconsistent with


their prior knowledge.


Tyler


Voss


(1982)


compare


prior


factual knowledge and


attitudes


affect


recall


recognition in


text


processing.


Subjects


tested


determining their


level


factual


items.


Knowledge


distinguished


from


attitudes,


which


are


measured


using a


test with


affective

factual


items.


items


Cued


when new


recall


and recognition are correlated with


information given is


incongruent


with subjects'


expectations.


correlation


is higher with


the affective


ites when


are










studies of


recognition memory,


Arkes ard Freedian


(1984)


canpare


performance


used

topic,


experts and non-experts.


to separate subjects


knowledge


A quiz on baseball


into high and low scoring groups.


knowledge


For another


local students are deemed experts


while


those


from


a western campus are


the nxm-experts.


Recognition is


higher


experts


contrasted


with


non-experts.


this


relationship


holds


only


when


the distractor


items on


the recognition


test


are


related


targets.


When the distractor


items are


related


targets,


expert subjects are


less


likely to correctly


recognize


items


seen


previously,


compa-ed


to non-experts.


One might


argue


that


on such


tests experts may encode a stimulus on


basis of


their


prior


knowledge


subsequently


-" p"


pre-learned


information


they


have


about


topic


in -emory,


rather than actually


recalling


presented material.


A similar argtOent


is made about


Chase and Simon


(1973)


findings reported earlier.


amxr


there


is evidence


literature


that higher


levels


knowledge


are


associated


with


better


recall


and recognition of


information.


more


Cxems


easily recall


familiar with a product class,


information about brands


inspected in


they


the past.


Having


been


exposed


advertising,


they


may more easily recognize


brands


in-store.


Knowledgeable


subjects'


ability


recall


recognise


information


enable


them to use such


information in


choice


context.


However,


before,


these


aggregate


studies


Ohio geography,


Serare are











that


disaggregate


viewsB


of knowledge


begin


to take


important steps


this direction.


Disaggregate Studies of Knowledge:


Search,


Effects on


Evaluation and Choice


contrast


preceding


aggregate


conceptualization of knowledge


involves


identification


different


types


of knowledge within a dcain.


A particular


type of


knowledge


can


be conceptually


linked


to a role or


function


that


it can


serve


search,


evaluation


choice.


This conceptual


linkage can


then be studied empirically.


example,


camera buyer may have knowledge about


the range of


available


brand


values


attribute,


the correlation


between


features,


the overall


evaluations of


the available brands.


For each


these


particular


types


functions.


p-odct


Thus


category


, for example,


knowledge,


knowledge of


specify


inter-attribute


correlations


can


used


reduce


search


effort,


attribute


range


knowledge


may provide a


basis


for the provisional


evaluation of


a piece


brand-attribute


information


on. Moreover,


role


that a


given


type


knowledge


plays


differ depending on


the specific


scenario or


the aspect of


consumer choice


tanks


being studied.


Some


recent


articles


have


tried


to develop typologies of


consfter


product


work


class knowledge.


this


These articles are


important


the stage


precursors of


for our own conceptual


disaggregate


conceptualizations,


dissertation.


I ___











Punj


Staelin


1983)


propose a descriptive model


of information


search


test


using


data


from


survey


auto


Ibers.


Self-ratings


amount


knowledge


are obtained.


types of


knowledge


are


identified.


Usable


Prior


Information


is operationalized


the specific knowledge of


models and dealers.


Prior kMeory Structure


is defined as


the consumer's knowledge of


cars


in general.


study


examines


impact of


these


two dimensions of knowledge


the anmont and nature of


the external


information and


time spent


searching


information.


The authors hypothesize


that external


search


negatively


related


to Usable


Prior


Information.


Thus,


a person who


knows


about available models and


the process of buying a car has


little


need


external


information


search


about


models


as the expected


benefits


search


are


limited.


This hypothesis


is supported by the


data.


External


search


is hypothesized


to be


positively related


to general


product


class


knowledge.


conaimer


knmws a


lot about autos


general


better


able


comprehend new information.


This may


increase


search.


Thus,


experienced


auto


bayer


who knows about


construction


engines


, types of


carburetors and


types of


suspensions


search


However,


oIre


information about


the authors do not find


irical


these aspects


support


than the novice.


for this proposition.


Russo


knowledge.


and Johnson


They


(1980)


this


develop a hierarchical


taxonuy


taxonomy


investigate


of product


content and











Best Brand


. Whole


Brand/Attribute Ranking or


Camper


ison


Whole Evaluations of Brand/Attribute


Single Pair
attribute)


Camps


risons/Rankings


(i.e.


of brand on an


Single Evaluation


This


(Brand Attribute


develops


Value)


an interesting progression of


coxmp


inferential


levels


the knowledge of


a product category as


developed


the context of


a choice.


orever,


it does not examine how


these


knowledge


types


are explicitly used in consuaer choice.


We will


contrast


our


work


with


that of Russo and Jnson more specifically in


Chapter 3.


Park


Lessig


(1981)


study the


impact of


product


familiarity on


conasier


decision


biases


and on


their use of heuristics.


They propose


approaches


for operationalizing and measuring product


familiarity.


These are:


much


person


knmws


about


the product


(i.e.


the actual


amount of


information contained in long term


mo


subjective


familiarity


assessment


(i.e.


much


a person


thinks he or she knows about


product).


their


familiarity


empirical


asking


study,


Park and Lessig


subjects


whether


(1981)


they


assess


have


subjective


searched


information


about


mxcrCeve ovens,


se oX e or omn one.


The respardents


assigned


ne


three


(high,


mIedium


low)


levels


familiarity,


based


their


responses.


Starting with a large set of


)wnwlc- anb at-+e nrfnveci 1111 flh 41 rm~ teh *nlnA a nkn: n


are


taxonxw


increasingly


Lrvrnrta


mi inlarv~


M~Qnr


I A- -_


4aaL











while


high


familiarity subjects use both non-functicmal and


futctiaral


attributes.


Those


with


medium


level


subjective


familiarity have


lower


levels


confidence


their


evaluations.


Thus,


subjective


familiarity


acquire


can


during


affect


evaluation


nature


choice.


information


Also,


that consumers


can differ,


depending on


the level


familiarity.


Alba


Hutch inson


(1987)


suggest


that


there


is a need for a


multidimensional


view of


consumer knowledge.


They define


coponren


knowledge.


Familiarity


is "the nberof


produit-related


experiences


that


have


been


yacxuimlated


by the xnwn


component,


expertise,


"the ability to perform product-related tasks


successfully.


" The


authors


suggest


propositions


relating


development of


the expertise caponent.


Olshavsky


(1982)


study


prepurchase


relationship


external


between


information


search.


These


authors


also


view beliefs as subjective


perceptions of


individual


that


or may not


true or correct.


Five


types of


conster


beliefs


are


identified.


These


are


attribute


aurrogates or


signals,


ability


judge,


benefits of


search,


variance of


offerings


variance


supply.


Thus ,


these belief


categories are


about


the consumers'


own skills and about


the market,


rather than being


product specific.


each


beliefs,


the authors hypothesize


the direction of


the resulting confidence


marketplace-related


beliefs


constaer


ar." T'e lexxx


DIuncan











level


identified


individual


earlier,


this


beliefs


study


rather


establishes


than the belief

that different


categories

it types of


knowledge can


influence haow consumers search


information.


Brucks


(1985)


conducts


avery


artant study design


to clarify


meaning


measurement of


product class knowledge and


to examine


effect


information search


behavior.


Knowledge


is distinguished


being


objective


subjective.


Objective


knowledge


what an


individual


actually


ry, while subjective knowledge


is what


individuals perceive


they know.


Brucks


classifies


objective


knowledge


into


following


five


types:


terminology,


attributes,


criteria


for evaluating


attributes


, (4)


perceived


correlation


between


attributes


situational


factors.


This


detailed


typology


provides


basis


understanding


role


different


types


knowledge


in search,


evaluation and choice.


Even


though


Braks


tests


hypotheses only about how objective and


subjective


knowledge


affect


search


and does not


test


for the


role of


specific

findings


types


objective


study


knowledge,


starting


useful


to review the


point for additional


research


losing a disaggregate view of


consumer knowledge of


a product class.


Brucks


measures subjective knowledge as ccnsuers'


self


assessments


confidence


in their


level


of familiarity.


Objective knowledge


types


measured by testing


for content.


In the experimental


task,


subjects


are










brand-attribute


information


maniplataed--subjets


choose


matrix.


intended


a sewing machine


for a


usage situation


person whose needs


are described as being either simple or


Comp


lex.


Objective


knowledge


statistically


significant


effects


search,


but only


in the


uomp


lex usage situation.


Objective knowledge


positively


related


to the number of


attributes searched.


This suggests


that


greater


knowledge


enables


consiBer


to ask questions about


attributes


(See


also


Miyake


Norman


1979)


Greater


objective


knowledge


alternatives.


associated


Moreover


with less

objective


information search on


knowledge


inappropriate


also associated with


greater


variability


number of


questions asked about available


alternatives.


These


results suggest


that objective knowledge


is related


to greater search efficiency.


discussing


findings,


Brucks


(1985


proposes


hypotheses


linking


each


five


types


objective


knowledge


to search


behavior


However,


81siXe


the five


types of knowledge were


founi


to be


highly correlated,


the hypotheses could not be


irically tested.


Subjective


knowledge


shows


relationship


with


number of


attributes


examined


positively


related


to the variability of


search.


camp


usage situation,


subjective knowledge shows a


negative


relationship


with


extent of


inappropriate search but


positively related


to the variability of


search.


asuBry


, Brucks'


findings


suggest


that


different


types










departure


point away from aggregate views


to more disaggregate views of


consumer knowledge and


its effects


in search,


evaluation and choice.


Disaggreegate Studies of Knowledge:


Sec a ry BEffect-


now


turn


studies


that


examine


cnOsumr


knowledge


in a


disaggregate


manner


identify


effects


secondary variables of


interest.


There


are


studies


that


have established links


between


different


performance.


types


knowledge


However,


number


subsequent


studies


recall and recognition


find relationships between


what consumers know and


the confidence


that


they have


in their choices.


Park


Lessig


(1981)


find


that consumers with high subjective


familiarity


are re confident


their choices.


Thus,


high familiarity


reduce


degree of


uncertainty that a consumer feels


. They also


find


that


conslmIers


with


subjective


familiarity


have


higher


confidence


than


those


with


moderate


familiarity


person


unfamiliar


with


microwave


oven,


he or she may


lack


the ability to


information


, and may not realise what information


is needed


rmke


thorough


evaluation.


This


person may have high confidence,


though


objectively


it may be misplaced.


A person with a moratdedegree


familiarity may realize


product class and hence have


that he or she does not know very tch about


loaerc nfidence.


Similarly,


confidence.


knowledge,


Brucks

comuaers


this


(1985) a

perceive


reduce


measures su

themselves


objective


uncertainty and lead


knowledge


having


to higher confidence


encode











better than another,


without knowing whether,


in fact,


there are any


differences.


high


priced


camrar


bought because


the person


believes


that


expensive


products are better


, even


though


in objective


terms,


chosen


brand


may not


perform better than a


less expensive


priori,


the conser may not be able


to verify the validity of


his beliefs about


summary


the product's quality.


studies employing disaggregate views of knowledge


examine


knowledge


content


related


coxxfidenoe.


While no


irical


wrk


exists


date,


it seems possible


to develop and


test


conceptual


linkages


between


different


types


knowledge


effects


other


variables of


interest,


such as recall and


recognition.


tMderatirn Effects of Knowledge Structure


have


considered


alternative


conceptualizations


different


types


product class knowledge and


their role


in consumer


search,


evaluation


choice.


It is appropriate


to acknowledge


that


manner


which


consumer's knowledge about a


product class


structured


affect


role


that


particular knowledge


item


plays.


This


proposition


follows


from


large body


of research


that


suggests


that


structural


organizational


factors


in memory affect


knowledge


activation and


thus


, how


it is used


in choice


tasks


(Bettman


1979


Biehal and Chakravarti


1986)


Johnson


zUsso


(1984)


find


that


product


familiarity


one


. A


secondary











brands


vs.


medi


ocre


brands.


searching


information or making


choices,


this


person


able


save


search


effort


by only


retrieving


brands


that


are


such


neory


category


only


considering


them.


concept


"evoked set"


(Howard and Sheth


1969)


is similar


to this


idea.


Russo


Johnson


(1980)


find


that at higher


inferential


levels,


knowledge


predominantly


brand


based


while


at lower


levels


mainly


attribute


based.


Sirce


conaer's a


external


choice


environment


is structured by brad,


that may facilitate using knowledge


that


structured


brand.


Thus,


consumer


forms


brand


evaluations,


these


should


facilitate subsequent choice


tasks and


this


type


of knowledge may be used more.


There


is, however,


some evidence


suggest


that


even


though


memo


structure


guide


information


processing,


consumers


can


sometimes


over-ride structural


constraints


when


task demands are strong enough


another


example of


(Biehal and hakravarti


the effects of knowledge structure,


1982)


Alba and


MPmnorstein


(1987)


have


shown


that


frequency


positive and


negative


attributes


brand


dominate other more significant


information


differences),


(siuc


attribute


when


conslurer


,rtance


degree of brand


motivated


is unable


process attribute


information.


Thus,


general,


one


argue


that


given


type of


knowledge


, differences


structure


late


the specific amy










different


situations--action pictures or portraits.


Differences


usage

may


situation


think


cameras


not be i

by brand,


aportant

e.g., "


for this constier,


and he or she


The Canon Sure-shot--for all


your


picture


taking


needs.


" In buying a camera for mse


in either situation,


brands may


second


recalled and


coiaser


there may be


little attribute search.


discriminate between


the usage situations.


Knowledge


structured


by attribute


for this consumer.


Thus,


action


pictures are to be


taken,


high shutter speed,


and having a motor


drive


would


relevant.


portraits,


a large-foramt camera


(i.e.


those


using


film larger than 35


would be needed.


This consider ay


start


considering


situation,


then


which


search for


attributes


are appropriate


information on


for the usage


those attributes.


Thus,


lack


of knowledge


leads


to ignoring differences


in usage situation,


brand-tbsed


structure


used and little attribute search way


Occur.


consumer


can


differentiate


between usage situations,


appropriate


attributes


identified,


prepurase


search


conduchrted by attribute.


simary,


depending


their


prior experiences,


cmsaers Gay


differ


in how their knowledge about a product class


is structured.


Such


differences


structure can


in turn have an i


Mpac


t on


the way a given


type


is used


to perform the


functions of


sernt,


evaluation and choice.


Moderating Effects of Other Bkg-romd Factors


addition


knowledge structure,


the role of


a particular type










pressure,


level


involvement etc.


In this concluding section of


this


chapter


, we consider some of


these effects.


study reviewed previously,


Johnson and Ruaso


(1984)


find


that


recall


learned


information shows a


task x


familiarity


interaction.


They


explain


this


finding


terms of


the effects of


familiarity on


search.


During


choice


task,


search


postulated


inverted-U-shaped


function of


familiarity


Highly familiar subjects are


selective


acquiring


information


search


less.


familiarity


subjects


have


little


basis


to conduct a search and also


show


less


search.


CkOnly the moderately familiar subjects have both


need


ability


to search.


In contrast


, during a judgment


task


greater the


familiarity the more easily new


information is encoded.


Thus


search


is expected


to increase monotonically with familiarity


recall


data support


this reasoning


. Thus


task


factors seen


to moderate


how knowledge


is used.


RPnj


Staelin


(1983)


assume


that cost and benefit assessments


underlie


autcaobile search


behavior


. They argue


that


if there


is a high


level

little


knowledge


gained


low.


of specific makes and dealers,


by search


If knowledge of


there


is relatively


, the expected benefits of


the product class


in general


search


is extensive,


search


have


greater


benefits because new information can be more


effectively


c

ended.


Mset


studies


search


behavior seem


to use


type of product as a


aore











less


information search.


Also,


patterns of


search


for such


products may


are


similar


across


people


than


durable


products


(See


example Newen and Staelin


1973


Brucks


1985)


Time


pressure


another


background


factor that can


iate


fwtction


of knowledge.


Consider a consumer who wants


to buy a stereo as


gift.


large


number


brands


are


available,


some sort of


screening


necessary


tendency to save search effort may be


accentuated


under time pressure.


For example,


a consumer ay be able to


save


search


effort


examining


ine


a pair of


correlated


attributes.


level


of consumer involvement may be another moderating


factor


consumr who


search


is not highly


evaluation


involved in


effort


a purchase may asnt


to reduce


search extensively


. This


consumer


cue


rely on some general


to the quality of


the product


prior knowledge


. This could lead


using price as


to the choice of


more


search


expensive


product


extensively


relative


Thus,


to when he or she


this


case


is involved


, a background


enough


factor such as


involvement may moderate how knowledge


is used in a casLmer task.


nmmar


ize,


this


chapter


examines


literature on consmaer


knowledge.


review


studies


that


take


both


aggregate


disaggregate


perspective


on consumer knowledge.


latter view


, which


distinguishes


specific


between


types


types of knowledge


knowledge


permits


search,


the study of


evaluation


the role


and choice.










potentially


more


useful


approach


to the study


of conuxxer product class


knowledge.


This


chapter


also


acknowledges


the role of


other


factors


that may


moderate


role of knowledge.


These


include knowledge structure.


role


badcground


factors


like


task


structure,


time


pressure,


search


costs and benefits and


involvement are pointed out.


Chapter


these


elements


are


combined


developed more


formally


conceptual


framework for studying the


role of


consuerr


knowledge.


bulk of


that chapter will


be devoted


to a discussion of


proposed


typology


consm uer


product


class


knowledge.


Each


type of


knowledge


typology


will


linked


functions,


using


examples


from


consumer environment.


Conceptually relevant


factors


that may moderate


these


relationship. will also be considered.
















CHAPI`W 3
m01BNI -R~~WL


preceding chapters examine


research


in which consxaer knowledge


studied


establishes


both


both


aggregate


lack of


di"ggregate


a rich disaggregate


levels.


taxoncoy of


The review


conanmr


knowledge


also a sparse set


irical


studies


* Thus,


there


is a


need


for a


taxonomy that examines different


types of


consner knowledge


their


functions


establishes


these


linkages


irically


Moreover,


explicit


consideration of


bkgrowund


task


factors as


well


as the


likely effects of knowledge structure would enrich


both


conceptual


empirical


contributions of


the research.


With


the above


mind,


conceptual


framework in Figure 3-1


is designed


to study


different


types


constrer


knowledge


can


serve


functions


search,


evaluation


choice.


This


chapter


next


presents


disaggregate


typology


consumer knowledge


that


is linked within


conceptual

information


role


framework

processing


background


various


scenar-los.


factors


functions


Where


structure


different


appropriate,


is considered


consaUer


the moderating


to provide a


tore


complete appreciation of


benefits of


adopting such an approach


to the study of


consumer knowledge effects.


Overview of


a Typology of Product Class Knowledge



























































Figure


Cacephal


Framewrk


Pwh- r,- P.Inc nr c- a aul~


IrPXX'I CLASS


S -RUCRE


PWRXUT CLASS

KNWLEDG TYE
aONTENT TYPES


a-'

IN:

SEARCH

EVALUATION

CHOICE


SBI[ONDARY
KFFULXTy :

RECALL

RECOGNITION

CONFIDENCE


BACSGROUND
FACUS:

TIME mSSU

LEVEL OF
IN VEMET


#P#-^*











work


propose


that an extended categorization of


consIUmer


knowledge of


a product class may be stated as


. Terminology specific


to the


Attribute values of brands i
Range of available attribute


follows:


product class.
n the product class.
values.


Correlations between attributes.


St&mary evaluations or affect
class


tweards


brands


in the


product


Importance of


attributes.


Usage situations appropriate


to the


product class.


these


types


knowledge,


four--terminology,


attribute


covariation,


importance)


criteria


usage


evaluating attributes


(similar to attribute


situations--are also proposed by Brucks


(1985)


Moreover


, these


hierarchical


can


taxonomy


also


related


suggested


seme


Russo


elements


Johnson


(1980)


Similarities


considered


differences


where


with respect


relevant,


to the other typologies will


continue our discussion of


typology.


seven


types


of product class


knowledge cover


information


that


typically


found


and used


in the consumer environment


kx occasion,


each


type


directly and externally available


to the consumer,


e.g.


terminology


brand


attribute


values


in medi-advertising messages,


seen


store


or in articles


such as


in Cansuer Reports


. In other


cases,


consumer may derive


the knowledge by processing


information


about


product


category


Thus,


range of


attribute values or


inter-attribute


correlations


class


can


learned by


vyami RI n O 1 rS


n cmi s~ naC n


0. w.V 1'tSSJ .- -


product


m krwr


w


I











order


further develop the meaning


of each


of these


types or


categories


product


knowledge,


we will


refer


to the brand-attribute


matrix


presented in Figure 3-2


While each of


iteas


in the


obviously


this


broader


example,


implications


it is nevertheless useful


than what might


be developed within


for illustrating


framework


being developed.


Terminology


relevant


product category


includes knowledge of


product

product


attributes


category


and also


the specific


usage


language used


of the product category


to describe


Thus


context


of our brand-attribute matrix


, "lens sharpness,


"shutter speed"


"aperture


size


" constitute examples of


producht-class


terminology


many


instances,


knowledge of


such


terminology may be viewed as


ability


associated


relate


with


these


attributes


product,


e.g.,


to specific


features or benefits


the relationship between


"shutter


speed"


and action


photography.


cell


entries


brand-attribute matrix


(Figure


3-2)


analog


knowledge


attribute values


in a product category


. These


raw


"information


items"


also


viewed


as the


"brand beliefs"


c mzponent


multi-attribute


attitude


model


(Wilkie and Pessemier


1973)


Knowledge


of the


range of


available attribute values


is onme


type of


distributional


information


that a consumer may possess about a product


class.


context


Figure 3-2,


it is viewed as


the difference


are


taxonomy
















ATiRIRInlW


Terminology
Range
Correlation
Sumary Attribute Importance
Summary Brand Evaluation
Brand Attribute Values
(Usage Situation is not shown)


- e.g. Attributes
- going down columns
- between ooltmus
-SAI
-SBE
- cell entries


BRANDS










end-points


range.


ConsLuers may know about


the density of


available


brand values


over different


parts of


range.


Knowledge


correlation


association


between


the brand


values


attributes


is another type of


product class knowledge.


Such


knowledge


specific


generic


product


class


e.g.,


(e.g.,


a price-quality association)


in our brand-attribute matrix,


with


better


lens


quality


scores


also


have


higher a am


shutter

values


values).


attribute


Knowledge


pairs


suxt associations between brand


ay be environmentally otmerved or


learned


directly by word of


mouth.


Suary

marginal v


brand


alues


evaluation


brands


thought


const.Ler' S


as knowledge of


awareness set.


In the


context


brand-attribute atrix


in Figure 3-2,


a consumer could


integrate


available


information for each brand and come up with an


evaluation


picked


brand.


uord-of-mouth


Alternatively,


the evaluation may have been


from reading Consumer


Reports and


consumer


not have an


underlying


integrated structural


knowledge of


brand-attribute


values.


Attribute


importance


viewed as


the marginal


values


that a


cons Bmer


place


attributes when evaluating brands


in a product


category.


These


take


form of


value weights applied


to beliefs


expectancries


about


brands


multi-attribute


modelling


framework)


have a rich


underlying structural


knowledge of


cameras


speed










distinction


between


weights


scale


values


is fairly clear


it has


often


been


difficult


to empirically distinguish


between


them


(Anderson


1981


, 1982)


Recent


research


also documented


the importance of


situational


factors


consumer


behavior


(Belk


1974


Kakkar and Lutz


1981


Dickson


1982)


This


category


knowledge


pertains


judgments


appropriateness


category


of brands


Alternatively


for specific


knowledge


purposes

usage


relevant to the

situations may


product

link to


knowledge


attribute


importance


or situation specific summary brand


evaluations


Srivastava


points


made


, Alpert


above


and Shocker


illustrate


1984).


important consideration


taking


disaggregate


view of knowledge


Clearly


at some


level


these


different


types of knowledge may be


linked in a cognitive structure and


derived


from


each


other


Nevertheless


, an attempt


to disentangle


them


functionally


would


lead


better understanding of how


they


might


guide consumer


information


processing


functions


in choice.


Figure


overall


conceptual


representation


this


very


critical


point.


following


section


develops


this model


for each


identified


element of


our knowledge


nents of


taxonomy and illu

the Product Class


states each

Knowledge T


with an example.


ypology


Terminology


Brucks


(1985)


states


having


knowledge


relevant


terminology


"makes


easier


questions


interpret


their


are


Compo










evaluating


brands


On the other hand,


a consider who knows


relevant


terminology


can


understand


ring of


ccmzication


aessatg


product class


information available


the environment.


consumer


unfamiliar


with


terminology relevant


cameras


either


limit


search


turn


other sources such as


friends

aperture


or salespersons


size.


for the meaning of


contrast


terms


a consumer who knows


like


lens sharpness or


the meaning of


these


terms


can make sense of


advertisement


information about camera models


store


fact


presented


search


in an


more


independently


Moreover,


this


latter consumer may use knowledge of


more


advanced


terms,


such as motor drive and manual


over-ride capability and


search for


information about


these more advanced features.


Bterminologys'

terminology


(1985)

and a


typology


Available


separate categories


attributes


In contrast,


for knowledge of


as envisaged here,


terminology


can


include


meaning


attributes,


and also


terms


relating


product


i us ed.


For example


, the user of


camera


should


know


meaning of


"load"


"rewind"


film.


Knowledge


by a person


attributes and


to formulate questions,


the way a camera is used will


interpret new


both be used


informtiom.


Brand-Attribute


Values


BAVs)


.Consuars


possess


basic


or "raw"


knowledge of


the value of


brand


product category on any particular attribute


Examples are


"The


Pen tax


K-1000


a price of


$150"


or "The Minolta MBxxua costs










for

each


further

other


evaluation


C-mpared


brand.


other


These values can be


criteria


order


related


to form brand


evaluations.


example,


price


Pentax


camera


may be


criterion


that attribute


(e.g


SIR cameras usually


cost


over


$120),


brand-attribute


value


another


alternative


(the Minolta,


in this case).


Brand


Attribute


Value


lowest


category


in a hierarchical


typology


of consner knowledge


proposed by Russo and Johnson


(1980)


they


point


out,


the processing of brand-attribute values


riso


with


each


other


using


elements of


processing heuristics


(Bettman and


Park


1980)


generates


more


cap


forms of knowledge.


Alternatively


ca p


information


obtained


frcm


the envi rert e.g


frcm


advertising


asking others


for recommendations


, recognizing that


such


knowledge


generated


same


prior


information


processing


conducted by the source.


Brand-attribute


values


also be used in conjunction with other


forms


knowledge


to generate


information


pertinent


to the consumer's


task.


Thus,


consumer


knows


price of


a brand


in a product


class


can


make


inferences


about


quality


basis of


knowledge of


a price-quality


correlation.


Attribute Range Knowledge


conssU r


know


that


power


output


for better stereos


intended


home


use


is between 20 and


100 watts


. This knowledge of




orjnnnBf~n











further


search


be more


likely


information.

if the consumer


Use of


range knowledge


is unfamiliar with


in this manner

product class


difficulty


understanding


meaning


given


item of


information.


Consider


a consumer who


is unfamiliar with


the terminology


relevant


characteristics


stereos.


not know the meaning of


output


power,


output


power


related


to the way a stereo


sounds.


seeing


an advertisement


that says


"The


Technics SX


615 has


output


watts,


this


consumer


find


difficult


comprehend


meaning


information


given


in the ad and may


search


other


brands


output


power


to establish


the context


for an


evaluation.


contrast


a consumer who


knows


that


the available


range of


power


output


values


20 to 100 watts can


use


that


information


to generate a


provisional


evaluation.


Thus,


stereo with 80 watts output may be


seen


being


quite


good


that attribute and


this evaluation may


guide


further


search


. The alternative may be searched further on other


attributes


price


, warranty


terms


distortion


order to


conclusively establish


its overall


acceptability


mechanism


which


range


knowledge


impacts


consumers'


evaluation,


search


and confidence


is proposed


in Assar and Chakravarti


(1984).


This


associated


1saljKs


that warrant


further research are


discussed


Chapter


A subsequent


section of


this dissertation will










Attribute Correlation


Consumers


learn


about


existence of


associations


between


attributes


products


froi advertising


stimsli


,by


experience or


from


other


people.


Some


correlations may


be general


and perceived


to exist


across


product


classes


(e.g.


"Higher


priced


products are of better


quality")


Other


correlations may be


specific


to a product class


(e.g.


"Higher priced cameras


have


better quality").


When


the correlation between


two attributes


is known


conslers can


this


make


predictions or


inferences about


information


that


missing


to be collected


. Thus


Brtuks


(1985)


suggests that


knowledge


attribute covariation enables


infe -


to be made about


attributes


without


external


search


. If information on one attribute


missing


the environment,


or is forgotten,


brand values on


the other


attribute


used along with


the correlation


to infer the missing


value


without


correlation


search


searching


, thus,


information


used


information.


to reduce


acquisition


Knowledge


the effort associated with


context


of a choice.


particular


, excessive search costs


or the


presence of


time


pressure may


accentuate


cons BIer a


tendency


to use correlational knowledge


the manner suggested above.


Similarly


consumer who


lacks a high


level


of involvement


in the


task


hand


may seek


to conserve


search effort


by only


examining one


pair


of correlated attributes


Clearly


, the use of


such a search


use











consumers


the specific way


in which such knowledge


impacts


upon


search,


evaluation and choice.


Summary Brand Evaluation


The

clearly

such k


stored

important


knowledge


overall


evaluations of brands


components of


statement


in a product category are


product class knowledge.


such as


An example of


"Nikon cameras are very good.


Cansumers


obtain


such


evaluations


from external


sources such as


advertising


nationss


others


can


hold


such


evaluations


without


underlying


knowledge


base of brand-attribute


values.

sugary


contrast,


more


evaluations


experienced


modify


external


olsytderis

ly derived


may form

MIes,


their own


based


specific experiences.


- L


brand evaluations


for one


brand may have effects on


extent


search


alternatives


are


mediated


specific


evaluated.


underlying


Moreover


alternatives


the nature of


categorical


structure


other


these effects -ay


in which such


sury brand


evaluations are stored.


Sutary


brand


evaluations


jmpac


t on


the extent of


search


in a


number


ays.


First,


brand


which


hasa


very favorable saM ry


evaluation

without


(e.g.


extensive


Nikon)

search


may

for


be chosen based on


information.


stored evaluation,


This may be more


likely given


external


conditions


like


tine pressure.


However,


an accessible stored


overall


evaluation


make


less


likely


that


individual


'ecMmer











Second,


overall


evaluation


brand


form


basis


inferences


about brand-attribute


values


that are missing


in the context


choice


task


(Dick


1985


Thus,


aperture


size


for a


particular


model


featured


in an ad


, it may be


inferred


to be


large,


consistent


with


favorable evaluation.


Alternatively


when a


conasLiur


in a store and at


temp


ting


to retrieve specific features of


brands


previously


encountered in another store,


an inference based on a


prior evaluation ray form the basis


for the


judgment.


Third,


other


affect


alternatives


associated


that


with


judgments


a brand my be


choices


transferred


are ade without


search.


example,


a mcara such as Nikon,


that


is categorised as a


good


brand


favorable


overall


evaluations


Nikon


models


while new models of


other brands may be


inferred as mediocre


, if


such a brand categorization exists.


Finally


* a


poorly


evaluated


brand


be ignored completely or


from


candidate


examined at all


during


search.


Moreover


, an explicit


infererxe


if aide


, for a mediocre brand


likely


mediocre under evaluative consistency processes


(see


Lingle and Ostrom


1981).


Sary


brand evaluations may also


impact on


information


processing


strategies


evaluation.


First,


favorable sumaary evaluation


exists


brand,


attribute


values


ay be used as criterion


levels


during


search


attribute


values of


other alternatives


in a


produce


dropped











consumer


may not do


pairwise


feature comparisons


as is implicit above.


Rather


, other


alternatives


be evaluated


in wholistic


fashion and


these overall


evaluations compared


to that of


the memory brand.


Thus,


brand


preceding


evaluations


discussion


guide consumer


suggests how knowledge of


information processing


overall


in search,


evaluation


choice.


While


these


knowledge


functions


are


very


important


deserve


critical


enquiry


, empirical


work


in these areas


is outside

Summary At


the scope of


tribute


this dissertation.


Importance


Knowledge


attributes


clearly


value


or the


critical


"importance"


for judging


brand


to attach

s in a pr


to relevant


oduct class.


Thus,


statement


"The


most


important


thing


a camera is


quality


lens


clearly


y priorities


the critical


factor


in the


choice of


a camera.


most models


of multi-attribute


judgments


, it is assumed


that


evaluation


attribute


level


can


be distinguished


from


the weight


given


that


attribute


that


when


alternatives


evaluated,


importance


held constant across


the entire


(See


example


, Anderson


1981,


1982)


These


"importance


weights"


that are


learned


from various sources


, may be abstract judgments of


value or may


derived


from


understanding


relationship


between


feature


a product and


the benefit associated with


possession of


that


feature.










attributes


for search


from a


larger set


Attribute


importance may also


used


sane


combinatorial


rule


for evaluation.


However


, Brucks


proposes


that


such criteria may be used as


reference points


, to decide


whether


alternative


acceptable.


this


sense


judging


alternatives,


overall


evaluations


criteria

rather


are cutoffs on

than attribute


attribute values,

importances in th


or on

true


sense of


term.


Sane


consrers


lack a basis


for differentiating


between attributes


Their


search


process


aore


contingency


driven


or "constructive"


Bettman


1979


Bettman and Zins


1979;


Biehal


and Ohakravarti


1986


attribute


importances are


learned by


experience,


this knowledge


used


tentatively


initial


stages


As consumers


improve


their


understanding


of the


relationship between attributes and product


performance,


there


may be more variance


the extent of


search across


attributes.


Thus,


knowledge


attribute


importance


is a basis


selectivity during


information search


Specific


patterns


attribute


importance


weights


may manifest


themselves


heuristics


used


evaluation.


Thus,


sharply


different


importance


weight


structures


produce


greater


use of


non-compensatory


strategies


search


(Biehal and Chakravarti


1982)


Such


effects are more


likely


in the


presence of background factors such


large


number


of alternatives,


time costs,


etc.


These


bckgrouzd


factors


their


effects


on search heuristics have


been extensively










e Situation


situation


which


product


intended


used


sometimes


important


consideration,


affecting


the pattern of


search


also evaluations.


Knowledge of


the usage situations appropriate


brand


prompt


matching


the product/brand


to situational


requirements


features


perforumrce


(Brucks


1985)


Situational


variations


mke


different attributes aore or


less


important and


also


affect


appropriate


criterion


levels


for those attributes.


Thus,


marketing


literature,


situational


factors


their


interactions


with


product and person factors are an


important basis


market


segmentation


(Dickson


1982;


Srivastava,


Alpert


Shocker


1984).


Consider


two different


usage situations


for a camera--taking action


pictures,


taking


portraits


Knowledge


situational


appropriateness


brands may


lead


to the choice of


a Nikon


for action


pictures


cases,


whereas


the action


Yashica


be preferred for portraits.


picture situation may pr


t consideration of


In other


shutter


speeds


whereas


size


film


more


important


consideration


portraits.


structure of knowledge affects


both


how information is searched as well as what


information


is searched.


final


illustration


of the


importance of


usage situations has


with


social


aspects of


consamption,


particularly for products such


food,


beverages and clothing.


In social


contexts,


social


influences










situation.


Thus


social


aspect


a situation


is one of


several


background


factors


that moderates


the search


evaluation and choice of


brands


product


category


sm8e


ways,


this


reconceptualization


the normative pts of


a conmstuer's belief


system for a product category.


Distinctions


from Other Knowledge


Typologies


proposed


typology


builds on Brucks'


(1985)


work


but does have


several


distinctions


from


typology proposed by Russo and Johnson


(1980)


will


explore


sm f


these


differences.


Russo and


Johnson's


typology


designed


explore


conaser


knowledge


familiar


products


Their protocol analysis reveals knowledge at several


"inferential


levels"


: (1)


Best Brand;


Whole Brand/Attribute Ranking


Comparison;


Whole


Evaluations


of Brand/Attribute;


Single


Pair


Ccaparisons/Rankings


brand


attribute


Single Evaluation


(Brand Attribute


Value).


First,


since


Risso and Johnson


focus on


"inferential


levels,


" they


clearly delineate


between


the different categories of knowledge


function.


Level


, for example


, includes


rankings or


compa


prisons of


both


brands


attributes.


might


argue


that


the st


ry brand


evaluations


attribute


i-portance


categories


develop


brand


versus


attribute knowledge distinction while collapsing what seems


like


arbitrary


distinction


between


Russo


Johnson' s


inferential


"levels"


2 and










distinction


would


made


based


degree


goodness--the


first


example


level


(Best Brand


whereas


the second


is at


level


(Whole Evaluation of Brand).


Second,


knowledge


category termed


"Brand-Attribute


Values"


typology


are


statements


about


level a brand


possesses on an


attribute


(e.g.


"The


Pentax


K-1000 costs $150")


bamso and Johnson's


lowest


inferential


level


(Single


evaluation--also


termed


Brand


Attribute


Value)


basic


statement


fact or perception.


Alternatively,


K-1000


focus


is inexpensive")


our typology


have

. Thus


an evaluative campxnent


(e.g.


as the above distinctions


is to discriminate


between knowledge


"The Pentax


indicate


types based


functions


that


they


play


choice.


In contrast,


RMso and


Johnson


(1980)


between


knowledge


types on


the basis of


inferential


levels


information


processing hierarchy


Further,


given


difference


objectives of


the two


typologies,


it is


understandable


that attribute


range


, inter-attribute correlation


, usage


situation


Johnson's


terminology knowledge show n clear mapping to Russo and


typology.


Functions of Knowledge


in Consumer


Information Processing


references


which


functions


used by a consumer


knowledge denote


the way in


. The various knowledge elements


in our


proposed


typology


serve


different


functions


in search


, evaluation and


choice.


This


categorization


permits a clearer enunciation of


the role


our


discriminate











missing


particular


information


role


can


that


made


given


from


type


in evaluating


knowledge


brands.


plays depends on a


variety


background


factors


, including structure


At the same


time,


also clear


that different


types of knowledge can also be used


perform


same


information


processing


fiction.


Thus,


both range


knowledge


prior


suarary brand evaluation can


be used


to form a


provisional


evaluation of


a new piece of


information


for a brand.


concluding


this


section,


develop


an inventory of


general


propositions


regarding


how different


types of knowledge may be used in


different


functions.


First,


search,


knowledge


used


make


sense


information,


enable


coinsmr


understand


meaning


attributes


terms


related


usage.


these


CBses,


knowledge


search.


eliminating


information.


other


need


These


cases


ccmp


effects


, knowledge redire


lete


search


search effort by


exhaustive acquisition of


patterns nay 'be acxentuated or


attenuated


background factors such as


time pressure


or a low


level


involvement.


Thus,


different


types of knowledge can sometimes have


opposing effects on search.


Secawx,


evaluation.


in evaluation


Thus,


,knowledge


provides criteria and contexts


terminology comprehension is low,


range knowledge


provides


context


evaluating


given


brand-attrikrte


value.


a


evaluations


can


used as a criterion


to ccmpare other


increases


brand











sUmary


brain


evaluation


may form


basis


for choosing a


brand


.The


chosen


brand


either be objectively superior


, the degree of


affect


very


positive,


both conditions myy


exist


Clearly,


srch a


choice,


the basis of


an overall


evaluation also significantly


reduces choice processing


effort.


stated


earlier


irical


portion


this dissertation


examines


use


range


correlational


knowledge


in search,


evaluation


choice.


next


chapter


reviews


literature


pertaining


range


knowledge


its effects and provides examples


from


conssumer


environment.


This


literature review motivates


irical study to be


reported in Chapter 5.


ade















CHAPIER


KNOWLXIE OF


ATTRIBUTE RANGE


IN A PRrDUC CATEGURY


EFFECTS ON EVALUATION AND SEARCH


indicated


in previous sections


, this chapter elaborates upon


inner


in which a consumer's knowledge of


range of


attribute values


available


evaluations


product


choices.


category


We will


affects


first


information


review the existing


processing,


literature


related


fields and


then develop a


proposed mechanism by which range


knowledge


affect


consutner


search


, evaluation and choice.


We will


also


forth


formal


propositions regarding these effects


that


were tested in


the empirical


study described in Chapter 5.


motivate


propositions


that


follow


consider the case of


prospective


car-buyer


is confronted with


information regarding


gas-mileage


rating


compact


car


Moreover


suppose


that


this


rather


information


naive)

n and


consumer is not

knows little about


abreast


the available


recent

range of


fuel-economy

gas mileage


values.


This consumer may be unable


to evaluate


information saying


that


mileage


rating


of the car


is 40 m.p.g.


However


, if he or she


knows


that


caopac


t car gas mileage ratings


range


from 32-40 m.p.g


m.p.g.


value becomes meaningful and


the car's gas mileage rating can


coded


"best


available"


product


category


. This


preliminary


evaluation


would


qualify


model


a viable choice











level


brand A


is lowest on


, a smoker may evaluate


cigarette


effects


favorably


of that


level


without


taking


tar content


account


Moreover


actual


, if brand B


health


is higher on


tar,


will


seen as wnrse on


that attribute and


the conrmier my


form


negative


overall


evaluation of


, again without


regard


to the


actual


health


effects.


this


situation,


presenting


rank


order


information


within


range


values


provide


basis


polarization


evaluations,


without


recognition


equally


deleterious health effects of


brands.


these


examples


show,


range


information


naturally


available


consUmnrs


their


environment.


This


information


strongly


influence


how consumers


form evaluations of


new


information and how they then search


for further


information in


context of


making a choice


in a product category.


Social


t Research on Range Effects


Research


social


judgment has


found


that


judgments of


a stimulus


affected


values


other


stimuli


present


the judgment


context.


These


other


stimuli


can


used as a basis


for comparison


(Sheri f


, Sherif


Nebergall


1956)


Suppose


a camera is


priced at


$150.


judgment


about


the price of brand A may differ


, depending on


whether


judgmennt


is made


in the context of


other brands


priced at


$100 or $200.


Knowledge


range


attribute


provides


one


type of


are


xaenn


memory










have


been


termed


"contrast


effects"


Sherif


and Hovland


1961


Ostrao


Up-haw


1968)


Thus,


stimuli


are


likely


to be


rated as more


similar


given dimension when


range on


the dimension


is wide,


relative


when


the range of


values on


that dimension


is known


to be


narrower.


Extending


our previous example


consider two camera brands A and B,


priced


$150


$200


respectively


the prices


for available


cameras


range


between


$100


$500,


then A and B are


likely to be


judged


more similar


to the case when prices range between


$150


$250


When


the wider range


is used as a context


, the stimulus


judgments


shift away from the upper end of


the range which serves as an


anchor


. This


, in effect


, is the well-known contrast effect.


B are seen as similar


, the consumer may perceive


them as


comparable


alternatives


that


are


in the


same


"ballpark"


on price.


narrOW


range


context,


A and B nay be seen as


less similar and


consumer


categories.


perceive


Thus


them


knowledge


falling


attribute


different


ranges


price


lead


differences


perception.


These


perceptual


differences


in turn may


affect evaluations and subsequent choices.


Range


effects have been


interpreted


in two ways


One explanation


that


those


effects


truly


reflect


changes


hw a person actually


thinks


about stimuli


* i.e.


, they are effects on mental


representations.


Others


have


argued


that


these


effects


changes


inhos~


, compared


in how










important


to distinguish between


these explanations


for any


specific


study


range


effects.


Chakravarti


and Lynch


(1983


have


reviewed


this


controversy


irical


evidence.


Their


review


shows


that


weight of


the empirical


evidence


is that


range effects


occur


response


language


, and


that


the effects are not carried over


other


judgments


to subsequent


behavior


While


representational


effects


are


(Sherif


intuitively


Hovland


plausible


1961


Helson


discussed


1964),


literature


they have been difficult


demonstrate


irically.


Range


effects


a representational


nature may be more


likely when


subjects


are unable to


camp


'rehend


the dimension of


judgment


(Hutchinson


1983)


This


viewed


a knowledge effect


that


represents an


interaction


between


terminology


knowledge and range knowledge,


two of


elements


in our taxonoay


Alternatively,


such effects may occur


ju gments


that


require


trading off


incormensurable dimensions


(Mellers


Birnbaun


1982)


These


conditions


which


favor


representational


effects


will


considered in more detail


when we discuss a eanisa


by which range


information may


influence


search.


When


prior


response


language driven rating


is nade salient on a


subsequent


occasion ,


that


rating may serve as a basis


behavior


. In


studying


contrast


effects


, Sherman


, Ahlm,


Beraan and Lynn


(1978)


found


subjects


rated


importance


recycling program higher in


context


trivial


issue,


and lower


in the context of


an important










Ecoomics Research


on Range Effects


economics


literature


, the dispersion of


alternative values


attribute has


been seen as a determinant of


search behavior


. The


models


typically


consider


dispersion


prices .


Stigler


(1961


price


dispersion


as a measure of


ignorance


in the market


consumers


are


well


informed,


sellers'


prices


are


likely


widely


dispersed.


sellers


dispersion


order


is large


lowest


, it will


price,


pay to search several


as long as the expected


saving


higher


than


cost


search.


amount


of search


conducted


to get a


lower price


is positively related


to the


variability


prices.


Rothschild


(1974)


proposed


that


with


increasing price dispersion,


intensity of


search


ir creases.


As prices became more dispersed,


expected


total


costs


including price and search costs)


will


decrease.


Kohn


Shavell


(1974


developed an analytical model


of the behavior


individual


price)


searching sequentially


distribution


. The attribute searched


which was either known with certainty


(e.g.


(the


static


case)


known with certainty


(the adaptive case


Search


continued


until


individual


found


alternative


for which


utility


was


greater


than


"switchpoint,


defined as


"the


level


utility


which


expected


immediate


cost


continued search


equalled


expected


future


gain


search.


" The


location of


switchpoint


as


related


to the characteristics of


the distribution, in


saw










latter


(adaptive


situation,


searching enabled


individual


learn


(risk)


about


distribution.


associated


preserving


mean


with


utility)


same


static


token


distribut ion


the switchpoint


increased


if the variability


B increased

, leading to


more search and compensating


increased risk.


Stigler


(1961)


proposed


two different search strategies--one where


consUmer


decides


number


of draws


before starting


(fixed


sample


size)


other


being sequential


search .


Gastwirth


(1976)


analytically


explored


optimal


number of


searches varies with


assumed


price


distribution


and search costs


With a


fixed sample


size,


optimal number of


searches


increases with search costs.


Chly


when


search cost


very


small does


the optimal


number of


searches vary


greatly


with


optimal


the

number


price distribution.


searches


However


, for sequential


is much more sensitive


to the


search,

fora of


the assumed price distribution.


Axell


1974)


proposed an adaptive model


for how consumers and firms


search


market


durable


products


(cars)


According to


this


model,


mean or variance of


price distribution


is not known


consumerr,


some


initial


estimate


of the


variance of


price


used


. This estimate is


updated as prices are sampled during search.


Wilde's


(1980a)


model


focuses


on search of


alternatives


. In this


model


, the


consumer


is assumed


to rely


on a subjective estimate of


joint


distribution


prices and quality


levels


that are available


I











Clearly


even


this


selective


review


the economics


literature


establishes


that


extent


search


postulated


to be greater when


there


However


attribute


more


, little


variability


empirical


affects


on an attribute,


work


behavior


range


is wider


focused on how variability


consider


in search,


on an


evaluation or


choice.


Studies


marketing


eco omani cs


evidence


variability


that consumers


attribute.


Though


information have


search

range


been cited in


behavior depends on


attribute


is not


explicitly


treated


this


literature


, the ranges of


price


(or other


attributes)


can


considered as one aspect of


the variability of


distribution of


an attribute.


Marketing Research on Range Effects


effect


attribute


variability on consumers'


search


behavior


been


irically examined


in two studies


Meyer


(1982)


modeled how


consumers


update


their


expectations/perceptions


about


alternatives


during sequential


search


He proposed


that


the probability


inspecting


alternative


is a


function of


the conslar's expected


value


or mean


for each attribute


, and


the variability of


each attribute


as perceived by the consumer.


Subjects


Meyer


1982)


study searched for


information about


apartments.


They were


given


one


item of


information


the value on


attribute


range


values


another


attribute


manipulated,


thus operationalizing variability


. The study


found


that an


one


was










with


Thus


probability


of searching alternatives


is higher


wider


ranges of


the given attributes.


Goldman


Johansson


1978


used


economics of


information


framework


develop


test


hypotheses


about


consuDers'


search


behavior


gasoline


positively


. They hypothesized


related


perceived


that


the propensity to search


variability


prices,


negatively


opportunity


costs of


time and


to the


importance of


choice of


filling station


. The data


in this study were


taken from a


panel


gasoline


. The extent of


price variability aes based


consumers


perceptions


similarities and differences among


brands


based on


price


levels and


price


fluctuations.


authors


found


that


perception of


variability


in prices had no


significant


effect


propens i ty


search.


As an explanation,


authors


proposed


that


monetary savings of


search are small


econanic


factors


ones.


have


However


little


extent


explanatory


price


power


, mp*e


variability


to non-economic


in this study was not


stated.


level


variability


, consumers would not expect


save


much and


they will


search


less


Further


, consumers may wish


well


known


brand or expect competent service


. These non-economic


factors


unke


choice of


gas station


to be more


important


than


desire to save


, and search may


be reduced once


the consu ter


finds a


station


with


reputation


maintain.


Consistent


with


this


explanation,


authors


found


that


a variable measuring


importance of


purchases










While


postulates


empirical


that


evidence


amount


is mixed

search


, the conceptual


positi


work clearly


related


variability


attribute


Coupled


with our


intuition


regarding how


consume rs


use


range


knowledge


, this evidence suggests


that


range


knowledge


search


play


an important contextual


, evaluations and choices among


brands


role


in a


, affecting consumers


product category


Range Knowledge


in Sequential


Search


the effects of


range knowledge on search and evaluation have


been


treated


separately


However


, range knowledge may produce dynamic


interplays


between


evaluation


search .


many


consumer choice


situations,


judgment-relevant


information


available


simultaneously


Rather


, information may be gathered or become available


sequentially


, and


consumer


have


decide which


pieces of


information


should be gathered.


economiUcas


literature


, there


is considerable support


idea


consumers searching sequentially


Early models assumed a known


underlying


distribution


discussed


searches of


fixed sample size.


However


more


recent mWdels


acknowledge


that


the distribution of


prices


typically


known,


that


searchers


can obtain


information


sequentially


(Stigler


1961)


Consequently


these situations


opt inal


search


pattern


adaptive


nature.


Co mse rs


obtain


information


sequentially


, learn


distribution of


alternatives and


decide


much


search


appropriate.


See,


for example,


Kdhn and











more


general


idea


that


when


information


is sparse


sequential


search


enables a consumer


to learn about


the available options.


nature


ability


interpret


further


search


the initially available


be mediated by


information.


the consumer' a


For example,


meaning


attribute


is not clear


, range knowledge may be


used


context


evaluate a given brand-attribute value.


If the


brand-attribute


value


suggests


brand is


best on an attribute


, it


be examined on other attributes


If the given


brand


is evaluated as


mediocre or poor


other brands may


be searched.


Attribute Range Knowledge:


Effects on


Evaluation and Search


now


move


formal discussion of


the mechanism


by which we


propose


range knowledge may


impact upon


evaluation and search.


Assar


and Chakravarti


(1984)


examined


the role of


range


information


evaluation


search


. They provided a mechanism for


range effects


conducted


preliminary


study


that


provides


basis of


empirical


work


to be presented


in this dissertation.


Consider


consumer examining single


lens


reflex


SIR)


cameras


. If


confronted


find


with


difficult


attribute


make


siuc


sense


as lens quality


a brand-attribute


, the conser may


value if


attribute


unfamiliar


other


hard,


even


this unfamiliar


consUmer


may be able


to evaluate a given camera'


lens quality score


range


provided


index


lens


quality


Thus,


range


i n fnrmsti on


ai d nrnvi i onal


hv nrnrvidina rA nnt.vt. Enr


vn Iaint i nna


ann


r










Range


information


permits a preliminary


evaluation of


a brand-attribute


value


encountered


during


choice.


This


preliminary


evaluation


mediate


subsequent


search


Thus,


preliminary


evaluation is


favorable


, the


brand


examined


further


(1981)


ymprool


that


standards


used


simplifying mechanic


evaluating


information.


Brand


alternatives


that


appear


attractive


initially


likely


be examined


further


In contrast,


if a brand


appears


unattractive


one


or two dimensions


, reasonably that


brand


will


chosen.


In the


latter case


, the consrtr may search other


brands


same or different attributes


Hence


range knowledge


affects


searTh,


there


ultimately


be effects on


the choices


that


consumers make.


Assar


information


Chakravarti


have


(1984)


also discussed


confidence


the effects


judgments


that


choices


range

. For


example,


absence


the context


provided by range


knowledge,


constumrs


lower their evaluations


Cespei


isate


the associated


uncertainty


the consumer


Moreover


is less


such


familiar with


effects are


likely to be accentuated when


the dimension of


judgment.


Assar


their


and

study


Chakravarti


This


(1984)


ensured


used SIR cameras as


that


the product class


their student subjects did not have


much


extra-experimental


knowledge


about


the meaning of


attributes and


range


values available


in the envircirat.


This


permitted


them


to experimentally manipulate range knowledge.


lynrh











told


that


range


of attribute


values


was consistently


three


times


wider

the


than


same


in the narrow range condition


center value


the scales


In the third condition


being anchored at


, subjects were not given


any range


information at all.


subjects


first


evaluated


profiled brand.


lens quality


index


score


this


brand was placed at


, towards


the upper end of


narrow


range


(30-45)


but not as


high on


the wide


range


(15-60)


was


expected


that


this


profiled brand would be evaluated ost


favorably


narrow


range


condition,


less


favorably


wide


range


condition,


least


favorably


when no attribute


range knowledge was


available.


associated


Subjects


with


were


their


also


asked


evaluation.


provide a confidence rating


expectation wcs


that subjects


would


more confident


when range knowledge was available


relative


when


it was missing.


second


task,


requiring search,


subjects were given a single


"seed"


was


brand-attribute


presented


in a 3


value,


brand


the lens quality


x attribute matrix


index of


(see


a brand


Table 4-2)


. This


.The


lens


quality


index value given was


45--the best available


value on


narrow


range.


the narrow range was used because


expected


effects


search


were


believed


to be most


likely at


this


level.


Without


range


knowldege,


consumer


expected


to have


difficulty


evaluating


given


value with confidence.


Hence


, the


consumer


expected


to search


for other


brands


, to develop a context










Table
Range


Pretest--Evaluat ion


Subjects were given


Task


this camera


profile:


Max.


shutter speed


index


= 1400


Price


Warranty


months)


Lens quality


index


Rated


the desirability


of the profiled camera on a


= extremely undesirable


Rated confidence


9-point scale


= extremely desirable)


in this desirability rating


= not at all


confident


= extremely confident)


Attribute Range Knowedge
Conditions Only


Provided


in the


Wide and Narrow Range


Attribute


Range Cond it ion


Wide Range


Narrow Range


Maximum Shutter Speed
Price ($)


500-2000
150-600


1000-1500


300-450


Warranty


(months)


Lens Quality


6-24
15-60


Index


12-18
30-45











Table
Range


4-2
Pretest--Search


Task


Subjects were asked
stocked 3 brands (A, B,
as a search attribute).
The "seed" value given
shown.


to
and

i was


imagine
C), all


they were
priced at


lens quality


index


in a camera store which
$400 (to eliminate price


for B


= 45


in matrix as


Max.


shutter speed index


Lens quality


index


Warranty period


Subjects indicated which
checking the appropriate cell
Subjects were asked to d
(Note: They did not actually g
Finally, they were asked
they wanted and how that would


piece of information
on the matrix.
describe how this new
et a new brand attribu
to indicate a second
be used.


they wanted next


item would be used.
te value).
piece of information


__45










favorable


provisional


evaluation


can


be made without searching other


branxis.


Hence,


as expected


that


the conumaer would be are


likely


search


given


brand


other


attributes,


since


it would be


identified as a


likely


choice candidate.


wide


range


condition


to 60),


the provisional


evaluation


was


still


expected


favorable but unlikely to be as


extreme


before.


consumer' s


search


patterns


would


less


predictable


this


instance.


The same


brand could be examined


further


lens quality scores of


other


brands may be searched


to locate a


better


brand.


Alternatively


, even


other


brands might


be searched on


other attributes.


Assar


Chakravarti


(1984)


found


that


, as expected,


evaluation


higher


the narrow and wide range conditions and significantly


lower


have


brand's


conditions.


the no-range condition


subjects


evaluations


Also


This


discount


were


S as


suggests


their


different


expected


confidence


that a lack


evaluations.


of context


profiled


the narrow and wide


ratings iere


range


lower and


variability


evaluation


was


higher


absence


range


information


, relative


to the


two range conditions


. Thus,


the absence of


a context


led to uncertainty


in evaluation.


analyzing


search


sequences


subjects who searched


remaining pi

as following


eoes of brand B


purely


information


brand-based sea


(see Table 4-2)

rch pattern. Th


were classified


ose who searched


was











range


condition.


There


was


difference


between


extents


brandwise


search


in the wide


range condition and


the condition where no


range


information


was available


Attributewise search as hypothesized


most


likely


in the no


range case


, less


likely


in the wide


range


case


least


likely


in the narrow range case.


Here however


, the range


manipulation


not have a significant effect,


though


the proportions


were ordered as predicted.


data


from


this


preliminary study supported


the key


elements of


proposed


mechanics for how range knowledge affects evaluations and


search


behavior


. The


availability


range


information


may enable


subjects


form


preliminary


evaluations


which


in turn may mediate


search.


Issues


for Research


provocative


findings of


the preceding study suggest additional


research


down


exact


nature and locus of


range knowledge


effects.


While


the study showed how range knowledge-induced


effects on


provisional


evaluation


could mediate search


(a behavioral


criterion)--


several


design


limitations


constrained


more


explicit


inferences


regarding these effects


These are discussed below.


First


there


is a need


to examine alternative specific accounts of


range knowledge


influences evaluations.


Two competing acxaonts need


resolved.


one,


when range


information


is missing


consumers


experience


uwcertainty in evaluating a brand and may compensate by











evaluated


less


favorably


the absence of


range knowledge


than when


present.


However


, a


brand


with


unfavorable


characteristics may


receive


somewhat higher


favorable evaluation


in the absence of


range


information.


This


suggests


that


absence of


range knowledge,


evaluations may


gravitate


toward


the middle of


the response scale.


study


using


brand-attribute


values


spanning


range


dimension


will


allow


separation


between


the explanation


invoking the


discounting


evaluations


versus


the case when subjects are tmwilling


make


extreme


evaluations


When


range


information


is not given


, the


pattern


ratings


low brand-attribute values will


be diagnostic.


discounting explanation would suggest


lower evaluations


relative


those


wide


narrnow


range conditions


(n the other hand,


higher


evaluations


these


brands will


support


the explanation


that


subjects are unwilling to make extreme evaluations.


Second,


the Assar-Chakravarti


(1984)


study


brandwise search


pattern


demonstrated


only when


the seed brand-attribute


value was


top end


of the


range.


It remains


to be seen how search


patterns


affected


by choice of


"seed"


value


. It may be


that


the extreme


upper


a range may serve as a psychological anchor


permitting a


representation


"best


available" )


that


leads


to brand based


search


This


may not hold


for values


in the


interior of


range


that


only


permit


less


extreme


favorable


evaluations.


Other


possible


psychological


anchor


points


range


also need


to be explored.


are











next


chapter


presents


the experimental


study


conducted along


lines


developed


above


We manipulated


range


knowledge


in a manner


similar


previous


Assar-Chakravarti


1984)


study- -but


this


time


studied


effects


over


alternative


brand-attribute


values


that


spanned


range of


attribute


values under consideration.


passing,


one


might


note


that


there are numerous


possibilities


interesting


studies


other aspects of


range knowledge effects.


instance,


some


ci rct-ustances,


sae


person may find


different

of lens


ranges o

quality


the same attribute


available


relevant


the environment


Suppose


known


that the range

to be 15 to 60.


However,


available


choice alternatives may be restricted


(say


between


lens quality values of


30 and


some alternatives are


out of


stock.


such


situation


, the


consumer


ecologically


available


range


to 60)


to form an evaluation


Also,


if the


consumer


free


to choose or to


reject


the entire set


, this ecological


range


would affect search.


A camera with a


lens quality


index of


45 my


seen


mediocre


that


attribute,


that


brand may not


considered


for choice,


nor searched further.


However


relevant


if the consumer must make a choice


range


from what


the choice set may be used


is available


to form provisional


evaluations.


this


case


brand with


the lens quality


index of


seen


best available


. The evaluation


in absolute


terms











These,


other


interesting


issues


such as


the effect on


the use


range knowledge when search


yields


values outside a


previously known


range


(See


Meyer


1981


Hagerty


Asker


1984


, suggest


issues


warranting


empirical


further study


work


However


this


such


issues are outside


dissertation


are


left


the scope of


for future


research.
















CHAPTER 5


AN SMPIRICAL


INVESTIGATION OF


ATIRI 1TE RANGE


KNOWLEDGE EFFE S:


CONCEPTUAL HYPOIHESES


AND STUDY DESIGN


previous


chapter


outlined


constaer'


knowledge


attribute


ranges


in a product category may guide evaluation and search.


proposed


make


that


provisional


the consumer may use range knowledge as a context

evaluations of brand attribute values encountered


sequential


mediate


these


search


patterns of


processes,


process


sequential


S


. These


search


provisional


information.


effects


evaluations may then


As a consequence


occur


level


confidence


that consumers have


this chapter


, the proposed mechanism is developed more


formally


First,


scenario


described


which


range knowledge may affect


provisional


evaluations of


discussion


a brad's attribute values


impact


. This


these evaluations on


is followed

information


search


behavior


Alternative accounts of how these evaluations may occur


are

Next,


first


discussed and corresponding


the likely effects of


these


formal hypotheses are


provisional


presented.


evaluations on search are


discussed


along with


formal


hypotheses


for the same.


Finally,


effects


are


on confidence


followed


in evaluations are hypothesized.


description of


These hypotheses


a study that was designed


to test


in their evaluations.


secondary











Range


Effects


on Provisional


Evaluations:


A Proposed Mechanism


Very


often,


complex


technically


sophisticated


product


categories,


consumers


have


deal


with


attributes


that


difficult


caup


rehend


or described


using unfamiliar


terminology


example,


many


beginning


amateur


photographers may be considering


purchase


single-


lens


reflex


(SLR)


cameras


that


typically have a


number


lens


elements


optical system.


The consumer may be


unable


assess


performance


features of


these elements


, and even


some


externally


determined assessment


is provided


in the


form of


"lens


quality


index,


consIuer


not be able


to comprehend or


evaluate


meaning


this


terminology


itself


Thus,


if the


conslzmer


encounters


advertisement


for an SIR camera


that says


that


lens


quality


index


she may find


it difficult


evaluate


given brand-attribute value and consequently be uncertain


about


what evaluation


to attach


to it.


The Evaluation Mechanism


consumers may deal


with


this uncertainty


is not entirely clear


What


involved


development


reporting


of a subjective


evaluation


-canp


rehended


piece of


information.


Although some


findings


have


been


equivocal


between negative and neutral


evaluations


, Yamagishi


Hill


1983)


, much


evidence


suggests


that


uncertainty


viewed negatively


in the context of


the evaluation


4ool.~~~~~~~~~I.-1- fl%_~ r. nI 4. A -- ..-a .*n


are


nru~


C, ~,,, Lt,,,,,~,l


d


*


__














and/or


poorly


understood.


This discounting


behavior my be applied


consistently


regardless of


the brand-attribute


value


presented.


Depending


risks


that


constImer


perceives


situation,


one


alternatively


argue


that


uncertainty could lead a


consumer


to avoid making extreme evaluations of


information


that


is not


fully


comprehended.


This


tendency would


reflect


in moderate subjective


provisional


evaluations


being formed and reported.


illustrate


brands


these


with


alternatives,


lens quality index


consider


ratings of,


SIR camera


, 30 and


consumer


does


evaluations


report


understand


this


these


scale


point


form


scale as


subjective


, say


2and


respectively


in Fig


Under the uniform discounting mechanism,


both


these


responses


would


be lower than an evaluation


that might


given


contextual


information


(such


range


knowledge)


were


available


making


judgment.


However


, under


the alternative


mechanis ,


where


conssuer


was


attempting


avoid


forming and


reporting


extreme


evaluations,


one


expect


the lawer


lens quality


index


of 30 to receive a


rating of


whereas


the 45 may be rated


as before.


would serve as


Perhaps


the natural


the mid-point of


anchors


the scale or the end-points


for these reported judgments.


situation


changes considerably when


the consumer has knowledge


range


of lens quality


index scores.


This


range


then


provides a


scale


which


inScming


brand-attribute


value may be placed and














Brand A


Brand


- 30


= Lens Quality


Index


Discounting


Avoidance of
Extremes


Alternative


Figure 5-1
Evaluation Mechanisms











desirable


same


part


token,


of the


smaller


range


attribute


and should


values


receive higher


are


likely


ratings.


receive


By the


lower


ratings.


Attribute


values


middle


range may receive


moderate


scores.


arguments


above have


been presented


intuitively


Past


research


(see


Parducci


1965;


Parducci


and Perrett


1971)


also suggests


that


reported


ratings


brand-attribute


also


values


affected


over the range and


distribution of


the number of


response scale


categories.


this


argment,


the ratings received would also vary by


width of


the scale


. Given a wide and a narrow scale centered at


same


lens


quality


index


score,


evaluation


of given brand-attribute


values


should


differ


systematically based on


the mapping


of the given


attribute


value


to the natural anchors provided by the range knowledge.


Search Patterns Contingent on Overall


Evaluation


many


juxig nnt


contexts,


consue rs


have


judiment-relevant


information


available


the same


time


. Sometimes,


information


items


are acquired


sequentially


following the acquisition


initial


"seed"


item of brand-attribute


information.


As discussed


above,


range


information


enables


consner to


form a


preliminary


evaluation


"seed"


brand-attribute


value.


This evaluation may


lediate


search behavior


We will


examine how this mediation prices may


develop


function


range


knowledge


the attribute value


encountered.











among


three


alternatives


(say


B and C)


for which


information


available


three


attributes--shutter


speed,


lens


quality


warranty


Say the consumer


learns


that


brand B has a lens quality


index


Since


value


consumer


be difficult


lacks a context


to interpret


, this seed brand-attribute


. The consumer may then examine


lens


quality


index


values


brands


A and C to form some basis


evaluating


given


brand-attribute


value


.Thus


, in this


situation,


consumer may search by attribute


look at other


brands on


same


attribute)


to form a context


for evaluating the seed value of


lens


quality


inrdex


for brand B.


Note


that


this


tendency for attribute


based


search


will


persist


regardless


level


initial


brand-attribute


value.


When


range


knowledge


is available


this


information can be used


form


will


a provisional


mediate


evaluation of


subsequent


brand


pattern


. This


of search


provisional


Since


evaluation


the provisional


evaluation


will


depend


on the


actual


nature of


the range knowledge as


well


level


brand-attribute value encountered


, we will


examine several


alternative


cases.


First,


consider


case


where


the seed


value


is a


lens quality


index


score of


50 for


brand B and


known


range of


lens quality


index


scores


, 30-50.


this


particular


case,


seed


brand-attribute


value


coded


the "best available"


brand


would


rate


prime


choice


candidate on


basis


of this











known


range


were


wider


20-60


but assumed


to be


centered


value of


lens quality score of 50 would


still


extreme


evaluated


previously


favorably


Hence


while


this

there


evaluation

is still a


would not


good chance


be as

that


brand

the


will


consumer


searched further


may search


there


the other brands


is a greater


(A and C)


likelihood


on the


that


lens quality


attribute


see


whether


they


are


better


than


Thus,


in this


particular


case,


there


likely


lower


probability


brand-based


search


than


in the previous case and a higher


likelihood of


attribute-based


search.


case where


lens quality


index scores


still


range between


where


seed attribute value


, 45,


may be


viewed


similarly


case


just


discussed.


Since


evaluation


favorable,


not extreme,


brand-based search


is somewhat


less


likely


attribute-based search


for a


brand with a better


lens quality


index


score


is more


likely to occur.


Thus


, in


rmn


, the search


predictions


are conditioned


upon


favorableness


of the


provisional


evaluation of


the seed brand-attribute


value


The more extreme


favorable


provisional


evaluation of


a given


seed


brand-attribute


value


(given


the known


range of


attribute scores)


more


likely


brand-based search.


likelihood of brand-based


search


will


decline


lower brand-attribute values as


provisional


evaluations


become


less


favorable.


propensity


search