price knowledge during grocery shopping: what we learn and what we forget

R

A

Trssspsupum©

K

pritaiS

aawh

vD

B

0h

ARTICLE IN PRESS+ModelETAIL-501; No. of Pages 15

Journal of Retailing xxx (xxx, 2014) xxx–xxx

Price Knowledge During Grocery Shopping: What We Learn andWhat We Forget�

Birger Boutrup Jensen ∗, Klaus G. Grunert 1

MAPP Centre for Research on Customer Relations in the Food Sector, Business and Social Sciences, Aarhus University, Denmark

bstract

Past research on consumer price knowledge has varied considerably partly due to differences in how and when price knowledge is measured.his paper applies a multi-point, multi-measure approach to reconcile differences in past price knowledge research by examining systematic

elationships between time of measurement and type of measures applied. Examination of consumer price knowledge before, during, and aftertore visit sheds light on what is measured at the individual points in time: episodic price knowledge and/or reference prices? With a between-ubjects design interviewing 1,204 respondents, the authors investigate three price knowledge measures (price recall, price recognition, and dealpotting) demonstrating that these are hierarchically related. Results suggest that reference prices dominate before store visit, but also that episodicrice knowledge, surprisingly, is still accessible at the store exit. These findings enable the authors to reconcile diverging results from past research,howing how consumer price knowledge evolves and suggesting that the vast majority of consumers learn about prices, whether consciously ornconsciously, during grocery shopping. Thus, when applying a multi-point, multi-measure approach, consumers appear to know more about
rices than suggested by past research. Determinants of price knowledge are also examined and the results indicate that price knowledge buildsp not only because of active search but also due to accidental exposure to prices and with low degrees of conscious processing. Implications foranagers are discussed.
2014 New York University. Published by Elsevier Inc. All rights reserved.

tail co

1vta2pt

eywords: Episodic price knowledge; Grocery shopping; Reference prices; Re

Introduction

Economic theory largely assumes that consumers know therice of products they purchase with reasonable accuracy. Whenetailers and manufacturers set prices strategically, they alsomplicitly assume that consumers know the price of the productshey purchase (Urbany, Dickson, and Sawyer 2000). Meanwhile,

Please cite this article in press as: Jensen, Birger Boutrup, and Grunert, Kland What We Forget, Journal of Retailing (xxx, 2014), http://dx.doi.org/1

generation of research suggests that consumer price knowledges poor, even for items they have just chosen (e.g., Dickson andawyer 1990; Vanhuele and Drèze 2002; Wakefield and Inman

� The authors thank Tino Bech-Larsen and Hans Jørn Juhl for providing valu-ble feedback on earlier versions of this paper. We also want to acknowledge

major Danish retail chain for its role in the data collection process. Finally,e are grateful to the editor and three anonymous reviewers whose commentselped to improve the quality of this paper.∗ Corresponding author at: MAPP, Business and Social Sciences, Aarhus Uni-ersity, Bartholins Allé 10, Building 1323 – Office 216, DK-8000 Aarhus C,enmark. Tel.: +45 87165401.

E-mail addresses: [email protected] (B.B. Jensen), [email protected] (K.G. Grunert).1 Address: MAPP, Business and Social Sciences, Aarhus University,artholins Allé 10, DK-8000 Aarhus C, Denmark. Tel.: +45 87165007.

datcttiedsoaim

022-4359/$ – see front matter © 2014 New York University. Published by Elsevier Ittp://dx.doi.org/10.1016/j.jretai.2014.01.001

nsumer behavior; Retail pricing

993). However, results of studies on consumer price knowledgeary considerably, thus making it difficult to draw any consis-ent conclusions. For instance, as shown in Table 1, shoppers’bility to recall exact prices has been found to vary between.0 and 61.3 percent. Such wide variations may be attributedartly to differences in sociocultural or macroeconomic condi-ions, but their origin is more likely linked to differences in studyesign (see Estelami and Lehmann 2001; Estelami, Lehmann,nd Holden 2001). In particular two design issues stand out:he timing of the measurement in relation to the buying pro-ess (before store entry, point of selection, after store visit) andhe type of measures applied (recall, recognition, judgment, rela-ive ranking). Other design issues possibly causing this variationnclude differences in consumer-related factors and product cat-gory selection. The overall objective of this paper is to reconcileifferences in previous price knowledge research by examiningystematic relationships between time of measurement and typef measures applied, while controlling for effects of consumer

aus G., Price Knowledge During Grocery Shopping: What We Learn0.1016/j.jretai.2014.01.001

nd product-related factors, leading to an improved understand-ng of mechanisms governing consumer price knowledge and itseasurement.

nc. All rights reserved.

dx.doi.org/10.1016/j.jretai.2014.01.001


mailto:[email protected]

mailto:[email protected]


ARTICLE IN PRESS+ModelRETAIL-501; No. of Pages 15

2 B.B. Jensen, K.G. Grunert / Journal of Retailing xxx (xxx, 2014) xxx–xxx

Table 1Time of measurement and measures applied in past price knowledge research.

Before During After

Price recall Gabor & Granger (1961)b 51% Conover (1986) (Study 1) 51%Urbany & Dickson (1991)b 53% Conover (1986) (Study 2) 26% Conover (1986) (Study 2)a 27%Krishna et al. (1991)b 15% Dickson & Sawyer (1990) 47% McGoldrick & Marks (1987) 29%Vanhuele & Drèze (2002)(Study 2)

2% Wakefield & Inman (1993) 55% McGoldrick, Betts, andWilson (1999)

40%

Le Boutillier et al. (1994) 61% Rosa-Díaz (2004) 20%Vanhuele & Drèze (2002)(Study 1)

10%

Price recognition Vanhuele & Drèze (2002)(Study 2)

42%

Price judgment Urbany & Dickson (1991)b 51–54%Vanhuele & Drèze (2002)(Study 2)

33%

Notes. Price judgment here refers to measurement of consumers’ ability to judge an item’s price compared to its normal price. References in bold represent studiesusing a multi-point approach. In addition to the above-mentioned studies, a number of laboratory experiments have applied price recall and relative price ranking tostudy price knowledge (e.g., Mazumdar and Monroe 1990; Zeithaml 1982).

a Conover (1986) conducted phone interviews with the same respondents two days after the initial simulated shopping experiment and therefore as an exitm

viiwrdasswsasrwcwdip

rTaaoto

apb

pacTiuktamvtp

aptWntipr

easurement the study is a borderline case.b In-home survey.

With regard to timing, Table 1 shows that almost all pre-ious studies measure consumer price knowledge at one pointn time only, whether inside or outside the store. Exceptionsnclude Vanhuele and Drèze (2002) who measured price recallith two different samples at the entrance and at the shelf

espectively, and Conover (1986) who used a within-subjectsesign to measure immediate (simulated shopping experiment)nd delayed price recall (phone interview two days later) withome very surprising results.2 Otherwise, previous studies mea-ure price knowledge at a single point in the buying process,hich complicates cross-study comparisons as they may mea-

ure different aspects of consumer price knowledge. Monroend Lee (1999) argue that in-store studies mainly measure in-tore attention to price (i.e., short-term memory dominates theesults), whereas long-term price knowledge becomes the focushen measuring outside the store. However, as past studies have

oncentrated on single-point measurement, we do not know tohich extent differences in time of measurement account forifferences in measured price knowledge, and whether theres any systematic relationship across different measurementoints.

Another major reason for the variation in previous studyesults relates to how price knowledge was measured. Fromable 1 it is evident that price recall has been the single mostpplied measure, and such studies have generally found thatround half of the shoppers recall the accurate price at the point


f selection: considerably less before and after the store visit,hus suggesting rather poor price knowledge. However, relyingn price recall alone may underestimate the degree of consumer

2 Surprisingly, Conover (1986) found no evidence of fading price recall after two-day delay; for half the products, price recall even improved, which heroposes “may reflect heightened attention to those prices once price questionsegan, and rehearsal of them afterward” (p. 593).

pkq

1

rice knowledge. Monroe and Lee (1999) distinguish explicitnd implicit price memory and argue that some consumers pro-ess prices during product choice without being aware of it.hese consumers do not recall the exposure event or the price

tself, but they still may be able to judge the price of the prod-ct chosen according to its attractiveness. Such shallow pricenowledge would not be uncovered in an explicit price memoryest such as a price recall or a price recognition test. Monroend Lee therefore recommend that future studies include tests toeasure different levels of price knowledge. Table 1 reveals that

ery few studies have taken up this recommendation, and it ishus likely that past research has only revealed part of consumerrice knowledge in the buying process.

To improve our understanding of consumer price knowledgend reconcile past research, multiple measures are required andrice knowledge must be assessed at multiple points in time inhe buying process. This is what we did in the present study.

e conducted three price memory tests (price recall, recog-ition, and judgment), and we did so before, during and afterhe store visit. Thus, we examined all nine cells of Table 1n a single study which allows us to pull back the curtain onrice knowledge acquisition much more effectively than in pastesearch.

Research questions

As noted, we extend previous research by applying a multi-oint, multi-measure approach to examine consumer pricenowledge. This design enables us to address the followinguestions:


. How price knowledgeable are consumers and how does thisknowledge differ between the three stages of a store visit?Our design can uncover whether consumers know more aboutprices than suggested by past research. In addition, this design



al of R

2

3

dobrp

C

eaietotsab

a

laioWlpcoMdpmtPpapuprHeaBwl

Hk

mrbd

–

Inman 1993). Price recall is assumed to be accurate if con-sumers can draw on episodic price knowledge, and accurateprice recall may hence result if consumers have engaged in

use the term ‘reference price’ without explicitly distinguishing “internal” and“external,” we refer to internal reference prices.

4 The ability to price rank an item relative to other items from the same (orrelated) product category differs from the ability to judge an item’s price com-

B.B. Jensen, K.G. Grunert / Journ

enables better indications as to what is measured at the threestages in the buying process. When is price knowledge dom-inated by episodic price knowledge and reference pricesrespectively?

. Are the measures of price knowledge (price recall,price recognition, and price judgment) hierarchicallyrelated? We assume a hierarchy of price informationprocessing—intentional, incidental and unconscious—thatcorresponds to different levels of consumer price knowledgeand that can be captured by different price knowledge meas-ures. Our design also allows the testing of the robustness ofthese relationships across measurement points.

. How do brand and store loyalty, category purchase fre-quency, price range, and deal share add to the explanationof the variance in consumer price knowledge? Since pastresearch on determinants of price knowledge has concen-trated on price recall (and the point of selection), variousprice-search related consumer traits have received most atten-tion (e.g., Dickson and Sawyer 1990; Wakefield and Inman1993). Less attention has been paid to factors affecting con-sumers’ exposure to prices (both in terms of frequency andvariation) which may influence consumer learning of pricesdue to differences in their history of exposure. Such learn-ing might be captured by less cognitively demanding pricememory tests. Thus, our design enables us to improve ourunderstanding of consumer price knowledge by exploringfactors like brand and store loyalty, category purchase fre-quency, price range, and deal share.

Combining different types of measures and applying them atifferent times in the buying process enables us not only to rec-ncile the various results of previous price knowledge researchut we can also shed new light on the conflict between theesults of scanner-based reference price studies and those ofrice knowledge surveys (cf. Monroe and Lee 1999).

Levels of price information processing, price knowledgeand its measurement

onceptual framework

We distinguish between episodic and semantic price knowl-dge (Chang 1986; Tulving 1972). The former is knowledgebout prices that a consumer has observed at a certain pointn time at a certain location, whereas the latter refers to gen-ral knowledge about product prices, that is, without referenceo any specific episode. Reference price is the central elementf semantic price knowledge. Researchers commonly assumehat consumers judge an item’s purchase price against an item-


pecific internal reference price (Briesch et al. 1997; Kalyanaramnd Winer 1995; Mazumdar, Raj, and Sinha 2005) which isased on prior purchase experience.3

3 Internal reference prices are based on previous experience (memory based)s opposed to external reference prices that are stimulus based. Whenever we

pmtsekst

etailing xxx (xxx, 2014) xxx–xxx 3

Both episodic and semantic price knowledge are affected byearning of new price information. Also, learning of episodicnd learning of semantic price knowledge are related, as learn-ng of new episodic price knowledge should lead to an updatingf the reference price (e.g., Briesch et al. 1997; Winer 1986).e distinguish three types of price information processing that

ead to learning of new price knowledge: (a) intentional (activerice search), (b) incidental (price noticed by chance at producthoice), and (c) unconscious (minimal attention to and encodingf price stimuli making price memory implicit) (Mazumdar andonroe 1990; Monroe and Lee 1999). Both intentional and inci-

ental price information processing may lead to new episodicrice knowledge, although the strength of the memory traceay differ. If consumers engage in intentional price processing,

hey may later be able to recall the exposure event (Monroe,owell, and Choudhury 1986). With incidental price processing,rice leaves relatively weak memory traces, though it should beccessible immediately after product choice. Finally, even whenrices are perceived as irrelevant, they may still be processednconsciously, though these price stimuli are registered onlyeripherally, leaving weak traces in memory, and cannot beecalled even immediately after product choice (Alba, Wesleyutchinson, and Lynch Jr. 1991; Monroe and Lee 1999). Refer-

nce price research assumes that reference prices are updatedutomatically during price information processing (see e.g.,riesch et al. 1997; Mazumdar, Raj, and Sinha 2005), whichould also imply that all three types of price processing will

ead to reference price changes.

ierarchical relationship between measures of pricenowledge

According to Table 1, price recall is the single most appliedeasure to assess consumer price knowledge, although price

ecognition, price judgment and relative price ranking have alsoeen used to some degree.4 To various extents these measuresraw on both episodic and semantic price knowledge.

A consumer may know the actual price of an item by heart.Such knowledge typically has been measured by a pricerecall test in which the respondent states the price of a prod-uct unaided (e.g., Dickson and Sawyer 1990; Wakefield and


ared to its normal price (i.e., its own price history). For instance, a consumeray be able to price rank brands, but in case of an item-specific price increase

his knowledge would not help him/her unless s/he also holds item (or category)-pecific reference prices. Since we focus on the evolvement of item-specific pricexpectations and episodic price knowledge, as well as different levels of pricenowledge, we exclude relative price ranking. Thus, we would not expect a clearystematic relationship between on the one hand relative price ranking and onhe other hand price recognition and price recall.




cessin

–

–

ciidotrCurspsr

stsut

P

seosRo1Hremristatstptpp

cwwbp

Fig. 1. Relationships between level of price information pro

intentional or, to a lesser extent, incidental price informationprocessing. If episodic price knowledge is inaccessible, con-sumers may answer a price recall test based on their referenceprice, and accuracy may depend on when this reference pricewas updated and on factors affecting price variation.

Consumers who cannot recall a specific price still may beable to recognize a price if given visual cues. This kind ofprice knowledge can be measured by a price recognition testin which respondents select a price from among a range ofhypothetical prices (Monroe, Powell, and Choudhury 1986).The likelihood of correct price recognition should be higherif consumers can draw on episodic price knowledge, but ascues are provided, a weak memory trace should be sufficientto trigger correct recognition as compared to unaided recall.

Even if consumers fail to recall or recognize the actual price,they may be able to judge whether a price is attractive com-pared to the normal price. That is, they may have implicitprice knowledge in the form of a sense of familiarity withthe normal price. Such familiarity can be enhanced by uncon-scious price processing which may prime the reference price.Monroe and Lee (1999) argue that such shallow levels of priceknowledge can be detected only if consumers receive suffi-cient cues in a test of implicit memory, such as in a judgmenttask (e.g., a deal spotting test). Consumers should be able toanswer a deal spotting test correctly even when they cannotdraw on episodic price knowledge.

The three measures of price knowledge to the right in Fig. 1an be argued to form a hierarchy. Thus, the three measures differn difficulty, they differ in the extent to which an accurate answers aided by the presence of episodic price information, and theyiffer in the way the probability of an accurate answer dependsn the occurrence of the three types of information processingo the left in Fig. 1. The absence of retrieval cues makes freeecall the most demanding memory test (Singh, Rothschild, andhurchill 1988). Hence, if a consumer is able to recall a partic-lar price, because s/he can draw on episodic price informationesulting from intentional, or at least incidental price processing,


/he should also be able to activate the same knowledge whenrovided with retrieval cues, as in price recognition and dealpotting tests. Likewise, if a consumer can recognize (but notecall) a particular price, s/he should be able to activate the

rWse

g, types of price knowledge and price knowledge measures.

ame knowledge and apply it when answering a deal spottingest. Therefore, price recall, price recognition, and deal spottinghould be hierarchically related in that they map onto the samenderlying dimension, namely, the strength of the memory tracehat the price information has left in the consumer’s memory.

rice knowledge before, during and after store visit

We expect that consumer price knowledge changes duringtore visits. Before a store visit, it consists primarily of refer-nce prices, which are commonly conceptualized as a weighingf past item-specific prices with the most recent price expo-ures having the greatest impact (see e.g., Briesch et al. 1997).esearch also suggests the existence of reference price zones aspposed to exact reference points (e.g., Kalyanaram and Little994), which implies that reference prices cannot be accurate.ence, consumers are expected to arrive at the store with inaccu-

ate price expectations. Some episodic price knowledge may alsoxist in memory, but grocery shopping tends to be a low involve-ent task, so accessible episodic price knowledge is likely to be

are before store entry. In contrast, the price of a just chosentem should be available as episodic knowledge at the point ofelection if the consumer has engaged in intentional or inciden-al price information processing during product choice (Dicksonnd Sawyer 1990; Monroe and Lee 1999). Therefore, the propor-ion of consumers with correct price recall and price recognitionhould be higher at the point of selection as compared to athe store entry. Also, as consumers’ reference prices should berimed and updated with the current price at the point of selec-ion, even if only unconscious price processing takes place, theroportion of consumers able to spot good and bad deals at theoint of selection should be higher than at the store entrance.

Price information processing during store visit may affectonsumers’ price knowledge on leaving the store compared tohen entering the store in two ways. First, price knowledgeill be more accurate because the reference price will haveeen updated with the current price as a consequence of in-storerice information processing. Second, episodic price knowledge


esulting from the store visit may still be accessible at the exit.e therefore expect consumers to perform better on each mea-

ure of price knowledge at the store exit compared to at the storentry.



al of R

cetgeia(ra

Fd

aapcprL2nitfcass

aptarictasbbtpl(nv

tlkcp

lescartiuiplv

tqetipbestkea

ppptootkem

B

dWmidsa(d


Comparing price knowledge immediately following producthoice with price knowledge at the store exit, the updated ref-rence price should still be accessible when leaving the store,hus resulting in no expected differences in the ability to spotood and bad deals between the two measurement points. How-ver, the episodic price knowledge that may have resulted fromntentional or incidental price processing should already be lessccessible compared to the point directly after product choiceDickson and Sawyer 1990). As a result, we expect correct priceecall and price recognition to be less likely at the store exit thant the point of selection.

actors affecting the level of price knowledge before,uring, and after store visit

Past research has largely assumed that some consumersre naturally price sensitive, hence they actively search fornd thus “know” prices (i.e., can recall prices). As a resultast research on determinants of price knowledge has con-entrated on various price-search related consumer traits likerice and value consciousness (and related variables, e.g., self-eported price checking) and their effect on price recall (e.g.,ichtenstein, Ridgway, and Netemeyer 1993; Mägi and Julander005; Wakefield and Inman 1993). Price and value conscious-ess are included in this study as control variables; however, ournterest in analyzing determinants concentrates on factors relatedo consumer exposure to prices both with regard to exposurerequency and with regard to the variation in prices which theonsumer is exposed to. Applying a multi-point, multi-measurepproach, we extend past research by exploring how brand andtore loyalty, category purchase frequency, price range and dealhare can add to the explanation of price knowledge.

Brand loyalty is often characterized by a limited choice setnd limited information search, resulting in less exposure torices both with regard to number of exposures and varia-ion in the prices that consumers are exposed to (e.g., Helsennd Schmittlein 1994; Krishnamurthi and Raj 1991). This mayesult in less episodic price knowledge as less intentional and/orncidental information processing will take place. However, itould also result in more accurate price expectations becausehe limited choice set also means that there is less price vari-tion to keep track of. As price knowledge measured beforetore entry draws predominantly on consumers’ reference price,rand loyal consumers’ price knowledge measured there shoulde more accurate. However, we do not expect a similar effect athe two other measurement points. The limited choice set makesrice comparisons less likely during store visit, although brandoyal consumers may be price sensitive to quantity decisionsKrishnamurthi and Raj 1991). Thus, brand loyalty should haveo effect on the level of price knowledge during and after storeisit.

Up to now, store loyalty has received very limited atten-ion in price knowledge studies. Assuming price variation is


ower within than across stores, store loyal consumers’ pricenowledge is less fragmented as it comes from fewer storesompared to that of less store loyal consumers. As referencerice knowledge dominates the entrance measurement, store

uimi


oyalty is expected to positively affect the level of price knowl-dge before store entry. Store disloyalty has been associated withearch for low prices and price specials across stores, includingherry picking (e.g., Kumar and Leone 1988; Talukdar, Gauri,nd Grewal 2010). Furthermore, less store loyal consumers areelatively less familiar with the layout of a particular store andhus more prone to being guided by in-store information, includ-ng price signs (Mulhern and Padgett 1995); also they are morencertain about a particular store’s prices, which may resultn intentional price processing for cross-store comparison pur-oses. Hence, store loyalty may have a negative effect on theevel of consumer price knowledge during and after the storeisit.

Category characteristics may influence price knowledge dueo their implications for price exposure. High purchase fre-uency categories will result in more frequent and recentxposures to prices leading to more accurate price expecta-ions. In addition, high frequency categories may create a greaterncentive to search for lower prices, encouraging intentionalrice processing and hence increasing the likelihood of accessi-le episodic price knowledge. Previous studies suggest a positiveffect of buying frequency on price knowledge at the point ofelection (Le Boutillier, Le Boutillier, and Neslin 1994) and athe store entry (Vanhuele and Drèze 2002). The level of pricenowledge before, during and after the store visit is thereforexpected to be higher as the higher frequency of exposure mayffect both episodic price knowledge and reference prices.

Similarly, product categories with a high deal share and/orrice range may motivate consumers to engage in intentionalrice information processing resulting in accessible episodicrice knowledge while at the same time making it more difficulto keep price expectations accurate due to increased likelihoodf varying price exposures. We therefore expect a higher levelf price knowledge at the point of selection and at the store exit,hat is, when the measures are based mainly on episodic pricenowledge, but we do not expect a similar effect before storentry where performance on price knowledge measures is basedainly on reference prices.

Method

asic design

We interviewed shoppers before they entered the store, imme-iately after their product selection, and as they left the store.e did not employ a within-subjects design where all threeeasurements would be done with the same shopper, because

nterviewing the same shopper about price knowledge before,uring, and after store visit likely would lead to inflated situation-pecific price involvement (Conover 1986). Instead, we applied

between-subjects design and compared three random samplesbefore, during, and after store visit) with considerable effortevoted to ensuring the comparability of the samples and meas-


res: (a) identical design of the price memory tests, (b) thenterviews being carried out in the same stores and at approxi-

ately the same time of the day and the week, (c) standardizednterview procedures and training, (d) the same interviewers



6 B.B. Jensen, K.G. Grunert / Journal of R

Table 2Focal product categories and their relation to category characteristics.

Product categories Category characteristics

Deal share Buying frequency Price range

Margarine L L LToothpaste H L LMuesli L H LJuice H H LOlive oil L L HSandwich chocolatea H L HKetchup L H HGround coffee H H H

Note. H indicates that a product category scores relatively high on the charac-t

c(p

S

aBpsaitwpWkw

viwltrrItpicrt

vPaviB

wo

fiesafreseaa

tWqnnbna(itescvirwfiswl

(1atujiAvt

M

the following order: (1) price recall, (2) price recognition, and

eristic, whereas L indicates a relatively low score.a Thin slices of chocolate to put on slices of bread.

arrying out the interviews according to a rotation principle, ande) identical product categories across the three measurementoints.

tore selection, product selection, and fieldwork

The data collection covered a two-week period in August in large city in Denmark. Since prior studies (e.g., McGoldrick,etts, and Wilson 1999) show that store formats can affectrice memory test outcomes, we conducted our study in twotores (hypermarket and supermarket) differing in size (15,600nd 1,850 m2, respectively), location, customer base, and pricemage. Furthermore, because the types of shoppers differhroughout the day and week, as does the shopping occasion,e considered potential differences in consumer informationrocessing resulting from this (Bell, Corsten, and Knox 2011;alters and Jamil 2003) and scheduled interviews to match

nown variations in store traffic spanning both weekdays andeekends.Our selection of product categories was guided by systematic

ariations in buying frequency, price range, and deal share. Buy-ng frequency referred to the relative rate of turnover, price rangeas measured as the price difference between the highest and

owest priced item in the category, and deal share representedhe share of turnover in the category accounted for by tempo-arily reduced prices. Category managers from the participatingetail chains supplied the data for these three characteristics.n discussions with category managers and store managers andhrough exploratory store checks, we identified 29 potentialroduct categories; these were narrowed down to eight result-ng from practical considerations such as the location of theategory in the store. Each product category in Table 2 thus rep-esents relatively high or relatively low levels on each of thehree characteristics creating a 2 × 2 × 2 factorial design.

The data collection involved a combination of personal inter-iews and a questionnaire that respondents completed at home.ersonal interviews allowed for the presentation of visual stimulidapted to the respondent’s product choice. In addition, inter-


iews are well suited for measuring consumers’ price awarenessmmediately after product choice (Dickson and Sawyer 1990).y combining the interviews with a take-home questionnaire,

(us

etailing xxx (xxx, 2014) xxx–xxx

e reduced the duration of the interviews significantly, whichught to result in a higher response rate in the store.

Every third group (single person or family) that passed axed point in the entrance/exit area was intercepted, whetherntering or exiting the store, and asked to identify the primaryhopper and participate in an interview. At the store shelves,n interviewer approached the first person who chose an itemrom a focal product category. On completion of the interview (orefusal to participate), interviewers waited at least two minutes tonsure potential respondents would be unaware of the interviewituation before approaching the next potential respondent. Atach interview point, respondents were offered a chance to win

gift certificate to use in the store (conditional on participationnd returning a completed questionnaire).

At the entrance, because consumers do not always plan whato buy at the item level (Bell, Corsten, and Knox 2011; Inman,

iner, and Ferraro 2009), we asked the following screeninguestion: “Have you considered buying [category] at [storeame] today?” If the respondent indicated such an intention oroted that the question prompted a recollection of the need touy, this category provided the starting point of the interview. Ifot, the interviewer continued to the next category. To qualify forn entry interview, respondents thus had to indicate a manifestplan to buy) or latent (need exists but no plan to buy) purchasentention. As it is easier to recognize a picture of a specific itemhan recall the item’s name, we developed interview folders forach category with pictures of the items (in total 113 SKUs)old by each store. The interviewers used a list with the focalategories listed according to a predefined order (the order wasaried several times) and asked respondents for their purchasentention in the category one at a time. For the first category theespondent intended to buy, the interviewer provided a folderith pictures of all items in that category; respondents identi-ed the specific item they planned to purchase or, if they had nopecific plan, the item they purchased most often. The picturesere sorted in alphabetical order by brand, and then by SKU

evel, the brand name serving as an additional cue.At the shelves, interviewers discreetly observed shoppers

applying a similar procedure as that of Dickson and Sawyer990) and interviewed them about the item chosen immediatelyfter they placed the item in their shopping cart. At the exit,he interviewers again used an ordered list of the focal prod-ct categories and, one at a time, asked shoppers if they hadust bought one of them; if so, they identified the item in annterview folder similar to the one used in the entry interviews.s recommended by Estelami and Lehmann (2001), we inter-iewed shoppers about one item from a single product categoryo reduce the task burden.

easures and stimuli

We applied three measures of consumer price knowledge in


3) deal spotting. The sequence of administering the three meas-res was determined by decreasing demands with regard to thetrength of the memory trace. As we moved from one measure



B.B. Jensen, K.G. Grunert / Journal of Retailing xxx (xxx, 2014) xxx–xxx 7

d app

ti

tviwt[inibacb

fWpptP1tititmpoat

pdddcpt

aiDo

taTdtf

tapb(pictro+

anftmsiobcs

Fig. 2. Example of a stimulus car

o the next, we introduced an increasing number of visual cuesn order to facilitate retrieval of memory traces.5

First, respondents stated the actual store-specific price ofhe focal item on that particular shopping trip. Before the storeisit, the price recall question asked was: “What do you thinks the price of this brand of [category] at [store name] today?”hereas after the store visit, the price recall question referred

o the purchased item: “What was the price of this brand ofcategory] at [store name] today?” Because some actual pricesncluded promoted prices, we also asked respondents to state theumber of units to which the recalled price corresponded (i.e.,f they bought/intended to buy a multi-item offer, this shoulde reflected in the number of units stated). At the shelves wesked: “What was the price of this brand of [category] you justhose?” We accepted “don’t know” responses as recommendedy Estelami and Lehmann (2001).

Next, the price recognition test was conducted according to aorced choice method (Monroe, Powell, and Choudhury 1986).

e showed respondents a card containing several alternativerices (distractor cues) and “forced” them to choose one of therices from the card. The number of alternatives on the card andhe similarity of the stimuli affect recognition accuracy (Monroe,owell, and Choudhury 1986; Singh, Rothschild, and Churchill988); our price recognition test featured five plausible pricesogether with the actual price to reduce the risk of correct guess-ng (see Fig. 2). The intervals between the prices depended onhe item’s price level but ranged from 3 to 10 percent (largerntervals for products with higher actual prices).6 In addition,he calculated distractor prices resembled plausible prices; theajority of Danish grocery prices end in DKK .95, that is fixed

ercentage intervals would make the hypothetical prices stand


ut and increase the likelihood of correct guessing. The pricesppeared in two consecutive rows of three prices each to reducehe risk of bias due to centripetal behavior, and the position of

5 One might worry that by providing possible prices, the recognition questionrovided indications about the actual price that can facilitate answering theeal spotting question. However, the level of difficulty of our recognition testiminished such risk as the target price is embedded in a list of five plausibleistractor cues. In support of this, 45.1 percent of the respondents could notorrectly spot a deal at the entrance despite having been exposed to the sixrices in the recognition test. This proportion diminished significantly duringhe store visit after price exposure.

6 We adapted the intervals to the absolute level of the actual price, i.e., if thectual price was less than DKK 10, the interval was typically DKK .50–1.00,f the actual price ranged from DKK 10 to DKK 30, the interval was typicallyKK 1.00, and if the actual price was higher than DKK 30, we used intervalsf DKK 2.00.

aotecT

taa(d

bdi

lied in the price recognition test.

he correct price was evenly distributed across stimuli cards tovoid selection bias (Monroe, Powell, and Choudhury 1986).he depiction of the prices used a frame that replicated theistinctive shelf price tag format used by the stores as well asheir actual numeral format. These combined efforts aimed toacilitate retrieval of weak memory traces.

As the final measure of price knowledge, we used a judgmentask that measured deal spotting ability. Respondents judged thettractiveness of four prices (±10 and 20 percent from the normalrice) relative to the normal (non promoted) price at the store. Arief introduction to the test explained the term “normal price”defined as the price normally charged by the retailer when theroduct is not price promoted) and the response options. Thendividual prices were then presented one at a time as visualues followed by the question: “Is this price higher or lowerhan the normal price of this brand of [category]?” Half theespondents considered prices +20%, +10%, −10%, and −20%f the normal price, and the other half saw prices −20%, −10%,10%, and +20% of the normal price.

To determine how consumer price knowledge evolves during normal shopping trip, we used the actual price (promoted orot) in the store during the study as a standard of comparisonor consumer estimates in the price recall and price recognitionests. As regards special-priced items, one might argue that weeasure something which respondents cannot know before the

tore visit. However, in Denmark most price promotions appearn sales flyers distributed to the majority of households and/or inther media so respondents may have processed this informationefore arriving at the store, in particular for products they reportonsider buying in the store on that shopping trip. That is, totudy differences in consumer price knowledge before, during,nd after store visit the actual price must be used as the standardf comparison for each measurement point. The deal spottingest, however, did not rely on explicit recollection of the pricexposure from the current shopping trip as it aims at measuringonsumers’ ability to spot deviations from the normal price.hus, the deal spotting test used the normal price in the store.

Each full interview lasted four to five minutes. On comple-ion, respondents received a questionnaire to complete at homelong with a reminder of the incentive and reassurance of theirnonymity. The questionnaire items measured price involvementprice and value consciousness), brand loyalty, store loyalty,emographics, and other measures not relevant for this study.


As for price consciousness, we apply the definition offeredy Lichtenstein, Ridgway, and Netemeyer (1993, p. 235): “theegree to which the consumer focuses exclusively on pay-ng lower prices”. Consumers may also exhibit greater price




Table 3Matching samples before, during and after store visit.

Before During After Test

DemographicsAge (mean) 47.88 48.84 48.04 F = .5; p = .60Gender (females) 70.6% 68.9% 69.3% χ2 = .3; p = .87Income (mean) 476,401 493,314 475,739 F = .6; p = .55Perceived budget

constraints2.16 2.17 2.25 F = .3; p = .72

Household size 2.74 2.85 2.77 F = .8; p = .45Consumer characteristics

Brand loyalty .09 −.02 −.07 F = 2.2; p = .11Store loyalty 4.97 4.88 4.88 F = .4; p = .66Price consciousness .06 −.03 −.03 F = .9; p = .41Value consciousness −.07 .06 .01 F = 1.5; p = .23

Design variablesHypermarket proportion 51.4% 52.4% 50.5% χ2 = .3; p = .88Margarine proportion 12.6% 11.8% 13.0%Toothpaste proportion 13.1% 13.0% 12.8%Muesli proportion 10.8% 11.8% 12.8%Juice proportion 12.8% 11.8% 12.5%Olive oil proportion 13.1% 12.3% 11.5% χ2 = 1.8; p = .99Sandwich chocolate

proportion12.1% 12.6% 12.5%

Ketchup proportion 12.8% 13.7% 12.5%Ground coffee proportion 12.6% 12.8% 12.5%

Notes. The midpoints of each age and income category were used as substitutes tocalculate age and income means respectively (Nickols and Fox 1983). Perceivedbt

iribatHas22saar

setqaasrns

Table 4Differences in consumer price knowledge before, during, and after store visit.

Before During After

Correct price recall 7.6% 43.8% 37.5%Correct price recognition 21.4% 60.7% 55.5%Correct deal spotting 54.9% 71.6% 66.8%

Deal oblivion 16.1% 12.6% 12.5%

Price recall accuracya 19.3% 7.9% 9.4%

n = 395 n = 420 n = 389

p

dnmb

Mv

rirc2ptobawqcgrtapOeswei

tcompared with at the shelf (recall χ = 137.9, p < .001; recogni-tion χ2 = 129.7, p < .001; and deal spotting χ2 = 24.5, p < .001)and at the exit (recall χ2 = 100.9; p < .001; recognition χ2 = 96.8;

udget constraints were measured on a seven-point Likert scale consisting ofhree items adapted from Urbany, Dickson, and Kalapurakal (1996).

nvolvement due to a general concern about the quality/priceatio, which is referred to as value consciousness. The seventems for the measurement of these two factors were developedased on Lichtenstein, Ridgway and Netemeyer. For brand loy-lty, we used six items developed based on the literature onhe role of brand loyalty in price information processing (e.g.,elsen and Schmittlein 1994; Zeithaml 1982), while store loy-

lty was measured by four items related to different aspects oftore loyalty previously addressed in the literature (e.g., Mägi003; Mulhern and Padgett 1995; Urbany, Dickson, and Sawyer000). Both the price involvement and loyalty measures usedeven-point Likert scales anchored by (1) “strongly disagree”nd (7) “strongly agree.” The coding of the brand and store loy-lty items was specific to the product category and the store,espectively (see Appendix A).

Results

We interviewed 1,204 shoppers: 395 before, 420 at thehelves, and 389 after the store visit. The interviews were almostvenly distributed across the eight product categories and thewo stores. Of the 1,204 respondents, 1,040 returned a completeduestionnaire (86.4 percent): 341 of the respondents interviewedt the entrance, 360 of the respondents interviewed at the shelves,nd 339 of the respondents interviewed at the exit. We found noignificant differences in consumer price knowledge between


espondents who returned the questionnaire and those who didot. Sample matching is critical to the validity of a between-ubjects design. According to Table 3, there are no significant

dak

a Price recall accuracy measures the absolute deviation relative to the correctrice (see e.g., Estelami and Lehmann 2001).

ifferences between the before, during and after samples on aumber of variables potentially relevant to the price knowledgeeasures, which suggests that the efforts to ensure comparability

etween the samples worked as intended.

easures of price knowledge before, during and after storeisit

We show the percentage of responses that indicated cor-ect price knowledge before, during, and after the store visitn Table 4. Between 7.6 and 43.8 percent of the respondentsecalled the correct price during the shopping trip. With visualues, in the form of a card with six alternative prices, between1.4 and 60.7 percent of the respondents recognized the correctrice (cf. a 16.7 percent chance that the respondent would selecthe correct price randomly). For the deal spotting test, a majorityf respondents could identify a good or bad deal that differedy 10 percent from the normal price (i.e., correct responses toll four deal spotting items) at each measurement point. Thereas a 6.3 percent likelihood of answering all four deal spottinguestions correctly by guessing. The segment of deal obliviousonsumers (not even able to spot a deal at the 20-percent mar-in) was very small, ranging from 12.5 to 16.1 percent of theespondents,7 and so the vast majority was able to spot a deal athe 20-percent margin. We operationalized price recall accuracys the absolute deviation from the actual price expressed as aercentage of the actual price (Estelami and Lehmann 2001).n average, among respondents entering the store, price recall

stimates deviated 19.3 percent from the actual price. Thus con-umers’ price expectations appeared quite inaccurate on entryhile mean price recall errors at the shelf (7.9 percent) and the

xit (9.4 percent) were much lower after an opportunity to updatetem-specific price knowledge.

In line with this, the percentage of correct responses in thehree price memory tests was significantly lower at the store entry

2


7 The proportion of deal oblivious shoppers does not change significantlyuring grocery shopping (χ2 = 2.9; p = .234); if a shopper enters the store with

low perception of the self-relevance of prices, the opportunity to update pricenowledge should have no effect.



al of Retailing xxx (xxx, 2014) xxx–xxx 9

pcs

beran

Gk

ucoshiscuR

esr((vsRvRaisiarrdac

c.anmsr

Table 5Determinants of the level of price knowledge before, during, and after store visit.

Before During After

B Exp(B) B Exp(B) B Exp(B)

Threshold 1 −.65 −.65 −.95Threshold 2 1.18 .41 .12Threshold 3 2.58 1.23 .90

VariablesPrice consciousness .08 1.08 .16 1.17 .28** 1.32Value consciousness .18 1.20 .35** 1.42 .25* 1.28Brand loyalty .23* 1.26 −.08 .92 −.13 .88Store loyalty .11 1.12 −.10 .90 .09 1.09Price range −.63** .53 .54* 1.72 .28 1.32Deal share .16 1.17 .66** 1.93 −.22 .80Buying frequency .39 1.48 .64** 1.90 .62** 1.86

Chi-square p < .001 p < .0001 p < .0001Cox and Snell R2 .04 .11 .07Nagelkerke R2 .04 .12 .07Test of parallel lines p = .93 p = .58 p = .33

Note. We measured shopper demographics (gender, age, household size, andincome) and perceived budget constraints and used them as control variables,but they had no explanatory power.

Ds

mukrvta

iaTnalova


< .001; and deal spotting χ2 = 11.7; p = .001). In fact, the per-entage of correct price recall increased almost fivefold from thetore entry to the exit.

Interestingly, however, no significant differences emergedetween measured price knowledge at the shelves and at thexit. Hence, the proportions of consumers with correct priceecall (χ2 = 3.3; p = .069), price recognition (χ2 = 2.2; p = .138),nd deal spotting ability (χ2 = 2.1; p = .145) at the shelves wereot significantly higher than at the store exit.

uttman scalogram analysis of relationship between pricenowledge measures

Our results in Table 4 suggest that the price knowledge meas-res may be hierarchically interrelated in that the percentage oforrect responses decreases with the difficulty of the price mem-ry test. To verify this interpretation, we conducted a Guttmancalogram analysis. In a Guttman scale, a positive response to aigher-order item implies the same response to all lower-ordertems, such that individual response patterns form a joint ordinalcale. Perfect forms rarely occur empirically, but this analysisan reveal how closely a set of items (i.e., price knowledge meas-res) corresponds to a unidimensional continuum (Menzel 1953;obinson 1973).

First we assessed how often responses to the price knowl-dge measures—price recall, price recognition, and dealpotting—conformed to a perfect Guttman scale by using theigid Goodenough-Edwards technique to count scale errorsMcIver and Carmines 1982). The coefficient of reproducibilityCR) score was .95 before, .90 during,8 and .92 after the storeisit in excess of conventional minimum criteria (CR ≥ .90) andignificantly higher than the chance level (Minimum Marginaleproducibility is .75 before, .63 during, and .62 after storeisit) at each measurement point (McIver and Carmines 1982;obinson 1973). Because the CR alone is insufficient to indicate

Guttman scale, we also calculated the coefficient of scalabil-ty (CS), that is, the extent to which the item responses can becaled on one dimension. The CS scores—.79 before, .73 dur-ng, and .79 after the store visit—were well above the minimumcceptable level of .60–.65 (Menzel 1953). In summary, theseesults supported the pattern from Table 4: deal spotting, priceecognition, and price recall represent measures of increasingifficulty on a joint ordinal scale. This conclusion was robustcross the three measurement points, thus supporting a hierar-


hical relationship between measures.

8 Although, the CR score of .902 at the shelf equals the conventional minimumriteria (CR ≥ .90), it is still significantly higher than the chance level (MMR is63), while the corresponding CS score (.73) also by far exceeds the minimumcceptable level. In addition, these results were based on the most rigid tech-ique to count Guttman scale errors. The less rigid Guttman technique producesarkedly higher CR (.98 before, .95 during, and .97 after store visit) and CS

cores (.90 before, .87 during, and .91 after store visit), thus emphasizing theobustness of our conclusions.

vasrsis

ocn

* Significant at p < .05.** Significant at p < .01.

eterminants of price knowledge before, during and aftertore visit

To take into account the ordinal relationship between theeasures revealed by the Guttman scalogram analysis, we

sed ordinal regression analysis with four levels of pricenowledge—correct price recall, correct price recognition, cor-ect deal spotting, and none of the above—as the dependentariable to analyze determinants of price knowledge. We believehat the great parsimony with this approach overcompensates forny potential loss in information.

In preparation we ran principal component analyses ontems measuring price involvement, store and brand loy-lty to avoid multi-collinearity problems (see Appendix A).he analysis produced four components: price conscious-ess (Cronbach’s alpha = .82), value consciousness (Cronbach’slpha = .71), brand loyalty (Cronbach’s alpha = .91), and storeoyalty (Cronbach’s alpha = .76). Thus, we used the factor scoresf these four components in the ordinal regressions. Table 5 pro-ides the parameter estimates for the models before, during, andfter store visit.

The findings in Table 5 show a positive effect of price and/oralue consciousness on the level of price knowledge during andfter store visit, but not before store visit. Price and value con-ciousness may both lead to more elaborate price processing,esulting in accessible episodic price knowledge not only at thehelf, but also at the exit. The absent effect at the entrance maynstead reflect time elapsed or exposure to conflicting pricesince the last store visit.

We found no significant effect of brand loyalty on the level


f price knowledge at the shelf or at the exit, thus brand loyalonsumers appear to pay just as much attention to price ason-brand loyal consumers. In addition, we found a significant



1 al of R

pbeltc(sas

weelWbach

ssptimalof

cwotnklu

fa

hs

ra(f

HnoS

ahdpprstc(taropTkt

(popcpoTftplebppaw

egreat accuracy, our results suggest that some consumers mayhave learned about prices unconsciously. One might arguethat it is merely the information-rich store environment that

9 There may be several ways in which episodic price knowledge is used to auto-matically update reference prices. Yet, analysis of such mechanisms is outsidethe scope of this paper.10 Some respondents might have noticed the particular item’s price at checkout

as products were scanned or if checking their receipt. However, the interview

0 B.B. Jensen, K.G. Grunert / Journ

ositive effect of brand loyalty before the store visit, perhapsecause brand loyal consumers keep fewer prices in mind andxperience repeated exposure to these prices. However, storeoyalty had no effect on the level of price knowledge at any ofhe measurement points. One explanation could be that mostonsumers patronize several stores on a regular basis todaye.g., Mägi and Julander 2005), while consumers primarilyhopping at one store also may be somewhat price involvednd, for instance, scan shelves for specials in that particulartore (Urbany, Dickson, and Sawyer 2000).

Product categories characterized by high buying frequencyere associated with significantly higher levels of price knowl-

dge at the shelf and the exit. Greater economic incentive tongage in active price search in such categories increases theikelihood of episodic price knowledge at the shelf and the exit.

e found no significant (p = .09) effect of buying frequencyefore store visit, although it had the expected sign and met

p = .10 criterion. More frequent and recent price exposuresould result in more accurate pre-visit price expectations inigh-frequency categories.

In addition, high deal share categories were associated withignificantly higher levels of price knowledge at the point ofelection. We found no effect of deal share at the exit though,erhaps because promotional prices are more salient in memoryhan normal prices for high deal share categories, thus resultingn uncertainty regarding which price to retrieve from long-term

emory (Johnson 1994). It is less clear why we find no effectt the entrance. A negative effect of increased price variationinked to a high deal share may be balanced by a positive effectf increased likelihood of exposure to prices in store advertising,or example, sales flyers before store visit.

Price range exhibits two opposite effects in Table 5. Productategories characterized by a large price range were associatedith significantly higher levels of price knowledge at the pointf selection, possibly due to large price distributions increasinghe benefits of price search. Yet, we found a highly significantegative relationship between price range and level of pricenowledge before store visit. If the category functions as a priceevel cue of any given item, larger price ranges could cause morencertainty about the actual price of a particular item.

Discussion

We examined consumer price knowledge by three dif-erent measures before, during, and after store visit, whichllows us to answer the three research questions we put forward.

Question 1: How price knowledgeable are consumers andow does this knowledge differ between the three stages of atore visit?

Comparison to previous studies. When focusing on the price


ecall measure during and after store visit, our results (Table 4)re generally comparable with those found in past studiesTable 1). Before store visit, only Vanhuele and Drèze (2002)ound a lower percentage of correct price recall than we did.

totw


owever, except for their study, past pre-store-visit studies haveot taken place close to the shopping trip, nor have they focusedn a latent or manifest purchase intention contrary to our study.uch differences further complicate cross-study comparisons.

As for the other two measures, comparisons are only possiblet the entrance. Here Vanhuele and Drèze found a significantlyigher percentage of correct price recognition than we did, likelyue to key differences in the level of difficulty between the tworice recognition tests. Their recognition test consisted of threerices (actual price and two prices 10% above and below); ourecognition test was far more cognitively demanding depictingix prices with smaller percentage intervals. Yet, compared tohese authors, we found a considerably higher percentage oforrect deal spotting, which is in line with Urbany and Dickson1991). By applying a price recognition and a deal spottingest at the shelf and at the exit, we demonstrate that consumersre more price knowledgeable during grocery shopping thanevealed by the price recall test alone. The combined resultsf the price memory tests contradict the general opinion fromast research, viz. that of poor consumer price knowledge.hat is, by including measures of different levels of pricenowledge, we show that consumers know more about priceshan suggested by past research.

Episodic price knowledge. According to Monroe and Lee1999), measurements at the shelf measure in-store attention torice, whereas measurements at the exit reveal long-term mem-ry of price. Yet we find no significant decrease in accuraterice recall between these two measurement points. A largeonsumer segment thus appears to engage in intentional pricerocessing during grocery shopping resulting in strong mem-ry traces and accessible episodic price knowledge at the exit.hese findings suggest that episodic price knowledge does not

ade as fast as might be expected; the speed may depend onhe level of price processing. That is, the less elaborate therice processing, the faster it fades. Still, the significantly lowerevel of correct pre-visit price knowledge indicates that suchpisodic price knowledge fades and becomes difficult to retrieveefore the next store visit due to time or exposure to conflictingrices in between store visits. The findings suggest that in-storerocessed prices are not solely used to update reference prices9

s accessible episodic price knowledge at the exit seems far moreidespread than expected.10

Reference prices. Though our methodology may not bequipped to distinguish unconscious price processing with


ook place at the store exit, often several minutes later and after processing a lotf unrelated information since the checkout. Thus, no matter what, the answerso the price knowledge questions would originate from long-term memory ande therefore believe that our results warrant this conclusion.



al of R

amieaeed

aspetcphcrmtmpr

khafegWkecrfs

diearttkowpm

c

c

nctpritomricaoepoatp

pe

aeodvwepas

refckbhibap

sere


ctivates semantic memory making existing reference pricesore accessible, and that no learning in fact takes place. Yet,

n addition to the general argument about whether activation ofxisting knowledge is a form of learning or not, we also note

significant increase in consumers’ deal spotting ability fromntering to leaving the store. This result may emerge due to ref-rence prices being automatically updated with the current priceuring the store visit.

The vast majority of consumers thus appear to either learnbout prices consciously or unconsciously during groceryhopping, and such price learning may account for the referencerice effects reported in reference price studies (e.g., Briescht al. 1997; Winer 1986). Reference price models assumehat consumers retain memory-based reference prices to judgeurrent prices updating this knowledge as they shop. The poorrice knowledge generally reported in price knowledge studiesas challenged reference price models’ ability to predict brandhoice, as it indicates that most consumers do not notice oremember actual prices (Monroe and Lee 1999). By applyingeasures of different levels of price knowledge, we are able

o show that, overall, price-ignorant consumers represent ainority compared with consumers with fairly operational

rice knowledge. Our results thus seem in less conflict witheference price research.

Reference price research tends to assume item-specific pricenowledge and thus brand-specific tests of price knowledgeave been employed. However, according to Mazumdar, Raj,nd Sinha (2005), item-specific reference prices are not likelyor many products. Instead, consumers may have formed pricexpectations at a more aggregate level, for example, the cate-ory level, which might still be relevant for decision-making.e propose that associations to accessible item-specific price

nowledge may weaken before the next store visit due to timelapsed and/or exposure to conflicting price information, thusreating uncertainty about the price of any given item. In turn,eference prices may be accessible only at an aggregate levelor the next store visit. In fact, findings from Table 5 appear toupport this claim (see discussion of Question 3 below).

Reconciliation of findings. Our findings suggest that theifferences in the results of previous work can be reconciledn the following view of how consumer price knowledgevolves during grocery shopping: Accessible price knowledget the entrance primarily consists of category or item-specificeference prices; accessible episodic price knowledge is rare. Athe shelf, consumers may or may not consciously pay attentiono the price. If conscious price learning occurs, episodic pricenowledge becomes accessible at the exit. In addition, even ifnly unconscious learning occurs, the reference price is updatedith the current price. Before the next shopping trip, episodicrice knowledge, however, fades, making reference prices theost salient again.

Question 2: Are the measures of price knowledge hierarchi-


ally related?

Vanhuele and Drèze (2002) reported the finding thatonsumers who answer a price recall question correctly are

eei


ot necessarily able to use this knowledge to answer a lessognitively demanding question. Our study does not supportheir finding. On the contrary we show that the measures ofrice knowledge are hierarchically related, a conclusion that isobust across three measurement points. Vanhuele and Drèzenterpret their surprising finding as evidence that the questionsap different memory dimensions and relate it to researchn numerical cognition, which suggests that numbers can beentally represented in different forms (Dehaene 1992). Such

esearch, however, remains nascent in relation to prices in thatt rests on simplified tasks and number combinations unlike theomplex prices consumers typically face in a store (Monroend Lee 1999; Vanhuele, Laurent, and Drèze 2006). We, on thether hand, argue that different levels of price processing mayxplain differences in the level of price knowledge. Referencerices and/or episodic price knowledge may vary in strengthf memory trace depending on the level of price processingnd time elapsed since price exposure. The strength of memoryraces thus may explain the need for retrieval cues to activaterice knowledge from long-term memory.

Question 3: How do brand and store loyalty, categoryurchase frequency, price range, and deal share add to thexplanation of the variance in consumer price knowledge?

So far, we have concluded that timing and type of measurespplied can explain part of the variance in measured price knowl-dge between previous studies. Still, within the individual cellsf Table 1, we observe wide variations partly due to other studyesign differences. For instance, before store visit, in-home sur-eys produced significantly higher price recall scores comparedith store entrance interviews. In addition, we find significant

ffects of each of the three product characteristics on consumerrice knowledge. Thus, not controlling for these factors maylso have added to the variance in results between previoustudies.

We replicate the finding in earlier research that price-searchelated consumer traits affect the level of price knowledge, yetxtend past research by revealing that factors that affect therequency and variation of consumer exposure to prices alsoontribute to the explanation of differences in consumer pricenowledge. This is particularly evident at the entrance whererand loyalty and price range (and partly buying frequency)ave significant effects. By including less cognitively demand-ng price memory tests, we demonstrate that price knowledgeuilds up not only because of active price search, but also due toccidental exposure to prices and with low degrees of consciousrocessing.

The findings associated with value and price consciousnessupport the reconciled view of how consumer price knowledgevolves during grocery shopping as both may lead to elabo-ate price processing including price comparisons that makepisodic price knowledge accessible at the exit. The positive


ffect of these two predictors at the exit suggests that the consid-rable level of episodic price knowledge in long-term memorys attributable to a large degree of intentional price processing.



1 al of R

Tca

lljTabtctmtoieair1wpp

ldrsTipccnvioacTmbarm

slpaea

cpt5pppabcKrr

Aalctteacd

pap(scsifwabp

rcnefircr“fcbpe


he absence of an effect of these two factors at the entrance isonsistent with the notion that episodic price knowledge fadesnd thus becomes difficult to access over time.

Two findings support the claim related to the aggregationevel of reference prices. First, the lack of an effect of brandoyalty at the shelf indicates that brand loyal shoppers payust as much attention to prices in-store as do brand switchers.he positive pre-visit effect of brand loyalty thus cannot bettributed to differences in the level of in-store price checkingehavior. Rather, brand loyal consumers appear more likelyo retain item-specific reference prices because their limitedhoice sets give them fewer prices to keep track of and dueo repeated exposures to these prices. Brand switchers insteadight retain reference prices at more aggregate levels since

hey consider more divergent prices, and exposure to the pricef a particular item occurs only after longer time gaps. Second,n line with Vanhuele and Drèze (2002), we find a negativeffect of price range before the store visit. Large price rangesppear to create uncertainty about the actual price of any giventem, such that the category becomes an unreliable cue foretrieving price from long-term memory (Anderson and Bjork994). This characteristic is tied to the category level, so weould not expect this negative price range effect if referencerices were always item specific. Instead, for some consumersre-visit reference prices appear to be category specific.

Managerial implications

From a managerial point of view, our results suggest that aarge segment of consumers pay conscious attention to pricesuring grocery shopping thus underpinning the importance ofetailers’ in-store price communication efforts. The results alsouggest that some consumers learn about prices unconsciously.ogether, it seems that the vast majority of consumers learn price

nformation. While we provide evidence that in-store prices areerceived by most consumers and thus may influence brandhoice, we provide no evidence of their relevance for storehoice. The finding that specific prices are forgotten before theext store visit might suggest that objective prices are less rele-ant in explaining store choice. However, in case episodic pricenformation is integrated in an overall price image perception,ur findings could justify grocery retailers’ focus on price asn important marketing tool to promote their price image, espe-ially for retailers trying to trade at low/the lowest price levels.he positive effect of brand loyalty on pre-visit price knowledgeay indicate that for some products objective prices are remem-

ered and therefore could be relevant for forming a price imagend influence store choice. Hence, retail managers can use theesults to implement a mixed calculation policy with lower retailargins on high brand loyalty products.Time elapsed and conflicting price exposures from other

tores or brands in between store visits are likely to result inearned price information fading, and based on the average


rice recall estimates, which were off by more than 19 percentt the entrance, many shoppers hold fairly inaccurate pricexpectations entering the store. Retailers may therefore beble to implement minor stepwise price increases without most

S

ko


onsumers detecting it. This may be a cause for concern from aublic policy perspective, though the results of the deal spottingest may mitigate such concerns. With visual cues almost5 percent of the respondents were able to spot a price 10ercent higher than the normal price as a bad deal (almost 84ercent spotted a price 20% higher). Thus, studies relying onrice recall alone could lead to biased managerial decisionsbout the level of price increases because managers mistakenlyelieve they are undetectable. Stepwise price increases outsideonsumers’ reference price zones are likely to be detected (e.g.,alyanaram and Little 1994), and considering asymmetric

eactions to price changes this might result in quite negativeeactions (Kalyanaram and Winer 1995).

We find significant differences across product categories (seeppendix B) partly explained by price range, buying frequency

nd deal share. These factors were not measured at the individualevel; they were based on a categorization of the eight productategories into relatively high and low levels on each of thehree product characteristics. Nevertheless, the results suggesthat retailer pricing strategies should be adapted to different cat-gories accordingly. Minor stepwise price increases would alsoppear to be harder to detect in low frequency product categoriesharacterized by a relatively large price range as well as a lowegree of brand loyalty.

Contradicting the general view of brand loyal consumers asrice insensitive, our results indicate that they pay just as muchttention to price as non-brand loyal consumers supposedly asart of the quantity decision: to stock up on price specials or notKrishnamurthi and Raj 1991). For grocery items suitable fortock piling, brand managers should acknowledge the risk of pur-hase acceleration/forward buying where brand loyal consumerstock up their favorite brand and/or postpone their purchase ift is sold at normal price. This potentially implies that sale atull price drops resulting in loss of contribution margin. In otherords, a large loyal customer base should not lull brand man-

gers into false security, and so brand loyal consumers’ forwarduying must be taken into consideration when brand managerslan future promotion activities.

According to our results, store loyalty appears not to beelated to consumer price knowledge. Apparently, store loyalonsumers pay just as much attention to prices in-store as doon store-loyal consumers; the latter even include ‘cherry pick-rs’. This is in line with Urbany, Dickson, and Sawyer’s (2000)nding that almost 80 percent of “mostly one store shoppers”eported scanning the shelves for price specials. With very fewonsumers sticking to one store for their grocery shopping,etailers want to increase share of wallet and share of visits ofmostly one store shoppers” Our results imply the importanceor store managers of catering to this store loyal segment withompetitive prices and attractive bargains to keep them comingack and/or spending a larger share of their wallet in the store,erhaps even promote non-featured specials locally, which ben-fits this segment the most (Mägi 2003; Urbany, Dickson, and


awyer 2000).In case large retailers are interested in carrying out a price

nowledge survey of their own, they would most likely focusn one measurement method and a single point in the buying



al of R

pbtmcsamsiaipoerop

tdmWfaaltCiaewecdaposgt

dfsavar

ekspaFea

acbvospicv

wpipaeuisaasotuv

paaf(araLlpt


rocess. In selecting the measurement point, they would proba-ly avoid in-store measurements as they do not like disruptinghe in-store environment and in-store interviews are typically

ore time consuming (e.g., the wait between shoppers approa-hing low-frequency categories while store traffic is slow). Ourtudy provides insights useful for selecting the right methodnd time point. First, if managers are interested in getting infor-ation about reference prices, the deal spotting test is fairly

imple to do and would provide retailers with knowledge oftem and/or category-specific reference prices when carried outt the entrance. If, however, managers are interested in uncover-ng to which extent shoppers did indeed pay attention to in-storerice information (for example with regard to price specials),ur finding of no significant decrease in episodic price knowl-dge after store visit suggests that retailers could conduct a priceecall and/or a price recognition test at the exit and would stillbtain fairly accurate knowledge about their customers’ in-storerice search behavior.

Limitations and further research

Several limitations apply to this study and point to oppor-unities for further research. Although we found no significantifferences in measured price knowledge between the hyper-arket and the supermarket we studied, McGoldrick, Betts, andilson (1999) show that price knowledge can vary across store

ormats. Specifically, they found significantly lower price recallt an EDLP retailer (discounter) compared to two stores oper-ting with a Hi-Lo price format. We have not considered priceearning in a discount store setting, though the growing impor-ance of this format in Western Europe and North America (e.g.,leeren et al. 2010) suggests the need for research that examines

f price learning evolves differently in such a setting. We found negative effect of price range on the level of price knowl-dge before store visit. Assuming that discount stores operateith a limited price range in most product categories, one might

xpect more accurate price knowledge at the entrance of a dis-ount store. For instance, some product categories at a hardiscounter may be limited to few alternatives with limited qualitynd price differences. In such cases, category-specific referencerices could be fairly accurate. In addition, the positive effectf price range on the level of price knowledge at the point ofelection suggests that limited price ranges at a hard discounterenerally could result in less in-store attention paid to prices dueo moderate benefits of such activity.

Our choice of a between-subjects design reflects the seriousrawbacks of a within-subjects design, but it also prevented usrom recording the time elapsed between price exposure at thehelf and the store exit. Individual data might show how fastnd why episodic price knowledge fades before the next store


isit by revealing exactly which item was chosen, at which store,nd when. Consumer panel data might be an option, although aehearsal bias remains a potential concern (Conover 1986).

Umk


Previous price knowledge studies generally focus on ref-rence prices without distinguishing between episodic pricenowledge and reference prices. Though we do not mea-ure reference prices directly, that is, we infer referencerices from other measurements, we make a first attempt atddressing both episodic price knowledge and reference prices.uture studies should consider both types of price knowl-dge and try to separate them, perhaps applying experimentalpproaches.

It could be argued that basing the measurement of price recallt the store entrance on actual prices may affect our ability tolaim that episodic price knowledge from one store visit fadesefore the next as the actual price may differ between storeisits. However, price recall results excluding interviews basedn special-priced items (entrance 4.9%, shelf 35.5%, exit 32.7%)how exactly the same pattern as in Table 4. Given that normalrices of most groceries are fairly stable in the short run, thats, in between store visits, these findings seem to support ourlaim of fading episodic price knowledge before the next storeisit.

In line with Mazumdar, Raj, and Sinha’s (2005) observations,e note that at least some consumers appear to retain referencerices at an aggregate level before the store visit. More researchs, however, needed to confirm this prediction. We also measurerice knowledge for one store as opposed to across stores. Wessume that, with limited cognitive capacity, consumers gen-rally do not retain store-specific reference prices in memory,nless they exclusively patronize a single store to purchase antem. We do not know though when price knowledge is storepecific and when it spans stores. We found no effect of store loy-lty on consumer price knowledge at the entrance, which goesgainst expectations of store loyal consumers retaining store-pecific reference prices. Such expectations are, however, basedn the assumption that price variation within a store is lowerhan across stores. Future studies could verify this assumptionsing scanner data to reveal the level of price variation withinersus across stores.

Finally, there is no particular reason to expect that consumerrice knowledge evolves differently during grocery shoppingcross countries, though the level of price knowledge observedt particular points in time in the buying process may dif-er under varying sociocultural and macro-economic conditionsEstelami, Lehmann, and Holden 2001). For example, Vanhuelend Drèze (2002) explain the low proportion of correct priceecall in their study by suggesting that French consumers pay lessttention to prices than their counterparts in the US. Vanhuele,aurent, and Drèze (2006) found a negative effect of verbal

ength (number of syllables) and price usualness on immediaterice recall. Both these factors may vary across countries dueo the face value of currencies (e.g., Hungarian forint versus


S dollar) and the strength of price ending traditions. In sum-ary, we call for cross-country comparisons of consumer price

nowledge.



1 al of Retailing xxx (xxx, 2014) xxx–xxx

A ent, brand loyalty and store loyalty: Rotated factor pattern

I Price consciousness Value consciousness Brand loyalty Store loyalty

T .82 .12 −.06 −.06

I .77 .14 −.08 −.29I .74 .14 −.10 −.07I .73 .33 −.09 −.06I .05 .75 −.00 −.13

I .41 .73 −.07 .06

W .33 .72 −.11 .05

I −.03 −.08 .89 .08I −.02 −.00 .88 .13I −.15 −.13 .85 .07

I −.08 −.09 .78 .07

G −.11 −.23 .78 −.01I −.03 .05 .74 .06I −.22 .04 .09 .79I .15 −.16 −.05 .74I −.08 .08 .08 .73I −.36 −.06 .11 .72

C .82 .71 .91 .76

N ponent. (R) indicates that the item was reversely scored.

A ss product categories

After

ll Price recognition Deal spotting Price recall Price recognition Deal spotting

M 44.0% 56.0% 29.4% 43.1% 70.6%T 61.8% 67.3% 38.0% 43.0% 46.0%M 64.0% 84.0% 42.0% 60.0% 75.5%J 56.0% 74.0% 45.8% 63.3% 73.5%O 71.2% 73.1% 24.4% 68.9% 80.0%S 69.8% 81.1% 29.2% 47.3% 63.7%K 44.8% 63.8% 42.9% 50.4% 51.0%G 74

A

A

B

B

C

C

C

D

D

E


ppendix A. Principal component analysis of price involvemafter Varimax rotation

tems

he money saved by finding low prices is usually not worth the time and effort.(R)

grocery shop at more than one store to take advantage of low prices. am not willing to go to extra effort to find lower prices. (R)

compare prices of different stores to get the best price.

am very concerned about low prices, but I am equally concerned aboutproduct quality.

always check prices at the grocery store to be sure I get the best value for themoney I spend.hen grocery shopping, I compare the prices of different brands to be sure Iget the best value for the money.

switch between different brands. (R)

buy the same brand every time.

prefer to buy my favorite brand of regardless of the price of otherbrands.

don’t believe other brands of can fulfill my needs as well as my favoritebrand.

enerally, I compare different brands when I buy . (R)

postpone buying if my favorite brand is out of stock.

shop more frequently at than at other grocery stores.

t is often random where I grocery shop. (R)

know exactly where the different products are located in .

grocery shop at many different stores. (R)

ronbach’s alpha

ote. Loadings in boldface indicate the items that load the highest on each com

ppendix B. Differences in consumer price knowledge acro

Before During

Price recall Price recognition Deal spotting Price reca

argarine 10.4% 28.0% 64.0% 24.0%

oothpaste 7.7% 25.0% 55.8% 41.8%

uesli 4.7% 34.9% 67.4% 36.7%

uice 9.8% 19.6% 58.8% 52.0%

live oil 3.8% 15.4% 34.6% 38.5%

andwich chocolate 4.2% 18.8% 58.3% 53.8%

etchup 5.9% 13.7% 43.1% 32.8%

round coffee 14.0% 18.0% 60.0% 70.4%

References

lba, Joseph W., J. Wesley Hutchinson and John G. Lynch Jr. (1991), “Mem-ory and Decision Making,” in Handbook of Consumer Behavior, RobertsonThomas S. and Kassarjian Harold H., eds. Englewood Cliffs: Prentice-Hall,Inc, 1–49.

nderson, Michael C. and Robert A. Bjork (1994), “Mechanisms of Inhibition inLong-Term Memory: A New Taxonomy,” in Inhibitory Processes in Atten-tion, Memory and Language, Dagenbach D. and Carrs T., eds. San Diego:Academic Press, 265–325.

ell, David R., Daniel Corsten and George Knox (2011), “From Point of Pur-chase to Path to Purchase: How Preshopping Factors Drive Unplanned


Buying,” Journal of Marketing, 75 (1), 31–45.riesch, Richard A., Lakshman Krishnamurthi, Tridib Mazumdar and S.P. Raj

(1997), “A Comparative Analysis of Reference Price Models,” Journal ofConsumer Research, 24 (September), 202–15.

E

.1% 75.9% 46.9% 69.4% 73.5%

hang, Tien Minh (1986), “Semantic Memory: Facts and Models,” Psycholog-ical Bulletin, 99, 199–220.

leeren, Kathleen, Frank Verboven, Marnik Dekimpe and Katrijn Gielens(2010), “Intra- and Interformat Competition among Discounters and Super-markets,” Marketing Science, 29 (3), 456–73.

onover, Jerry N. (1986), “The Accuracy of Price Knowledge: Issues inResearch Methodology,” Advances in Consumer Research, 13, 589–93.

ehaene, Stanislas (1992), “Varieties of Numerical Abilities,” Cognition, 44,1–42.

ickson, Peter R. and Alan G. Sawyer (1990), “The Price Knowledge and Searchof Supermarket Shoppers,” Journal of Marketing, 54 (3), 42–53.

stelami, Hooman and Donald R. Lehmann (2001), “The Impact of ResearchDesign on Consumer Price Recall Accuracy: An Integrative Review,” Jour-


nal of the Academy of Marketing Science, 29 (1), 36–49.stelami, Hooman, Donald R. Lehmann and Alfred C. Holden (2001), “Macro-

Economic Determinants of Consumer Price Knowledge: A Meta-Analysis of


http://refhub.elsevier.com/S0022-4359(14)00002-5/sbref0005
















































































































































































































































































al of R

G

H

I

J

K

K

K

K

K

L

L

M

M

M

M

M

M

M

M

M

M

M

N

R

R

S

T

T

U

U

U

V

V

W

W

W


Four Decades of Research,” International Journal of Research in Marketing,18 (4), 341–55.

abor, Andre and Clive Granger (1961), “On the Price Consciousness of Con-sumers,” Applied Statistics, 10 (November), 170–88.

elsen, Kristiaan and David Schmittlein (1994), “Understanding Price Effectsfor New Nondurables: How Price Responsiveness Varies across Depth-of-Repeat Classes and Types of Consumers,” European Journal of OperationalResearch, 76, 359–74.

nman, Jeffrey J., Russell S. Winer and Rosellina Ferraro (2009), “The Interplayamong Category Characteristics, Customer Characteristics, and CustomerActivities on in-Store Decision Making,” Journal of Marketing, 73, 19–29.

ohnson, Hollyn M. (1994), “Processes of Successful Intentional Forgetting,”Psychological Bulletin, 116 (2), 274–92.

alyanaram, Gurumurthy and John D.C. Little (1994), “An Empirical Analysisof Latitude of Price Acceptance in Consumer Package Goods,” Journal ofConsumer Research, 21, 408–19.

alyanaram, Gurumurthy and Russell S. Winer (1995), “Empirical Gener-alizations from Reference Price Research,” Marketing Science, 14 (3),161–9.

rishna, Aradhna, Imram Currim and Robert W. Shoemaker (1991), “ConsumerPerceptions of Promotional Activity,” Journal of Marketing, 55 (2), 4–17.

rishnamurthi, Lakshman and S.P. Raj (1991), “An Empirical Analysis of theRelationship between Brand Loyalty and Consumer Price Elasticity,” Mar-keting Science, 10 (2), 172–83.

umar, V. and Robert P. Leone (1988), “Measuring the Effect of Retail Store Pro-motions on Brand and Store Substitution,” Journal of Marketing Research,25 (May), 178–85.

e Boutillier, John, Susanne S. Le Boutillier and Scott A. Neslin (1994), “AReplication and Extension of the Dickson and Sawyer Price-AwarenessStudy,” Marketing Letters, 5, 31–42.

ichtenstein, Donald R., Nancy M. Ridgway and Richard G. Netemeyer (1993),“Price Perceptions and Consumer Shopping Behavior: A Field Study,” Jour-nal of Marketing Research, 30 (2), 234–45.

ägi, Anne (2003), “Share of Wallet in Grocery Retailing: The Effects of Cus-tomer Satisfaction, Loyalty Cards and Shopper Characteristics,” Journal ofRetailing, 79 (2), 97–103.

ägi, Anne and Claes-Robert Julander (2005), “Consumers’ Store-Level PriceKnowledge: Why Are Some Consumers More Knowledgeable Than Oth-ers?,” Journal of Retailing, 81 (4), 319–29.

azumdar, Tridib and Kent B. Monroe (1990), “The Effects of Buyers’ Inten-tions to Learn Price Information on Price Encoding,” Journal of Retailing,66 (1), 15–32.

azumdar, Tridib, S.P. Raj and Indrajit Sinha (2005), “Reference PriceResearch: Review and Propositions,” Journal of Marketing, 69 (4), 84–102.

cGoldrick, Peter J. and Helen J. Marks (1987), “Shoppers’ Awareness of RetailGrocery Prices,” European Journal of Marketing, 21 (3), 63–75.

cGoldrick, Peter J., Erica J. Betts and Alexandra F. Wilson (1999), “Mod-


elling Consumer Price Cognition: Evidence from Discount and SuperstoreSectors,” Service Industries Journal, 19 (1), 171–93.

cIver, J.P. and E.G. Carmines (1982), Unidimensional Scaling, 2nd ed. BeverlyHills: Sage Publications.

Z


enzel, Herbert (1953), “A New Coefficient for Scalogram Analysis,” PublicOpinion Quarterly, 17 (Summer), 268–80.

onroe, Kent B. and Angela Y. Lee (1999), “Remembering Versus Knowing:Issues in Buyers’ Processing of Price Information,” Journal of the Academyof Marketing Science, 27 (2), 207–25.

onroe, Kent B., Christine P. Powell and Pravat K. Choudhury (1986), “RecallVersus Recognition as a Measure of Price Awareness,” Advances in Con-sumer Research, 13, 594–9.

ulhern, Francis J. and Daniel T. Padgett (1995), “The Relationship BetweenRetail Price Promotions and Regular Price Purchases,” Journal of Marketing,59 (4), 83–90.

ickols, Sharon Y. and Karen D. Fox (1983), “Buying Time and SavingTime: Strategies for Managing Household Production,” Journal of ConsumerResearch, 10 (2), 197–208.

obinson, John P. (1973), “Toward a More Appropriate Use of Guttman Scal-ing,” Public Opinion Quarterly, 37 (2), 260–7.

osa-Díaz, Isabel María (2004), “Price Knowledge: Effects of Consumers’ Atti-tudes Towards Prices, Demographics, and Socio-Cultural Characteristics,”Journal of Product & Brand Management, 13 (6), 406–28.

ingh, Surendra N., Michael L. Rothschild and Gilbert A. Churchill (1988),“Recognition versus Recall as Measures of Television Commercial Forget-ting,” Journal of Marketing Research, 25, 72–80.

alukdar, Debabrata, Dinesh K. Gauri and Dhruv Grewal (2010), “An Empiri-cal Analysis of the Extreme Cherry Picking Behavior of Consumers in theFrequently Purchased Goods Market,” Journal of Retailing, 86 (4), 336–54.

ulving, Endel (1972), “Episodic and Semantic Memory,” in Organization ofMemory, Tulving Endel and Donaldson W., eds. New York: Academic Press,381–403.

rbany, Joel E. and Peter R. Dickson (1991), “Consumer Normal Price Estima-tion: Market versus Personal Standards,” Journal of Consumer Research, 18(June), 45–51.

rbany, Joel E., Peter R. Dickson and Rosemary Kalapurakal (1996), “PriceSearch in the Retail Grocery Market,” Journal of Marketing, 60 (2), 91–104.

rbany, Joel E., Peter R. Dickson and Alan G. Sawyer (2000), “Insights intoCross- and Within-Store Price Search: Retailer Estimates Vs. ConsumerSelf-Reports,” Journal of Retailing, 76 (2), 243–58.

anhuele, Marc and Xavier Drèze (2002), “Measuring the Price KnowledgeShoppers Bring to the Store,” Journal of Marketing, 66, 72–85.

anhuele, Marc, Gilles Laurent and Xavier Drèze (2006), “Consumers’ Imme-diate Memory for Prices,” Journal of Consumer Research, 33 (2), 163–72.

akefield, Kirk L. and Jeffrey J. Inman (1993), “Who Are the Price Vigi-lantes? An Investigation of Differentiating Characteristics Influencing PriceInformation Processing,” Journal of Retailing, 69 (2), 216–33.

alters, Rockney G. and Maqbul Jamil (2003), “Exploring the Relationshipsbetween Shopping Trip Type, Purchases of Products on Promotion, andShopping Basket Profit,” Journal of Business Research, 56 (1), 17–29.

iner, Russell S. (1986), “A Reference Price Model of Brand Choice


for Frequently Purchased Products,” Journal of Consumer Research, 13(September), 250–6.

eithaml, Valarie A. (1982), “Consumer Response to in-Store Price InformationEnvironments,” Journal of Consumer Research, 8 (March), 357–69.












































































































































































































































































































































































































































































































































































































































































































































































































































































































































































price knowledge during grocery shopping: what we learn and what we forget

Documents