Effects of Objective and Evaluative Front-of-Package Cues on ......([email protected]) is professor of marketing, Sam M. Walton College of Business, University of Arkansas,

Effects of Objective and EvaluativeFront-of-Package Cues on Food Evaluationand Choice The Moderating Influence ofComparative and NoncomparativeProcessing Contexts

CHRISTOPHER L NEWMANELIZABETH HOWLETTSCOT BURTON

Many nutrition labeling studies only consider how consumers process health infor-mation about a single food product (ie in a noncomparative processing context)However consumers also often comparatively evaluate many different foodproducts at once in more complex shopping environments (ie in comparative pro-cessing contexts) Directly addressing these important differences the results oftwo online studies and two retail laboratory studies demonstrate that the effects ofdifferent types of front-of-package nutrition cues (objective vs evaluative) varyacross consumersrsquo processing contexts (comparative vs noncomparative) Whenconsumers evaluate a single food item in a noncomparative context objective nutri-tion cues that offer specific quantitative information lead to higher evaluations andintentions to purchase healthier products than do evaluative nutrition cues (whichprovide interpretive information about a productrsquos overall healthfulness andor nutri-ents) However these effects are reversed when consumers evaluate multiple fooditems simultaneously in a comparative context such that evaluative cues have amore positive impact on evaluations and purchase intentions of healthier productsThe authors integrate processing fluency and resource matching theoretical frame-works to explain why evaluative (objective) front-of-package cues are more influen-tial in comparative (noncomparative) processing contexts Implications for consumerhealth the food and retail grocery industries and public policy are offered

Keywords front-of-package nutrition labeling processing fluency resource

matching theory consumer health comparative and noncomparative processing

nutrition labeling consumer information disclosures food evaluations and choices

Christopher L Newman (cnewmanbusolemissedu) is assistant

professor of marketing School of Business Administration University of

Mississippi PO Box 1848 Oxford MS 38677 Elizabeth Howlett

(ehowlettwaltonuarkedu) is professor of marketing Sam M Walton

College of Business University of Arkansas Fayetteville AR 72701

Scot Burton (sburtonwaltonuarkedu) is distinguished professor and

Tyson Chair in Food and Consumer Products Retailing Sam M Walton

College of Business University of Arkansas Fayetteville AR 72701

This article is based on the dissertation of the first author and the other

two authors contributed equally to the development of the studies and

manuscript The authors thank the reviewers associate editor and editors

for their many insightful comments and suggestions The authors also

thank Melissa Cinelli and Robin Soster for their helpful comments on a

previous draft of this manuscript as well as Chris Berry for his help with

data collection This research was partially supported by the Robert Wood

Johnson Foundation through its Healthy Eating Research Program and by

the SEC Faculty Travel Grant Program The authors are very appreciative

of this support Supplemental materials (such as the stimuli used in each

study) can be found in the online-only version of this article

Laura Peracchio and Darren Dahl served as editors and Jennifer Argo

served as associate editor for this article

Advance Access publication September 27 2015

VC The Author 2015 Published by Oxford University Press on behalf of Journal of Consumer Research Inc

DOI 101093jcrucv050

by guest on August 9 2016

httpjcroxfordjournalsorgD

ownloaded from

INTRODUCTION

Identifying the more and less healthful packaged foodproducts can be a daunting endeavor for shoppers when

one considers that the typical supermarket carries over40000 different itemsmdashthree times more than in 1980(Nestle 2006) The friendly advice frequently offered toconsumers interested in making more healthful foodchoices at supermarkets is to ldquoshop the perimeterrdquo (MayoClinic Staff 2013) It is here that the less processed morehealthful foods such as produce eggs lean beef chickenand fresh seafood can be found However inside the pe-rimeter the healthfulness of processed packaged foods canbe somewhat more difficult to discern quickly andaccurately

This complex retail choice environment provides excel-lent opportunities for marketers to influence shoppersrsquo pur-chase decisions by offering cues to simplify theirevaluative and choice processes (Bettman Luce andPayne 1998) With consumers now making 82 of theirpurchase decisions inside the store (Point of PurchaseAdvertising International 2014) marketing promotions de-signed to influence consumersrsquo product evaluations at thepoint of purchase have become critically important For ex-ample many food manufacturers and retailers are usingfront-of-package (FOP) labeling to attract attention and in-fluence perceptions at the point of sale In contrast to thedetailed Nutrition Facts panel (NFP) shown on the back orside of most packaged foods an FOP label presents con-sumers with a single condensed metric of nutrition infor-mation that is presumed to require less effort and time toprocess

Numerous retailers manufacturers and nongovernmen-tal organizations have implemented a wide variety of FOPlabeling programs such as Walmartrsquos ldquoGreat for Yourdquo ini-tiative Wingmanrsquos Wellness Keys and Unileverrsquos ChoicesProgrammed However these promotional programs varygreatly in terms of both information content and formatSome provide consumers with summarized objective nutri-tion information taken from the NFP others offer evalua-tive nutrition information that provides an interpretation ofa productrsquos overall healthfulness Although the stated pur-pose of FOP nutrition labeling is to ldquoeducate consumersand help them make healthier food choicesrdquo (FederalRegister 2010 see also Food and Drug Administration[FDA] 2015) the types of FOP labeling programs that bestengage and ultimately influence consumersrsquo product evalu-ations and decision-making processes are still poorly un-derstood (Andrews Burton and Kees 2011 Institute ofMedicine [IOM] 2010) Moreover we are not aware of anyprior nutrition labeling research that has specifically con-sidered how the effects of these different systems mightvary according to the particular type of processing task en-countered by consumers (ie noncomparative vs compar-ative evaluative tasks) (Hieke and Taylor 2012)

Therefore across two online experiments and two retaillaboratory experiments we utilize an objectiveevaluativecue framework and a comparativenoncomparative pro-cessing context framework (Oakley et al 2008 Olsen2002) to assess the interactive effects of FOP cue type andprocessing context on consumersrsquo evaluations and inten-tions to purchase healthy food products We also expandon the processing fluency literature (Jacoby and Dallas1981 Lee and Labroo 2004 Schwarz 2004 Whittlesea1993) to explain the processes underlying these effects Inthe following section we first present prior literature andbackground information on fluency different types of FOPcues (objective vs evaluative) and different types of pro-cessing contexts (comparative vs noncomparative) Wethen utilize resource matching theory (RMT) (Anand andSternthal 1987 Keller and Block 1997 Meyers-Levy andPeracchio 1995) to integrate these concepts and to guidethe development of our specific hypotheses

THEORETICAL FOUNDATION

Processing Fluency and Cue Type

Consumers view a product as an array of cues that canbe processed with varying degrees of effort speed andaccuracy (Novemsky et al 2007) Prior research has dem-onstrated that consumers base their judgments not only onavailable cue information but also on the ease with whichthey can process that information and generate relatedthoughts (Schwarz 2004) More specifically processingfluency is defined as the subjective feelings of ease or dif-ficulty with which external information can be processed(Schwarz 2004) Fluency can arise from either the pro-cessing of physical characteristics of a stimulus such asmodality or shape (perceptual fluency) or from the mean-ing of a stimulus (conceptual fluency) (Lee and Labroo2004 Tulving and Schacter 1990 Whittlesea 1993)These two types of fluency represent distinct constructsand have unique antecedents and consequences (Cabezaand Ohta 1993 Lee 2002) For example Labroo and Lee(2006) note that a perceptually fluent brand can be easilyrecognized and identified by consumers whereas a con-ceptually fluent brand is one whose meaning and associa-tions come to mind more easily Thus conceptual fluencyrelates to higher order reasoning and interpretive pro-cesses (Jacoby Kelley and Dywan 1989 Winkielmanet al 2012) Consumers often place more weight on cueinformation that feels easier to process when formingjudgments and making decisions (Shah and Oppenheimer2007)

In this research we propose that alternative types ofFOP nutrition cues lead to varying levels of conceptualfluency across different processing contexts and as a re-sult they have divergent effects on consumersrsquo food evalu-ations intentions and choices To better understand

750 JOURNAL OF CONSUMER RESEARCH

ownloaded from

and predict these differences we first offer an objectiveevaluative cue framework Objective cues provide productinformation that is impartial measurable and objective butgenerally lack a specific interpretive component(Prabhaker and Sauer 1994) Consider the nutrition iconused in the ldquoFacts Up Frontrdquo FOP labeling initiative re-cently developed by the Grocery ManufacturersAssociation (GMA 2013) This cue presents specific ob-jective quantitative information about calories and othercritical nutrients (ie saturated fat sodium and sugars)taken directly from the NFP that allows consumers to as-sess the healthfulness of a single product in a somewhatsimpler information environment (compared to the entireNFP) In contrast evaluative cues provide consumers withinterpretive information with respect to the overall productor a specific product attribute For example evaluativehealth cues can provide consumers with an interpretationof a productrsquos overall healthfulness (eg a more or lesshealthy choice) or certain product attributes (eg low orhigh fat low or high calories) These cues are designed tohelp consumers evaluate products more easily and quicklyby presumably decreasing the cognitive burden of cue in-terpretation and utilization especially in complex settingssuch as supermarkets where consumers encounter manyproducts at once (Feunekes et al 2008) Examples of eval-uative health cues include Walmartrsquos ldquoGreat for Yourdquo andthe IOMrsquos proposed ldquoHealthy Starsrdquo FOP icons

Noncomparative and Comparative ProcessingContexts

We propose that it is critical for researchers firms andpublic health officials to consider whether the effects ofdifferent FOP cues (objective vs evaluative) vary accord-ing to the type of processing context that consumers en-counter (comparative vs noncomparative) Prior researchindicates that consumers can engage in either comparativeor noncomparative processing (Oakley et al 2008 Olsen2002) and that these different processing contexts influenceattitudes intentions and behaviors in unique ways (Hseeand Leclerc 1998 Hsee et al 2013 Naylor Lambertonand West 2012 Nowlis and Simonson 1997) In compara-tive processing contexts where multiple options are avail-able consumers evaluate a product relative to the otherproducts and brands that are present (eg ldquoHow healthfulis this specific cereal relative to other available cereals inthis setrdquo) In order to do so they must first assess all ofthe different available options and then make direct com-parisons between these alternatives In noncomparativecontexts however consumers only need to evaluate asingle product in isolation (eg ldquoHow healthful is this onespecific cerealrdquo) Consumers are not burdened with evalu-ating other products or making explicit comparisonsin these simpler processing contexts (Hsee and Leclerc

1998 Olsen 2002 van Horen and Pieters 2012)Thus comparative processing is often more cognitivelychallenging for consumers than noncomparative processing(Kardes et al 2002)

FOP nutrition cues should help consumers make evalua-tions in these different processing contexts although tovarying extents (Chakravarti and Janiszewski 2003 Slovicand MacPhillamy 1974) The previously discussed differ-ences between cue types (objective vs evaluative) suggestthat a cuersquos impact on perceived fluency (and ultimatelyhealthfulness evaluations and purchase intentions) shouldvary according to the type of processing context encoun-tered by the consumer Drawing from RMT we next pro-pose that evaluative cues will lead to higher perceivedfluency in comparative processing contexts whereas ob-jective cues will lead to higher fluency in noncomparativecontexts (where evaluative tasks are less difficult)

HYPOTHESES

Resource Matching Theory

According to RMT information processing is most effi-cient when the resources available for processing bestmatch those required for a given task The effectiveness ofprocessing can be compromised when there are too few ortoo many resources available for the specific task (Anandand Sternthal 1987 Keller and Block 1997 Meyers-Levyand Peracchio 1995) Within the context of the current re-search RMT suggests that perceived fluency should behighest when the type of FOP cue that is provided best sup-ports or matches the specific processing demands requiredby a given processing context (comparative or noncompar-ative) Processing efficiency should decline (ie fluencyshould diminish) however when there is a suboptimalmatch between FOP cue type and the processing contextThus the ldquoprocessabilityrdquo of information largely dependson the congruence between the information format and thespecific processing task (Bettman Payne and Staelin1986 Payne Bettman and Johnson 1992)

Noncomparative Processing Resource Matchingand Fluency for a Single Product

Based on RMT and our conceptualization we expect thatthe provision of either an objective or evaluative FOP nutri-tion cue in a noncomparative processing context will en-hance the perceived fluency of a single stimulus or product(termed product-level fluency here) compared to when noFOP cues are available However the RMT perspective fur-ther suggests that an objective cue should more strongly im-pact perceived product fluency than an evaluative cue forseveral reasons The cognitive resource demands associatedwith evaluating a single product in a noncomparative con-text are considerably lower than those required to relatively

NEWMAN ET AL 751

ownloaded from

evaluate and compare many different products in a morecomplex comparative processing context As such con-sumers have sufficient cognitive resources available to dedi-cate to processing the specific detailed informationconveyed by an objective cue in a noncomparative contextThey can use their available resources to assess individuallevels of the nutrients offered in the objective cue (caloriessaturated fat sodium sugars) allowing for a more thoroughand complete evaluation of the product In contrast themore general evaluative cue information does relatively lit-tle to assist consumers with this task Evaluative cues do notfully indicate a productrsquos nutritional quality since only veryfew (if any) actual nutrient levels are provided In additionthe interpretation they offer is typically based on only a fewselect nutrients (Berning Chouinard and McCluskey 2008)Consumers are less likely to need this interpretive assistancein noncomparative contexts given that they only need toevaluate a single product in isolation (rather than many dif-ferent products in a set) These views are consistent withprior nutrition research which has shown that consumersrely more on the detailed information provided in the NFPwhen evaluating a single food product than on general eval-uative nutrition or health claims (eg ldquolow in fatrdquo or ldquohearthealthyrdquo) (Keller et al 1997 Mitra et al 1999) Thereforeobjective cues should serve as a relatively better match (ielead to higher levels of product-level fluency) thanevaluative cues for less resource-demanding noncompara-tive tasks

Comparative Processing Resource Matchingand Fluency for a Set of Products

By contrast we expect that evaluative FOP cues willlead to higher fluency for a set of products (termed set-level fluency here) than objective cues in comparativeprocessing contexts While product-level fluency refers tofluency of a single product (eg a single brand of cereal)set-level fluency considers the ease of relative comparisonsfor a set of products (ie product A is healthy product Bis less healthy than A product C is less healthy than A andB) and refers more to the fluency for the set as a wholeConsumers must compare options in comparative process-ing contexts in order to make evaluations necessary for cat-egorization and discrimination (Dhar Nowlis andSherman 1999 Nosofsky 1986 Tversky 1977) Althoughthese relative comparisons are often cognitively demand-ing they are easier when they can be made along a singlesimilar dimension (Novemsky et al 2007 Slovic andMacPhillamy 1974) This can be seen in the fact that con-sumers tend to rely more on alignable attributes than non-alignable attributes when comparing multiple options(Chakravarti and Janiszewski 2003) Therefore if all prod-ucts in a set offer the same type of cue informationbased on a common uniform dimension (eg

healthfulness) information asymmetry should be attenu-ated and consumers should be better able to make compari-sons across the set of alternatives based on the given cueinformation (Nosofsky 1986 Tversky 1977) We proposethat this shared information made salient by the cue willrender health information about the set as a whole moreconceptually fluent Thus we expect that the provision ofeither an objective or evaluative cue to have some positiveeffect on the overall fluency of a set of products (ie set-level fluency) in a comparative context relative to whenno FOP cues are available

As previously alluded to however comparative process-ing is often arduous and time consuming (Kardes et al2002) and it requires considerably more cognitive resourcesthan simpler noncomparative processing tasks Drawingagain from RMT evaluative cues should be a better matchfor comparative tasks than objective cues because they sim-plify the cognitive challenges associated with evaluatingand relatively comparing many different food options Thatis evaluative cues provide consumers with ldquocognitive short-cutsrdquo by interpreting the healthfulness of each product alonga common baseline By contrast objective cues do not offerthis interpretive assistance but instead provide specific de-tailed information about multiple nutrients that is difficultfor consumers to process repeatedly across many productoptions We therefore expect evaluative cues to serve as arelatively better match than objective cues (ie lead tohigher levels of set-level fluency) for more resource-demanding comparative tasks Relatedly we also expectthat the provision of both types of cues together will not en-hance fluency beyond that created by the single cue thatbest matches the given processing context (ie evaluative[objective] cues in comparative [noncomparative] contexts)That is when an optimal match between the type of cue andprocessing context is already present we anticipate thatthere will be little benefit from the addition of another cue(particularly when the additional cue is less appropriate forthe specific processing task) Based on RMT and our con-ceptualization we predict

H1a In a noncomparative processing context the pres-

ence of an objective cue alone will result in greater

product-level fluency than the presence of an evalu-

ative cue alone Additionally the presence of both

an objective cue and an evaluative cue will not result

in increased fluency relative to an objective cue

H1b In a comparative processing context the presence

of an evaluative cue alone will result in greater set-

level fluency than the presence of an objective

cue alone Additionally the presence of both an

evaluative cue and an objective cue will not result

in increased fluency relative to an evaluative cue

ownloaded from

Interaction of FOP Cue Type and ProcessingContext and the Mediating Role of Fluency

As shown in Figure 1 we next propose that the type ofFOP cue (objective vs evaluative) and the type of process-ing context (comparative vs noncomparative) will interactto influence consumersrsquo health-related fluency perceptionsas well as their evaluations and purchase intentions of ob-jectively more healthful products Drawing again fromRMT and our previously discussed conceptual frameworkwe anticipate the effects of FOP cues to differ across com-parative and noncomparative processing contexts Morespecifically we expect an objective cue to have a morepositive influence on perceived fluency healthfulness eval-uations and purchase intentions in noncomparative con-texts (that are less resource demanding) than incomparative contexts Conversely an evaluative cueshould have a stronger positive impact in more cognitivelychallenging comparative settings than in noncomparativesettings

Relatedly prior research also suggests favorable effectsof fluency on consumersrsquo product evaluations and purchaseintentions (Labroo and Lee 2006 Lee 2002 Lee andLabroo 2004 Novemsky et al 2007) Fluency has alsobeen shown to be positively related to product judgments(Shen Jiang and Adaval 2010) brand attitudes (Lee andAaker 2004) and product extension evaluations (Torelliand Ahluwalia 2012) Therefore as shown in Figure 1 weadditionally propose that product fluency will mediate theeffect of an objective cue on consumersrsquo evaluations andintentions to purchase healthy products in noncomparativeprocessing contexts (but not in comparative contexts) Thatis the provision of an objective cue should increase the flu-ency of a productrsquos health-related information which

should in turn positively impact healthfulness evaluationsand purchase intentions By contrast we expect that set flu-ency will mediate the effects of an evaluative cue in com-parative processing contexts (but not in noncomparativecontexts) This is consistent with the proposition that ob-jective cues serve as a better match (ie lead to higher flu-ency) for tasks in simpler noncomparative contextswhereas evaluative cues are a better match in more com-plex comparative settings Here we offer our specific hy-potheses regarding the cue type by processing contextinteractions and the conditional mediation effects

H2a The processing context moderates the effect of the

presence of a FOP cue An objective cue will lead to

higher perceptions of product-level fluency health-

fulness evaluations and purchase intentions for

healthier products in a noncomparative processing

context than in a comparative processing context

H2b The processing context moderates the effect of the

presence of a FOP cue An evaluative cue will lead

to higher perceptions of set-level fluency healthful-

ness evaluations and purchase intentions for health-

ier products in a comparative processing context

than in a noncomparative processing context

H3 The favorable effect of an objective FOP cue on (1)

healthfulness evaluations and (2) purchase intentions

of healthier products is mediated by product-level flu-

ency in a noncomparative processing context but not

in a comparative processing context

H4 The favorable effect of an evaluative FOP cue on (1)

of healthier products is mediated by set-level fluency

in a comparative processing context but not in a

noncomparative processing context

FIGURE 1

PROPOSED CONCEPTUAL MODEL

FOP Cue Type

Objective vs Evaluative

Processing Context

Comparative vs Non-Comparative

Healthfulness

Evaluations and

Purchase Intentions

Processing Fluency Type

Product-Level vs Set-Level

NEWMAN ET AL 753

ownloaded from

The remainder of this article is organized as follows Wefirst present pilot study findings to provide initial empiricalsupport for the proposed differences between cue typesbased on the RMT conceptual framework In studies 1Aand 1B we then examine the concepts of product-level andset-level fluency in noncomparative and comparative pro-cessing contexts respectively Here we conduct controlledtests to initially establish the effects of objective and evalu-ative cues on consumersrsquo perceptions of fluency (hypothe-ses 1a and 1b) We next extend these results in study 2 byexplicitly manipulating the processing context in a morerealistic retail lab setting to allow for comparisons of thecuesrsquo effects across the different types of processing con-texts (hypotheses 2A and 2B) We additionally examinethe potentially mediating role of fluency (hypotheses 3 and4) Study 3 is a concluding retail lab study that further ex-pands on the specific mechanisms underlying cue effects incomparative processing contexts

PILOT STUDY

The purpose of the online pilot study was to assess em-pirically the proposed differences between the tested objec-tive cue and the tested evaluative cues The 54 adultparticipants were recruited nationally through AmazonMechanical Turk (MTurk) (61 female ages ranged from19 to 74) All participants were shown three different cuesthat were used in our main studies an objective cue that of-fers calorie and nutrient levels a tiered evaluative cue thatranges from zero to three stars and a dichotomous evalua-tive cue that indicates whether a product is a healthful orunhealthful choice (see online appendix A) They re-sponded to the same series of questions about each of thethree cues (that were presented in random order) For ex-ample they responded to a set of questions about the firstcue presented and then they answered the same set ofquestions about the second and third cues presented

Participants reported the extent to which they perceivedeach cue as specific (very generalvery specific) detailed(not detailed at allvery detailed) and interpretive (not in-terpretive at allvery interpretive) We also measured theirperceptions of how cognitively demanding it would be touse each cue in a comparative processing context and in anoncomparative processing context specifically (ldquoUsingthe nutrition icon above to evaluate many different prod-ucts [a single product] in a comparative [noncomparative]manner would requirerdquo with end points of little efforta lotof effort and little attentiona lot of attention) (adaptedfrom Keller and Block 1997) We additionally assessedparticipantsrsquo perceptions of how well each cue would fitthe task requirements of the two processing contexts exam-ined in our studies (ldquoThe nutrition icon above is most rele-vant for evaluating many different products [a singleproduct] in a comparative [noncomparative] mannerrdquo with

end points of strongly disagreestrongly agree) (adaptedfrom Mantel and Kellaris 2003) Prior to answering anyquestions about specific cues participants were asked howcognitively demanding they perceived it would be to com-plete comparative and noncomparative tasks in general(ldquoIn general evaluating many different products [a singleproduct] in a comparative [noncomparative] manner wouldbe ardquo with end points of very easy taskvery difficult taskand very simple taskvery complex task) All constructswere measured on 7 point scales (all rrsquos ranging from 77to 94 all prsquos lt 01)

Since we measured each participantrsquos perceptions of allthree icons we performed within-subjects analyses of vari-ance with follow-up contrasts (a table with means and testsof differences can be found in online appendix A) As ex-pected planned contrasts indicate that the objective cuewas perceived as more detailed (Mfrac14 661) than both theHealthy Stars evaluative cue (Mfrac14 326 F(1 53)frac14 20310plt 001) and the dichotomous evaluative cue (Mfrac14 206F(1 53)frac14 33494 plt 001) The objective cue was addi-tionally perceived as more specific (Mfrac14 667) than theHealthy Stars evaluative cue (Mfrac14 354 F(1 53)frac14 17959plt 001) and dichotomous cue (Mfrac14 226 F(153)frac14 21799 plt 001) Also as expected the objectivecue was perceived as less interpretive (Mfrac14 328) than theHealthy Stars cue (Mfrac14 437 F(1 53)frac14 992 plt 01) anddichotomous cue (Mfrac14 428 F(1 53)frac14 354 plt 04)

In addition contrasts confirmed that comparative taskswere perceived as more challenging (Mfrac14 446) thannoncomparative tasks in general (Mfrac14 234 tfrac14 603plt 001) Given these higher resource requirements incomparative contexts our conceptualization suggested thatevaluative cues would be perceived as a better match forcomparative tasks than objective cues Results confirmedthat the evaluative Healthy Stars cue (Mfrac14 507 F(153)frac14 979 plt 01) and evaluative dichotomous cue(Mfrac14 469 F(1 53)frac14 313 plt 04) were both viewed asbetter matches than the objective cue (Mfrac14 383) for tasksin comparative contexts Providing additional support forour rationale contrasts further revealed that using eitherthe Healthy Stars cue (Mfrac14 298 F(1 53)frac14 2154plt 001) or dichotomous cue (Mfrac14 245 F(1 53)frac14 2692plt 001) to evaluate multiple products in comparative con-texts was perceived to require less resources than using theobjective cue (Mfrac14 434) By contrast our conceptualiza-tion also suggested that objective cues would be perceivedas a better match than evaluative cues for noncomparativetasks because consumers should have sufficient cognitivecapacity to meet the relatively low resource requirementsassociated with noncomparative tasks Results confirmedthat the objective cue was viewed as a relatively bettermatch for evaluating a single product in a noncomparativecontext (Mfrac14 498) than both the Healthy Stars cue(Mfrac14 402 F(1 53)frac14 620 plt 02) and dichotomous cue(Mfrac14 419 F(1 53)frac14 282 plt 05)

ownloaded from

Collectively these findings provide empirical supportfor the proposed differences between objective and evalua-tive FOP cues and for the RMT conceptual frameworkThey also provide initial general support for our proposi-tion that perceived fluency should be highest when thetype of FOP cue information provided best supports ormatches the processing demands elicited by the specificprocessing context We now directly examine the effects ofdifferent FOP cues on fluency in noncomparative and com-parative processing contexts in studies 1A and 1Brespectively

STUDY 1A

Methods

Study 1A was a 2 (objective FOP cue present vs ab-sent) 2 (evaluative FOP cue present vs absent) be-tween-subjects online experiment The 207 participantswere recruited nationally through MTurk and randomly as-signed to one of the four experimental conditions The me-dian household income was $40000 to $49000approximately 63 were female and ages ranged from 18to 81 Manipulations were placed on the front of a hypo-thetical frozen pizza package (see online appendix B)Pizza was chosen to be consistent with prior health market-ing research that used a nutritionally mixed (moderate)product (Andrews et al 2011) The nutrient values avail-able in the NFP mirrored those of a similar pizza on themarket When the objective cue was present it was consis-tent with the nutrient information disclosed in the productrsquosNFP The evaluative cue used in this study (and in study1B) was dichotomous such that a product only qualified forthe icon if specific nutritional standards related to saturatedfat trans fat sodium and sugar levels were met (consistentwith the standards needed to qualify for Walmartrsquos Greatfor You icon or to be eligible for the IOMrsquos Healthy Stars)The nutrition profile of the product (available in the NFPto all participants) remained constant across all conditions

Dependent Measures

Manipulation checks were used to assess the awarenessof the FOP cues After responding to all dependent vari-ables participants were asked about each of the cue types(eg Did you see a nutritional icon on the front of thepackaged food item that was presented) with a picture ofthe specific icons included Both questions had ldquonordquo orldquoyesrdquo response categories

The primary dependent variable of interest in this initialstudy was product-level fluency Specifically we focusedon how easily participants were able to discern the health-fulness of the product given the presence or absence of ob-jective and evaluative FOP nutrition cues Product fluencywas assessed through four 7 point bipolar adjective scales(modified from Fang Singh and Ahluwalia 2007 Lee and

Aaker 2004) with end points of ldquostrongly disagreestronglyagreerdquo (ie ldquoGiven the information on the package it iseasy to determine how healthy the product isrdquo ldquoGiven theinformation on the package it is clear whether the productis high or low in its level of nutritiousnessrdquo ldquoI feel confi-dent about whether this product is a healthy or unhealthychoice based on the information on the packagerdquo and ldquoIt iseasy to understand whether this product is a healthy or un-healthy choice given the information shown on the pack-agerdquo) (afrac14 94) Higher values indicate higher levels ofperceived fluency

RESULTS

Manipulation Check

The checks revealed a high level of awareness of the cuemanipulations When the evaluative cue was present (ab-sent) 90 (81) reported seeing (not seeing) it(v2frac14 10555 plt 001) When the objective cue was pre-sent (absent) 88 (92) reported seeing (not seeing) it(v2frac14 13483 plt 001)

Effects of Objective and Evaluative FOPNutrition Cues on Product Fluency

The FOP objective cue X evaluative cue interaction wassignificant for perceived product fluency (F(1 203)frac14 593plt 02 g2

pfrac14 028) The plot of means can be found inFigure 2 Perceived product fluency was lowest in the con-trol condition (Mfrac14 319) and increased significantly wheneither the objective cue (Mfrac14 508 F(1 203)frac14 4365plt 001 Cohenrsquos dfrac14 125) or the evaluative cue(Mfrac14 380 F(1 203)frac14 448 plt 04 Cohenrsquos dfrac14 37) wasprovided in isolation However as expected the objectivecue (alone) led to higher perceived fluency than the

FIGURE 2

STUDY 1A EFFECTS OF OBJECTIVE AND EVALUATIVE FOPCUES ON PRODUCT-LEVEL CONCEPTUAL FLUENCY

Objective Cue Absent Present

Level F

luency

Evaluative Cue Absent Evaluative Cue Present

NEWMAN ET AL 755

ownloaded from

evaluative cue (alone) (508 vs 380 F(1 203)frac14 2013plt 001 Cohenrsquos dfrac14 85) Additionally both cues to-gether on the package did not increase fluency more thanthe objective cue alone did (468 vs 508 F(1203)frac14 179 pgt 15) That is the evaluative cue had no ad-ditional effect on product fluency when added to a packagethat already offered an objective cue This pattern of resultssupports hypothesis 1a and offers initial insight on ourproduct-level fluency concept

DISCUSSION

The purpose of study 1A was to examine the effects ofobjective and evaluative FOP cues on perceived productfluency The results provide preliminary support for ourproposition that objective cues enhance product fluencymore in noncomparative processing contexts than evalua-tive cues As previously noted extensive prior consumerresearch has focused on how calorie and nutrient informa-tion can affect health-related perceptions of a single prod-uct (Hieke and Taylor 2012) While consumers canevaluate a single product in a noncomparative mannerthey often also simultaneously evaluate and compare mul-tiple products within a set during a typical shopping experi-ence (Nedungadi 1990) Consequently many prior studieshave not sufficiently accounted for the more complex com-parative processing settings frequently encountered by con-sumers at the point of purchase Therefore in the nextstudy we consider the effects of objective and evaluativecues on consumersrsquo perceptions of set-level fluency in acomparative processing context (when evaluating multipleoptions) As suggested in hypothesis 1b we expect a pat-tern of results nearly opposite to those observed in study1A such that evaluative cues should have a greater influ-ence on set fluency than objective cues

STUDY 1B

Methods

Study 1B was a 2 (objective FOP cue present vs ab-sent) 2 (evaluative FOP cue present vs absent) be-tween-subjects experiment conducted online The 190participants were again recruited nationally through MTurkand randomly assigned to one of the four experimentalconditions The median household income was $40000 to$49000 approximately 62 were female and ages rangedfrom 18 to 76

Manipulations were again placed on the front of hypo-thetical frozen pizza packages (see online appendix C)Participants were presented with a pizza product that had anutritional profile identical to the pizza used in study 1A(and thus again qualified for the same dichotomous evalua-tive cue used in study 1A) However this product was pre-sented in combination with two other pizza products (that

did not qualify for the evaluative cue) to create a set ofthree frozen pizzas The nutrient values available in theNFPrsquos closely mirrored those of similar products on themarket and they were again made available to all partici-pants When the objective cue was present it was againconsistent with the nutrient information disclosed in theNFP for each product The nutrition profiles of the prod-ucts remained constant across all conditions

Dependent Measures

The dependent variable of interest in this study was set-level fluency That is instead of focusing on participantsrsquoevaluations of one specific product (as in study 1A) thisstudy measured health-related conceptual fluency across aset of products After initial pilot testing four set fluencymeasures with end points of ldquostrongly disagreestronglyagreerdquo were used (ie ldquoOverall given the information pro-vided on the packages in the set of three pizza products itis easy to determine which ones are the more healthy op-tionsrdquo ldquoBased on the information on the packages in theset of three pizza products I know which brands are thehealthy onesrdquo ldquoThe information presented on the packagesin the set of three pizza products makes it easy for me tochoose a healthy optionrdquo and ldquoFor the set of three pizzaproducts available I can easily tell which ones are morehealthy and which ones are less healthyrdquo) (afrac14 98) Highervalues indicate higher levels of perceived fluency Thesame manipulation check from study 1A was used again

RESULTS

Manipulation Check

The check again revealed a high level of awareness ofthe cue manipulations When the evaluative cue was pre-sent (absent) 91 (98) reported seeing (not seeing) it(v2frac14 14742 plt 001) When the objective cue was pre-sent (absent) 99 (95) reported seeing (not seeing) it(v2frac14 16507 plt 001)

Effects of Objective and Evaluative FOPNutrition Cues on Set Fluency

As expected the FOP objective cue X evaluative cue in-teraction was significant for perceived set fluency (F(1186)frac14 1929 plt 001 g2

pfrac14 094) The plot of means isshown in Figure 3 Perceived fluency was again lowest inthe control condition (Mfrac14 211) and increased signifi-cantly when either the evaluative cue (Mfrac14 506 F(1186)frac14 6982 plt 001 Cohenrsquos dfrac14 160) or objective cue(Mfrac14 437 F(1 186)frac14 3973 plt 001 Cohenrsquos dfrac14 125)was provided in isolation However as expected the evalu-ative cue (alone) led to higher perceived fluency than theobjective cue (alone) (506 vs 437 F(1 186)frac14 408plt 05 Cohenrsquos dfrac14 40) Additionally both cues together

ownloaded from

did not increase fluency more than the evaluative cue alonedid (508 vs 506 F(1 186)frac14 004 pgt 90) That is theobjective cue had no additional effect on set fluency whenthe evaluative cue was already available This pattern of re-sults provides support for hypothesis 1b

DISCUSSION

The findings from studies 1A and 1B offer initial insighton our product-level and set-level fluency concepts respec-tively Results support our proposition that objective cueshave a more positive impact on product fluency than evalua-tive cues in noncomparative processing contexts whereasevaluative cues have a more positive impact on set fluencythan objective cues in comparative contexts This contentionwas strengthened by the fact that the combination of bothcues did not increase fluency in either processing contextbeyond that created by the single cue that was expected toenhance fluency the most That is fluency was equivalent inthe noncomparative (comparative) processing context re-gardless of whether both cues were present or only a singleobjective (evaluative) cue was present

Next study 2 extends studies 1A and 1B in several im-portant ways As Hsee et al (2013) recently noted judg-ments and choices can vary greatly across comparative andnoncomparative contexts and a ldquofailure to recognize the dif-ference can lead to systematic and serious errorsrdquo (182)Also the online setting common to many nutrition labelingstudies may at times restrict the generalizability of reportedfindings (Hieke and Taylor 2012) To address these pointscollectively we explicitly manipulate the processing contextin a realistic retail store laboratory environment in study 2This allows us directly to compare each cuersquos effects on

fluency and other important outcomes related to consumerhealth (ie healthfulness evaluations and purchase inten-tions for healthier products) across the comparative andnoncomparative contexts (as outlined in hypothesis 2)

An additional objective of study 2 is to assess whetherfluency serves as an underlying mechanism for the effects ofthe different cue types on evaluations and purchase inten-tions for healthier products More specifically we aim todemonstrate that product fluency mediates the effects of ob-jective cues in noncomparative processing contexts whereasset fluency mediates the effects of evaluative cues in com-parative processing contexts (as indicated in hypotheses 3and 4) We further enhance the generalizability of our con-ceptualization by examining these effects in a different prod-uct category (single-serve meals) with a different evaluativeicon (the Healthy Stars icon recommended by the IOM)

STUDY 2

Methods

Study 2 utilized a 2 (processing context comparative vsnoncomparative) 3 (FOP cue objective vs evaluative vscontrol) between-subjects experiment Participants wererecruited from a research subject pool consisting of bothstudents and adults at a large public university resulting inmixed sample of 126 adults and students Approximately62 of this sample was male and ages ranged from 19 to43 (mean agefrac14 22) Research was conducted in theShopper Experimental Lab Facility (ShELF) a behavioralresearch lab designed to look like a retail store with a widerange of products (eg food cleaning supplies DVDs)and numerous arrangements (eg end caps aisles islands)as shown in online appendix D Each participant was ran-domly assigned to one of the six conditions

We closely followed the procedures outlined by vanHoren and Pieters (2012) to manipulate the processing con-text We first presented participants in the comparativeprocessing condition with a set of five meal products on aretail shelf (similar to study 1B) We asked them to exam-ine all of the products simultaneously in a comparativemanner and to base their healthfulness evaluations and pur-chase intentions for the single objectively healthiest mealproduct on how it compared with the other four productsThey also recorded their perceptions of set fluency whilerelatively evaluating the products in a comparative mannerIn order to further facilitate comparative processing theseparticipants answered all of the dependent measures withall five of the meal products always visible to them

Conversely we showed the healthy product of interestalone on a retail shelf to participants in the noncomparativecondition (similar to study 1A) We asked them to form ageneral impression of the product and to base their health-fulness evaluations and purchase intentions for the producton that impression (van Horen and Pieters 2012) They

FIGURE 3

STUDY 1B EFFECTS OF OBJECTIVE AND EVALUATIVE FOPCUES ON SET-LEVEL CONCEPTUAL FLUENCY

Evaluative Cue Absent Present

Objective Cue Absent Objective Cue Present

NEWMAN ET AL 757

ownloaded from

also recorded their fluency perceptions of this single prod-uct while evaluating it independently in a noncomparativemanner Participants were not able to see any other mealproducts when completing the dependent measures to pre-vent comparative evaluations After all of these dependentmeasures had been completed we took every participant toa separate area of the ShELF and provided them with thestimuli from the alternative processing condition (eg par-ticipants in the noncomparative condition saw all five mealproducts together in a single set on the shelf) After exam-ining the stimuli while in their manipulated processingmodes (van Horen and Pieters 2012) participants in thenoncomparative condition completed the set fluency mea-sure and participants in the comparative condition com-pleted the product fluency measure This enabled us tocompare the cuesrsquo effects on fluency across the differentprocessing contexts See online appendix E for an exampleof the stimuli used in each processing context

To manipulate the FOP cue we presented each productwith either the objective cue from studies 1A and 1B anew evaluative cue (the Healthy Stars cue proposed by theIOM) or no FOP nutrition information (control condition)We placed all cue manipulations on shelf tags in front ofthe meal products (FDA 2013) When the objective cuewas present it was consistent with the nutrient informationdisclosed in the NFP for each product The evaluativeldquoHealthy Starsrdquo cue offers calorie information for a prod-uct and assigns it with zero to three stars based on predeter-mined nutrient standards (more stars indicate healthierproducts) We used the IOMrsquos exact nutrient standards toassign a star level to each of the five meal products oneproduct was the objectively healthiest and qualified forthree stars three products were moderately healthy andqualified for one or two stars and another product was un-healthy and did not qualify for any stars Consistent withthe previous studies the nutrient values available in theNFPrsquos closely mirrored those of similar products on themarket and were available to all participants The nutri-tional profiles of the products again remained constantacross all conditions and any potentially confoundingpackage indicators of product healthfulness (eg ldquolow infatrdquo) were discreetly removed We counterbalanced thepresentation of products on the shelf to control for any po-sitioning confounds (eg prominence due to eye level orrightleft placement)

Dependent Measures

We assessed the perceived fluency of the objectivelyhealthier meal product (ie product-level fluency) with theitems ldquoGiven the information on the package it is easy todetermine how healthy this product isrdquo and ldquoInformationabout this product is easy to processrdquo with end points ofldquostrongly disagreestrongly agreerdquo (rfrac14 90 plt 01) Weslightly adjusted the previously used set-level fluency

measure to reflect the new product type and it again ex-hibited strong reliability (afrac14 96) We then performed testsof convergent and discriminant validity for the product andset fluency constructs The two factor confirmatory factorresults showed standardized krsquos ranging from 89 to 94 forproduct fluency and 81 to 94 for set fluency The v2 forthe two factor model was nonsignificant (v2frac14 121 dffrac14 8pgt 10) and significantly less than that of the one factormodel (v2

difffrac14 1161 dffrac14 1 plt 001) Results also showthe u2 (325) for the two measures was less than the aver-age variance extracted estimate of 83 These findings sup-port both discriminant and convergent validity

We assessed participantsrsquo healthfulness evaluations ofthe objectively healthy meal product with two 7 point bipo-lar adjective scales using end points of ldquonot at all nutri-tious highly nutritiousrdquo and ldquovery unhealthyvery healthyrdquoin response to the question ldquoPlease consider the nutritionlevel of the food product shown Do you believe that thefood product isrdquo (rfrac14 93 plt 01) We measured purchaseintentions for the healthy meal with two 7 point bipolar ad-jective scales with end points of ldquovery unlikelyvery likelyrdquoand ldquonot probablevery probablerdquo (rfrac14 98 plt 01) Lastlywe assessed the effectiveness of the processing manipula-tion on a 7 point scale with the items ldquoI based my productevaluations on how wel7l it compared to the other avail-able optionsrdquo and ldquoI based my product evaluations on myoverall impression of itrdquo with end points of ldquostrongly dis-agreestrongly agreerdquo (van Horen and Pieters 2012) Weused the same cue manipulation check from previous stud-ies and coded the dependent measures so that higher valuesindicate more favorable responses

RESULTS

Manipulation Check

Analysis of variance results indicated that participants inthe comparative condition based their evaluations on rela-tive comparisons to other options more so than those in thenoncomparative condition (560 vs 259 plt 001) whilethose in the noncomparative condition based their evalua-tions on their impression of the single product more thanparticipants in the comparative condition (597 vs 234plt 001) The checks also again revealed a high level ofawareness of the cue manipulation when the evaluativecue was present (absent) 96 (90) reported seeing (notseeing) it (v2frac14 9307 plt 001) and when the objectivecue was present (absent) 98 (91) reported seeing (notseeing) it (v2frac14 8968 plt 001)

Interactive Effects of FOP Cue Types andProcessing Contexts

The overall FOP cue X processing context interactionwas significant for both perceived product fluency

ownloaded from

(F(2 120)frac14 420 plt 02 g2pfrac14 065) and perceived set

fluency (F(2 120)frac14 360 plt 04 g2pfrac14 057) The plots of

means are shown in Figures 4 and 5 respectively ForFigure 4 our primary focus lies on the effects of the objec-tive cue on product fluency across the two processing con-texts (see hypothesis 2a) As expected the objective cueled to higher fluency in the noncomparative processingcontext than in the comparative context (600 vs 491 F(1120)frac14 532 plt 03 dfrac14 72) Also compared to the nocue control condition in the noncomparative context(Mfrac14 443) product fluency increased when either the ob-jective cue (Mfrac14 600 F(1 120)frac14 817 plt 01 dfrac14 132)or the evaluative cue (Mfrac14 521 F(1 120)frac14 325 plt 05dfrac14 60) was provided Also as expected the objective cueled to higher product fluency than the evaluative cue (F(1120)frac14 300 plt 05 dfrac14 73) These results are consistentwith the findings presented in study 1a

In Figure 5 our primary focus is on the effects of theevaluative cue on set fluency across the two processingconditions (see hypothesis 2b) As expected the evaluativecue led to higher fluency in the comparative processingcontext than in the noncomparative context (616 vs 481F(1 120)frac14 1235 plt 001 dfrac14 109) Additionally com-pared to the control condition in the comparative context(Mfrac14 387) set fluency increased when either the evalua-tive cue (Mfrac14 616 F(1 120)frac14 3241 plt 001 dfrac14 237)or objective cue (Mfrac14 469 F(1 120)frac14 284 plt 05dfrac14 62) was provided The evaluative cue also led tohigher set fluency than the objective cue (F(1120)frac14 1294 plt 001 dfrac14 129) as expected These re-sults are consistent with the study 1b findings

The overall FOP cue X processing context interactionwas also significant for healthfulness evaluations (F(2120)frac14 510 plt 01 g2

pfrac14 087) and purchase intentions

(F(2 120)frac14 843 plt 001 g2pfrac14 123) for the healthy

meal product We expected that the pattern of the cuesrsquo ef-fects on these outcomes across processing contexts wouldbe similar to that just described for fluency The plots ofmeans for both dependent measures are shown in Figure 6As anticipated results reveal that the objective cue led tohigher healthfulness evaluations (437 vs 364F(1 120)frac14 317 plt 04 dfrac14 56) and purchase intentions(543 vs 416 F(1 120)frac14 622 plt 01 dfrac14 86) in thenoncomparative processing context than in the comparativecontext Conversely the evaluative cue led to higherhealthfulness evaluations (543 vs 462 F(1 120)frac14 604plt 01 dfrac14 60) and purchase intentions (582 vs 443F(1 120)frac14 1150 plt 001 dfrac14 88) in the comparativecontext than in the noncomparative context Taken to-gether these findings support hypotheses 2a and 2b

Another objective of study 2 was to examine the indirecteffects of the cues on healthfulness evaluations and pur-chase intentions for the healthy meal (via fluency) acrossthe different processing contexts (see hypotheses 3 and 4)We performed a series of bootstrap analyses using 1000samples and 95 bias-corrected confidence intervals (CIs)in PROCESS Model 8 (Hayes 2013) Findings relevant tothe mediating roles of product fluency and set fluency areshown in Tables 1 and 2 respectively for the dependentvariables of healthfulness evaluations and purchase inten-tions (see models 2 and 3 in each table)

Results indicate that the indirect effect (IE) associatedwith the cue type X processing context interaction throughproduct fluency was significant for both healthfulness eval-uations (IEfrac1448 CI [1102 110]) and purchase in-tentions (IEfrac1433 CI [1009 003]) (ie neither CIcontained zero see Hayes 2013 Zhao Lynch and Chen2010) Similarly the IE associated with the same

FIGURE 4

STUDY 2 EFFECTS OF FOP CUES AND PROCESSINGCONTEXT ON PRODUCT-LEVEL CONCEPTUAL FLUENCY

Processing Context Noncomparative Comparative

Level F

luency

Objective Cue Control Evaluative Cue

FIGURE 5

STUDY 2 EFFECTS OF FOP CUES AND PROCESSINGCONTEXT ON SET-LEVEL CONCEPTUAL FLUENCY

Level F

luency

NEWMAN ET AL 759

ownloaded from

interaction through set fluency was significant for health-fulness evaluations (IEfrac14 16 CI [012 507]) and purchaseintentions (IEfrac14 15 CI [001 511]) These findings for-mally indicate that the mediating role of fluency is moder-ated by the processing context (Hayes 2013)

More specifically findings show that compared to theevaluative cue the IE of the objective cue through productfluency was significant in the noncomparative processingcontext for healthfulness evaluations (IEfrac14 23 CI [052552]) and purchase intentions (IEfrac14 15 CI [006 496])By contrast the IE of the objective cue through the samemediational path in the comparative processing contextwas not significant for either dependent measure (ie bothCIs contained zero) These results provide support for hy-potheses 3a and 3b Next compared to the objective cuethe IE of the evaluative cue through set fluency was signifi-cant in the comparative processing context for healthful-ness evaluations (IEfrac14 17 CI [027 402]) and purchaseintentions (IEfrac14 17 CI [001 378]) By contrast the IE ofthe evaluative cue through the same mediational path inthe noncomparative processing context was not significantfor either dependent measure These findings support hy-potheses 4a and 4b

DISCUSSION

Study 2 addressed some important limitations of manyprevious nutrition labeling studies by testing the interactiveeffects of FOP cues and processing contexts in a more real-istic retail laboratory setting By explicitly manipulatingthe processing context we were able to demonstrate thatobjective cues have a more positive impact on fluencyhealthfulness evaluations and purchase intentions forhealthier products in noncomparative processing contextsthan in comparative contexts Conversely we also showedthat evaluative cues lead to higher fluency healthfulnessevaluations and purchase intentions for healthy productsin comparative processing contexts than in noncomparative

FIGURE 6

STUDY 2 EFFECTS OF FOP CUES AND PROCESSINGCONTEXT ON HEALTHFULNESS EVALUATIONS ANDPURCHASE INTENTIONS OF A HEALTHY PRODUCT

Panel A Healthfulness Evaluations

Panel B Purchase Intentions

Healthfu

ations

TABLE 1

STUDY 2 MEDIATING ROLE OF PRODUCT-LEVEL FLUENCY FOR THE EFFECTS OF AN OBJECTIVE CUE RELATIVETO AN EVALUATIVE CUE

Model 1 Product-levelfluency (mediator)

Model 2 Healthfulnessevaluations of healthier

product (with added mediator)

Model 3 Purchase intentionsof healthier product

(with added mediator)

Independent variables Coefficient T value Coefficient T value Coefficient T value

Objective cue 79 173 38 93 85 170Processing context 61 158 63 184 127 302Cue context 170 275 115 203 233 338Product-level fluency ndash ndash 29 306 19 171

NOTEmdashAll coefficients are unstandardized and cue type was coded as 0frac14evaluative cue 1frac14 objective cue The indirect effect of the interaction through prod-

uct fluency was significant for both healthfulness evaluations (IEfrac1448 CI [1102 110]) and purchase intentions (IEfrac1433 CI [1009 003]) The indirect

effects of the objective cue through fluency were positive and significant in the noncomparative processing condition (but not in the comparative condition) provid-

ing support for hypotheses 3a and 3b plt 05 plt 01 plt 001

ownloaded from

contexts Study 2 additionally demonstrated that the effectsof an objective cue (but not an evaluative cue) are medi-ated by product fluency in noncomparative settings where-as the effects of an evaluative cue (but not an objectivecue) are mediated by set fluency in comparative settings

Our final study (study 3) serves to expand on the insightsgained in study 2 regarding how package cues influenceimportant consumer health-related outcomes Consumersoften engage in comparative processing in everyday retailchoice environments where they are confronted with manydifferent options at once (such as supermarkets)(Nedungadi 1990) Therefore this study more strongly fo-cuses on the direct effects and IEs of an evaluative cue onconsumersrsquo actual choices in a comparative processingcontext We focus exclusively on an evaluative cue instudy 3 since evaluative cues appear to have the strongestinfluence on fluency healthfulness evaluations and pur-chase intentions for healthy products in comparative set-tings We aim to show that an evaluative cue increasesperceived set fluency which should lead to greater per-ceived differentiation between more and less healthful op-tions in a set of products This increase in relativehealthfulness perceptions should in turn lead to healthierchoices from the available set We test this final assertionin study 3

STUDY 3

Methods

Study 3 utilized a 2 (evaluative FOP cue present vs ab-sent) 2 (product healthfulness more healthful vs lesshealthful) mixed experimental design The cue manipula-tion was again a between-subject factor product healthful-ness was a within-subjects factor Each participant wasrandomly assigned to one of the two cue conditionsParticipants were recruited from a research subject poolconsisting of both students and adults at a large public

university resulting in a mixed sample of 100 adults andstudents Approximately 56 of this sample was femaleand ages ranged from 18 to 44 (mean agefrac14 23)

We again collected data in the ShELF retail lab and usedthe same evaluative cue as in study 2 We placed the exper-imental cue manipulation on the front of nine microwav-able soup products (see online appendix F) According tothe IOMrsquos Healthy Stars nutrient standards three productswere deemed healthy and qualified for three stars threeproducts were moderately healthy and qualified for one ortwo stars and three products were unhealthy and did notqualify for any stars We again counterbalanced the prod-ucts in the set and discreetly removed any potentially con-founding healthfulness indicators on the front of thepackages (eg ldquolow in fatrdquo) All participants were pre-sented with the set of soup products at the retail shelf andwere allowed to handle and compare the products freelythroughout the questionnaire

Dependent Measures

Choice was the primary dependent measure of interestand assessed by the question ldquoWhich soup product wouldyou be most likely to purchaserdquo (coded as 1 if a three starproduct was selected and as 0 otherwise) The same set flu-ency measure from studies 2 and 3 was used and again ex-hibited satisfactory reliability (afrac14 94) The samehealthfulness perception measure from studies 2 and 3 wasalso used again to assess evaluations of a healthy three starsoup and an unhealthy zero star soup (rfrac14 73 and rfrac14 87respectively) Consistent with prior nutrition labeling re-search (Feunekes et al 2008 Viswanathan Hastak andGau 2009) we computed a healthfulness difference vari-able by subtracting the perceived healthfulness of the zerostar soup from the perceived healthfulness of the three starsoup This provided a relative measure of the perceiveddifferences in healthfulness of the healthier and unhealthierproducts in the set

TABLE 2

STUDY 2 MEDIATING ROLE OF SET-LEVEL FLUENCY FOR THE EFFECTS OF AN EVALUATIVE CUERELATIVE TO AN OBJECTIVE CUE

Model 1 Set-levelFluency (mediator)

Model 3 Purchaseintentions of healthier

Evaluative cue 06 28 06 30 52 211Processing context 01 02 83 193 128 248Cue x context 67 220 66 233 118 349Set-level fluency ndash ndash 23 242 22 192

NOTEmdashAll coefficients are unstandardized and cue type was coded as 0frac14 objective cue 1frac14 evaluative cue The indirect effect of the interaction through set

fluency is significant for both healthfulness evaluations (IEfrac14 16 CI [012 507]) and purchase intentions (IEfrac14 15 CI [001 511]) The indirect effects of the eval-

uative cue through fluency were positive and significant in the comparative processing condition (but not in the noncomparative condition) providing support for hy-

potheses 4a and 4b plt 05 plt 01 plt 001

NEWMAN ET AL 761

ownloaded from

RESULTS

Logistic regression results indicate a higher likelihoodof choosing a healthier product from the set when the eval-uative cue was on the packages (bfrac14 113 standard er-rorfrac14 43 plt 01) More specifically results revealed thatthe probability of choosing a healthier product from the setsignificantly increased from 48 when the evaluative cuewas absent to 74 when it was present (zfrac14 233 plt 02)In order to expand on the processes underlying this effectwe performed a series of bootstrap analyses using 1000samples and 95 bias-corrected CIs in PROCESS Model 6(Hayes 2013) to formally test the ldquocueset fluencyper-ceived differentiation between more and less healthfulproducts in the setchoicerdquo serial mediational path As ex-pected the indirect effect of the evaluative cue throughthis mediational path was significant (IEfrac14 05 CI [001244]) These concluding findings build on our prior studiesby explaining in more depth how evaluative cues assistconsumers with making heathier choices in complex com-parative processing environments such as grocery storesThe implications of this research for theory and consumerhealth are discussed in the final section

GENERAL DISCUSSION

The typical consumer makes an average of 200 to 300decisions related to food choice and consumption in agiven day (Wansink and Sobal 2007) However consumersrarely have the cognitive capacity to process all availableinformation and they often use only a limited number ofproduct attributes in their evaluations to avoid informationand choice overload (Bettman Luce and Payne 1998)Consumers therefore frequently rely on less complex com-munications such as FOP nutrition cues (Roberto et al2012) It is still not clear though if and how FOP nutritionlabeling can satisfy its stated intended purpose of helpingconsumers make healthier food choices (FDA 2015Federal Register 2010) Moreover prior research has givenlittle specific consideration to how the effects of varioustypes of FOP nutrition cues may differ across noncompara-tive and comparative processing contexts (in which con-sumers process health information for a single product inisolation vs a set of multiple products respectively) Thisrelationship between FOP cue type and processing contexttype is growing in importance as consumers continue toface an increasing variety of food choice tasks and environ-ments (eg evaluating a specific product online vs evalu-ating many different products at once on a supermarketshelf)

Drawing from Resource Matching Theory (RMT) weexamined the effects of different FOP nutrition cues (ob-jective vs evaluative) on consumersrsquo evaluations and in-tentions to purchase healthy products across differentprocessing contexts (comparative vs noncomparative) We

then integrated a processing fluency theoretical frameworkto help explain why these important interactive effects oc-cur Study 1A demonstrated that objective cues which of-fer specific quantitative nutrition information increase(product-level) fluency more than evaluative cues in sim-pler noncomparative processing contexts Converselystudy 1B showed that evaluative cues which provide inter-pretive information about a productrsquos overall healthfulnessandor nutrients enhance (set-level) fluency more than ob-jective cues in more complex comparative processing con-texts After establishing these initial cue effects we thenexplicitly manipulated the processing context in study 2Here we demonstrated that objective cues positively influ-ence fluency healthfulness evaluations and purchase in-tentions for healthy products more strongly innoncomparative contexts than in comparative contexts Incontrast we showed that evaluative cues lead to higher flu-ency evaluations and intentions in comparative settingsthan in noncomparative settings Study 2 results also dem-onstrated that product fluency and set fluency are impor-tant mechanisms that underlie the observed effects ofobjective and evaluative cues in noncomparative and com-parative processing contexts respectively

Lastly in study 3 we expanded more on the processesunderlying the effects of evaluative cues on consumersrsquo ac-tual food choices in comparative settings (such as grocerystores) Findings revealed that evaluative cues increase theperceived fluency of a set of products which in turn leadsto greater perceived differentiation between the more andless healthful products in the set As a result consumersare more likely to choose healthier items from the availableoptions Both studies 2 and 3 were conducted in a more re-alistic retail lab setting with different product categories toovercome potential restrictions to generalizability posed bythe online settings used in many previous experimental la-beling studies (Hieke and Taylor 2012)

Theoretical Contributions

The present research offers new conceptual insight intothe relationships between cue type processing contextsand fluency Drawing from RMT (Anand and Sternthal1987 Keller and Block 1997 Meyers-Levy and Peracchio1995) we highlighted the importance of distinguishing be-tween comparative and noncomparative processing con-texts when considering effects of different types of FOPcues on fluency and other critical health-related outcomesFindings supported our predictions regarding the moderat-ing role of the processing context objective cues enhancedfluency more in noncomparative contexts than in compara-tive contexts while evaluative cues increased fluency morein comparative settings than in noncomparative settingsWe additionally showed similar effects of cue type acrossprocessing contexts on consumersrsquo evaluations and pur-chase intentions for objectively healthy products

ownloaded from

Drawing from prior research on the role of fluency incue effects (Berger and Fitzsimons 2008 Lee and Labroo2004 Whittlesea 1993) we also introduced the concepts ofproduct-level and set-level fluency and identified them asimportant mechanisms that underlie the positive impact ofFOP cues on consumersrsquo food evaluations intentions andchoices We then extended these findings by showing thatthe IEs of the cues through fluency are conditional on thenature of the processing context More specifically wedemonstrated that the positive IEs of objective (evaluative)cues only occur in noncomparative (comparative) settingsWe therefore highlighted the processing context type as aninfluential moderator of the mediating roles of product andset fluency (as proposed in Figure 1) These findings com-bined with the other results of this research collectivelysuggest that set fluency appears to be both conceptuallyand empirically distinct from other measures that focusonly on the fluency of a single stimulus (ie product flu-ency) Future studies may consider the potential usefulnessof this construct and whether it can be adapted for use inother research contexts

Next our findings extended previous research that hasshowed the salience and importance of a cue increases(eg price origin) when multiple products are organizedbased on it (Hutchinson and Alba 1991) For exampleAreni Duhan and Kiecker (1999) demonstrated that whenwine bottles were physically reorganized and shelved ac-cording to region the sales of wines from preferred regionsincreased while wine sales from less favorable regions de-creased (relative to when they were displayed by variety)We similarly showed that the provision of a standardizedcue (ie a FOP evaluative cue) can influence consumersrsquoevaluations of a set of products with varying nutritionalprofiles and ultimately their intentions and choices It canbe argued that the consistent comparable information pre-sented by the cue encourages consumers to focus more onthis health-related information and thus is more likely toinfluence their evaluative and choice processes Howeverunlike the reshelving of wines just noted our findings arestrengthened by the fact that they were observed withoutphysically rearranging and grouping the products accord-ing to the cue (eg placing relatively healthier products to-gether in a ldquohealthy sectionrdquo) These effects also largelyheld across multiple types of food categories and experi-mental settings This highlights the importance of cue ex-posure and the IEs that they have through conceptualfluency on consumersrsquo evaluations intentions andchoicesmdashparticularly at the set level

Lastly note that the consistent IEs of cues through con-ceptual fluency observed here suggest support for a cogni-tive-based role of fluency rather than an affective one(Labroo Dhar and Schwarz 2008 Winkielman et al2003) Our results demonstrate that the positive impact offluency is not only a perceptual process but also part of acognitive linkage in which consumers critically distinguish

healthier products from less healthy ones based on the spe-cific information conveyed by the cues (MogilnerRudnick and Iyengar 2008) For example study 3 demon-strated that evaluative cue information indirectly helpedparticipants better (cognitively) differentiate between moreand less healthful products through enhanced set fluencyThese differences in healthfulness perceptions then in-creased the likelihood that participants made healthychoices These findings reveal that fluency affects not onlyevaluations of a single object but also evaluations of otherconceptually related objects (ie products in a set) Morebroadly they are also suggestive of the potential role thatset fluency can play in cognitive evaluation processes incomparative processing contexts

Implications for Consumer Health andPublic Policy

Our research additionally shows when and how differenttypes of FOP nutrition labeling systems influence con-sumersrsquo evaluations purchase intentions and choices ofhealthy products These findings have timely and importantimplications for consumer health and welfare the food andretail grocery industries and public policy The NutritionLabeling and Education Act (NLEA) had a specific goal toldquoprovide clear unambiguous nutrition informationrdquo thatwould ldquoassist consumers in maintaining healthy dietarypracticesrdquo However since its passage and implementationin 1994 obesity has reached epidemic proportions in theUnited States where two thirds of adults are now consid-ered overweight or obese (Flegal et al 2012)

Our research suggests that one potential reason that thestandardized NFP (developed in response to the NLEA)has not been effective in preventing the rise in obesity maylie in its failure to directly address the fact that consumersrsquoevaluative contexts and decision environments can varygreatly We argue conceptually and show empirically thatthe type of processing environment faced by consumers(comparative vs noncomparative) influences the extent towhich food package cues can positively impact their food-related evaluations and decisions Our findings suggest thatinformation that is more detailed and objective may benefitconsumers more when they noncomparatively assess thehealthfulness of a single product However when con-sumers make relative comparisons between many differentbrands in a set (eg in large supermarket environments)information that is more evaluative in nature is likely to bemore beneficial in assessing product healthfulness andmaking healthy choices

We believe that these critical differences should be di-rectly considered by policymakers particularly in light ofthe numerous different FOP labeling systems currently inthe marketplace (and those currently being developed) Ifthe specific goal of the health community and policy mak-ers is to help consumers make healthier choices then the

NEWMAN ET AL 763

ownloaded from

ability to easily identify the most healthful alternativesfrom broad choice sets at the retail level is crucial (IOM2010) The implementation of a standardized evaluativeFOP cue in these comparative processing settings wouldlikely be most beneficial to consumers and positively influ-ence the healthfulness of their choices More specifically astandardized evaluative cue should help shoppers betterdistinguish between relatively more and less healthfulproducts at the retail shelf thereby increasing the likeli-hood that they choose healthier products However in gen-eral our findings suggest there also may be some potentialbenefits of adding either type of FOP nutrition information(evaluative or objective) relative to food packages thatonly offer the NFP

Limitations and Future Research

The present research has several limitations that suggesta number of potentially promising future researchopportunities The objectiveevaluative cue comparativenoncomparative processing and productset fluency frame-works used here have not been directly tested in other con-texts to our knowledge Future research can furtherexamine these typologies for any possible context effectsand enhanced generalizability While we tested several dif-ferent FOP labeling systems a number of other objectiveand evaluative formats can also be assessed Subsequentresearch may also seek to manipulate both evaluative andobjective cues together across different processing contextsto expand on the experimental designs used here Next ourstudies did not consider the price implied or explicit FOPhealth or nutrient content claims (FDA 2015) or other pro-motions that may influence consumersrsquo evaluations andchoices Research on a broader array of cue types with aconsideration of price and other promotions as well as po-tential differences across various segments of consumers(eg levels of objective knowledge and expertise low in-come and education high body mass index consumers nu-trition processing motivation) is desirable Additionallywhile our results provide some initial support for the prod-uct and set fluency concepts more research is needed tofurther refine these measures and to assess their reliabilityand predictive validity in other contexts

Future research could also assess potential differences inprocessing contexts posed by a retailerrsquos online choice en-vironment and its in-store choice environment Objectivecues might be more beneficial to online shoppers (whomay more often engage in noncomparative evaluations ofsingle products) while evaluative cues may be more bene-ficial to in-store shoppers (who are likely confronted withmany options simultaneously on retail shelves) Relatedlywe varied the size of the product set in the comparative set-tings in our studies but future research could further assesshow the effects of cues may vary as the number of productsin the set increases (ie as the complexity of the

processing task increases) Doing so may identify impor-tant boundary conditions for the effectiveness of (evalua-tive) FOP cues in comparative settings Lastly while weconducted two of our studies in a retail lab that offers asomewhat more realistic environment than many online ex-perimental settings it is acknowledged that actual retailstore environments are more challenging comparative pro-cessing settings for consumers and offer opportunities foradditional studies and analyses in general Overall addi-tional research on the complex relationships between dif-ferent types of cues processing contexts and fluenciesbeyond those examined here is warranted

DATA COLLECTION INFORMATION

All three authors contributed to the design of each studyThe first author was primarily responsible for the data anal-ysis of each study The pilot study was collected online inSummer 2015 from MTurk Study 1A was collected onlinein Fall 2011 from MTurk and study 1B was collected on-line in Fall 2014 from MTurk Studies 2 and 3 were col-lected in the Shopper Experimental Lab Facility at theUniversity of Arkansas in Fall 2014 and Spring 2012 re-spectively with the assistance of three undergraduate labassistants under the primary supervision of the first authorThese samples were recruited through both undergraduateclasses and the labrsquos subject pool All three authors contrib-uted to the construction and editing of the manuscript

REFERENCES

Anand Punam and Brian Sternthal (1987) ldquoStrategies forDesigning Persuasive Messagesrdquo in Cognitive and AffectiveResponses to Advertising ed Patricia Cafferata and AliceTybout Lexington MA Lexington Books 135ndash59

Andrews J Craig Scot Burton and Jeremy Kees (2011) ldquoIsSimpler Always Better Consumer Evaluations of Front-of-Package Nutrition Symbolsrdquo Journal of Public Policy andMarketing 30 (2) 175ndash90

Areni Charles S Dale F Duhan and Pamela Kiecker (1999)ldquoPoint-of-Purchase Displays Product Organization andBrand Purchase Likelihoodsrdquo Journal of the Academy ofMarketing Science 27 (4) 428ndash41

Berger Jonah and Grainne Fitzsimons (2008) ldquoDogs on theStreet Pumas on Your Feet How Cues in the EnvironmentInfluence Product Evaluation and Choicerdquo Journal ofMarketing Research 45 (1) 1ndash14

Berning Joshua P Hayley H Chouinard and Jill J McCluskey(2008) ldquoConsumer Preferences for Detailed VersusSummary Formats of Nutrition Information on Grocery StoreShelf Labelsrdquo Journal of Agricultural amp Food IndustrialOrganization 6 (1) 1ndash22

Bettman James R Mary Frances Luce and John W Payne(1998) ldquoConstructive Consumer Choice Processesrdquo Journalof Consumer Research 25 (3) 187ndash217

Bettman James R John W Payne and Richard Staelin (1986)ldquoCognitive Considerations in Designing Effective Labels for

ownloaded from

Presenting Risk Informationrdquo Journal of Public Policy andMarketing 5 (1) 1ndash28

Cabeza Roberto and Nobuo Ohta (1993) ldquoDissociatingConceptual Priming Perceptual Priming and ExplicitMemoryrdquo European Journal of Cognitive Psychology 5 (1)35ndash53

Chakravarti Amitav and Chris Janiszewski (2003) ldquoTheInfluence of Macro-Level Motives on Consideration SetComposition in Novel Purchase Situationsrdquo Journal ofConsumer Research 30 (2) 244ndash58

Dhar Ravi Stephen M Nowlis and Steven J Sherman (1999)ldquoComparison Effects on Preference Constructionrdquo Journalof Consumer Research 26 (3) 293ndash306

Fang Xiang Surendra Singh and Rohini Ahluwalia (2007) ldquoAnExamination of Different Explanations for the MereExposure Effectrdquo Journal of Consumer Research 34 (1)97ndash103

Federal Register (2010) ldquoFront-of-Pack and Shelf Tag NutritionSymbols Establishment of Docket Request for Commentsand Informationrdquo httpwwwgpogovfdsyspkgFR-2010-04-29html2010-9939htm

Feunekes Gerda I J Ilse A Gortemaker Astrid A WillemsRene Lion and Marcelle van den Kommer (2008) ldquoFront-of-Pack Nutrition Labelling Testing Effectiveness of DifferentNutrition Labelling Formats Front-of-Pack in Four EuropeanCountriesrdquo Appetite 50 (1) 57ndash70

Flegal Katherine M Margaret D Carroll Brian K Kit andCynthia L Ogden (2012) ldquoPrevalence of Obesity and Trendsin the Distribution of Body Mass Index Among US Adults1999-2010rdquo Journal of the American Medical Association307 (5) 491ndash97

Food and Drug Administration (2013) ldquoFDA Note toCorrespondentsrdquo httpwwwfdagovNewsEventsNewsroomPressAnnouncementsucm209953htm

mdashmdash (2015) ldquoFront-of-Package Labeling Initiative Questions ampAnswersrdquo httpwwwfdagovFoodIngredientsPackagingLabelingLabelingNutritionucm202734

Grocery Manufacturers Association (2013) ldquoFacts up FrontFront-of-Pack Labeling Initiativerdquo httpwwwgmaonlineorgissues-policyhealth-nutritionfacts-up-front-front-of-pack-labeling-initiative

Hayes Andrew F (2013) Introduction to Mediation Moderationand Conditional Process Analysis New York Guilford

Hieke Sophie and Charles R Taylor (2012) ldquoA Critical Reviewof the Literature on Nutrition Labelingrdquo Journal ofConsumer Affairs 46 (1) 120ndash25

Hsee Christopher K and France Leclerc (1998) ldquoWillProducts Look More Attractive When Presented Separatelyor Togetherrdquo Journal of Consumer Research 25 (2)175ndash86

Hsee Christopher K Jiao Zhang Liangyan Wang and ShirleyZhang (2013) ldquoMagnitude Time and Risk Differ SimilarlyBetween Joint and Single Evaluationsrdquo Journal of ConsumerResearch 40 (1) 172ndash84

Hutchinson J Wesley and Joseph W Alba (1991) ldquoIgnoringIrrelevant Information Situational Determinants ofConsumer Learningrdquo Journal of Consumer Research 18 (3)325ndash45

Institute of Medicine [IOM] (2010) Examination of Front-of-Package Nutrition Rating Systems and Symbols Phase 1Report Washington DC National Academies Press

Jacoby Larry L and Mark Dallas (1981) ldquoOn the RelationshipBetween Autobiographical Memory and Perceptual

Learningrdquo Journal of Experimental Psychology General110 (3) 306ndash40

Jacoby Larry L Colleen M Kelley and Jane Dywan (1989)ldquoMemory Attributionsrdquo in Varieties of Memory andConsciousness Essays in Honour of Endel Tulving edHenry L Roediger III and Fergus I M Craik Hillsdale NJPsychology Press 391ndash422

Kardes Frank R David M Sanbonmatsu Maria L Cronley andDavid C Houghton (2002) ldquoConsideration SetOvervaluation When Impossibly Favorable Ratings of a Setof Brands Are Observedrdquo Journal of Consumer Psychology12 (4) 353ndash61

Keller Punam Anand and Lauren G Block (1997) ldquoVividnessEffects A Resource-Matching Perspectiverdquo Journal ofConsumer Research 24 (3) 295ndash304

Keller Scott B Mike Landry Jeanne Olson Anne MVelliquette and Scot Burton (1997) ldquoThe Effects ofNutrition Package Claims Nutrition Facts Panels andMotivation to Process Nutrition Information on ConsumerProduct Evaluationsrdquo Journal of Public Policy andMarketing 16 (2) 256ndash69

Labroo Aparna A Ravi Dhar and Norbert Schwarz (2008) ldquoOfFrog Wines and Frowning Watches Semantic PrimingPerceptual Fluency and Brand Evaluationrdquo Journal ofConsumer Research 34 (6) 819ndash31

Labroo Aparna A and Angela Y Lee (2006) ldquoBetween TwoBrands A Goal Fluency Account of Brand EvaluationsrdquoJournal of Marketing Research 43 (3) 374ndash85

Lee Angela Y (2002) ldquoEffects of Implicit Memory on Memory-Based Versus Stimulus Based Brand Choicerdquo Journal ofMarketing Research 39 (4) 440ndash54

Lee Angela Y and Jennifer L Aaker (2004) ldquoBringing the Fameinto Focus The Influence of Regulatory Fit on ProcessingFluency and Persuasionrdquo Journal of Personality and SocialPsychology 86 (2) 205ndash18

Lee Angela Y and Aparna A Labroo (2004) ldquoThe Effect ofConceptual and Perceptual Fluency on Brand EvaluationrdquoJournal of Marketing Research 41 (2) 151ndash65

Mantel Susan Powell and James J Kellaris (2003) ldquoCognitiveDeterminants of Consumersrsquo Time Perceptions The Impactof Resources Required and Availablerdquo Journal of ConsumerResearch 29 (4) 531ndash38

Mayo Clinic Staff (2013) ldquoDASH Diet Tips for Shopping andCookingrdquo httpwwwmayoclinicorgdash-dietart-20045913

Meyers-Levy Joan and Laura A Peracchio (1995) ldquoUnderstandingthe Effects of Color How the Correspondence BetweenAvailable and Required Resources Affects Attitudesrdquo Journalof Consumer Research 22 (2) 121ndash38

Mitra Anu Manoj Hastak Gary T Ford and Debra JonesRingold (1999) ldquoCan the Educationally DisadvantagedInterpret the FDA-Mandated Nutrition Facts Panel in thePresence of an Implied Health Claimrdquo Journal of PublicPolicy amp Marketing 18 (Spring) 106-17

Mogilner Cassie Tamar Rudnick and Sheena S Iyengar (2008)ldquoThe Mere Categorization Effect How the Presence ofCategories Increases Choosersrsquo Perceptions of AssortmentVariety and Outcome Satisfactionrdquo Journal of ConsumerResearch 35 (2) 202ndash15

Naylor Rebecca Walker Cait Poynor Lamberton and Patricia MWest (2012) ldquoBeyond the lsquoLikersquo Button The Impact ofMere Virtual Presence on Brand Evaluations and PurchaseIntentions in Social Media Settingsrdquo Journal of Marketing76 (6) 105ndash20

NEWMAN ET AL 765

ownloaded from

Nedungadi Prakash (1990) ldquoRecall and Consumer ConsiderationSets Influencing Choice Without Altering BrandEvaluationsrdquo Journal of Consumer Research 17 (3) 263ndash76

Nestle Marion (2006) What to Eat New York North Point PressNosofsky Robert M (1986) ldquoAttention Similarity and the

IdentificationndashCategorization Relationshiprdquo Journal ofExperimental Psychology General 115 (1) 39ndash57

Novemsky Nathan Ravi Dhar Norbert Schwarz and ItamarSimonson (2007) ldquoPreference Fluency in Choicerdquo Journalof Marketing Research 44 (3) 347ndash56

Nowlis Stephen M and Itamar Simonson (1997) ldquoAttribute-TaskCompatibility as a Determinant of Consumer PreferenceReversalsrdquo Journal of Marketing Research 34 (2) 205ndash18

Oakley James L Adam Duhachek Subramanian Balachanderand S Sriram (2008) ldquoOrder of Entry and the ModeratingRole of Comparison Brands in Brand Extension EvaluationrdquoJournal of Consumer Research 34 (5) 706ndash12

Olsen Svein O (2002) ldquoComparative Evaluation and theRelationship Between Quality Satisfaction and RepurchaseLoyaltyrdquo Journal of the Academy of Marketing Science 30(3) 240ndash49

Payne John W James R Bettman and Eric J Johnson (1992)ldquoBehavioral Decision Research A ConstructiveProcessing Perspectiverdquo Annual Review of Psychology 43(1) 87ndash131

Point of Purchase Advertising International (2014) ldquo2014 MassMerchant Shopper Engagement Studyrdquo httpwwwpopaicomwp-contentuploads2014022014-Mass-Merchant-Study-Media1pdf

Prabhaker Paul R and Paul Sauer (1994) ldquoHierarchicalHeuristics in Evaluation of Competitive Brands Based onMultiple Cuesrdquo Psychology amp Marketing 11 (3) 217ndash34

Roberto Christina A Marie A Bragg Marlene B SchwartzMarissa J Seamans Aviva Musicus Nicole Novak andKelly D Brownell (2012) ldquoFacts Up Front Versus TrafficLight Food Labels A Randomized Controlled TrialrdquoAmerican Journal of Preventive Medicine 43 (2) 134ndash41

Schwarz Norbert (2004) ldquoMeta-Cognitive Experiences inConsumer Judgment and Decision Makingrdquo Journal ofConsumer Psychology 14 (4) 332ndash48

Shah Anuj K and Daniel M Oppenheimer (2007) ldquoEasy Does ItThe Role of Fluency in Cue Weightingrdquo Judgment andDecision Making 2 (6) 371ndash79

Shen Hao Yuwei Jiang and Rashmi Adaval (2010) ldquoContrastand Assimilation Effects of Processing Fluencyrdquo Journal ofConsumer Research 36 (5) 876ndash90

Slovic Paul and Douglas MacPhillamy (1974) ldquoDimensionalCommensurability and Cue Utilization in ComparativeJudgmentrdquo Organizational Behavior and HumanPerformance 11 (2) 172ndash94

Torelli Carlos J and Rohini Ahluwalia (2012) ldquoExtendingCulturally Symbolic Brands A Blessing or a CurserdquoJournal of Consumer Research 38 (5) 933ndash47

Tulving Endel and Daniel L Schacter (1990) ldquoPriming andHuman Memory Systemsrdquo Science 247 (4940) 301ndash6

Tversky Amos (1977) ldquoFeatures of Similarityrdquo PsychologicalReview 84 (4) 327ndash52

van Horen Femke and Rik Pieters (2012) ldquoWhen High-SimilarityCopycats Lose and Moderate-Similarity Copycats Gain TheImpact of Comparative Evaluationrdquo Journal of MarketingResearch 49 (1) 83ndash91

Viswanathan Madhubalan Manoj Hastak and Roland Gau(2009) ldquoUnderstanding and Facilitating the Usage ofNutritional Labels by Low-Literate Consumersrdquo Journal ofPublic Policy and Marketing 28 (2) 135ndash45

Wansink Brian and Jeffery Sobal (2007) ldquoMindless Eating The200 Daily Food Decisions We Overlookrdquo Environment andBehavior 39 (1) 106ndash23

Whittlesea Bruce W A (1993) ldquoIllusions of FamiliarityrdquoJournal of Experimental Psychology Learning Memoryand Cognition 19 (6) 1235ndash53

Winkielman Piotr David E Huber Liam Kavanagh and NorbertSchwarz (2012) ldquoFluency of Consistency When Thoughts FitNicely and Flow Smoothlyrdquo in Cognitive Consistency AFundamental Principal in Social Cognition ed B Gawronskiand F Strack New York Guildford Press 89ndash111

Winkielman Piotr Norbert Schwarz Tetra Fazendeiro and RolfReber (2003) ldquoThe Hedonic Marking of Processing FluencyImplications for Evaluative Judgmentrdquo in The Psychology ofEvaluation Affective Processes in Cognition and Emotioned Jochen Musch and Karl C Klauer Mahwah NJErlbaum 189ndash217

Zhao Xinshu John G Lynch and Qimei Chen (2010)ldquoReconsidering Baron and Kenny Myths and Truths AboutMediation Analysisrdquo Journal of Consumer Research 37 (2)197ndash206

ownloaded from

ucv050-TF1

ucv050-TF2

INTRODUCTION

Identifying the more and less healthful packaged foodproducts can be a daunting endeavor for shoppers when

one considers that the typical supermarket carries over40000 different itemsmdashthree times more than in 1980(Nestle 2006) The friendly advice frequently offered toconsumers interested in making more healthful foodchoices at supermarkets is to ldquoshop the perimeterrdquo (MayoClinic Staff 2013) It is here that the less processed morehealthful foods such as produce eggs lean beef chickenand fresh seafood can be found However inside the pe-rimeter the healthfulness of processed packaged foods canbe somewhat more difficult to discern quickly andaccurately

This complex retail choice environment provides excel-lent opportunities for marketers to influence shoppersrsquo pur-chase decisions by offering cues to simplify theirevaluative and choice processes (Bettman Luce andPayne 1998) With consumers now making 82 of theirpurchase decisions inside the store (Point of PurchaseAdvertising International 2014) marketing promotions de-signed to influence consumersrsquo product evaluations at thepoint of purchase have become critically important For ex-ample many food manufacturers and retailers are usingfront-of-package (FOP) labeling to attract attention and in-fluence perceptions at the point of sale In contrast to thedetailed Nutrition Facts panel (NFP) shown on the back orside of most packaged foods an FOP label presents con-sumers with a single condensed metric of nutrition infor-mation that is presumed to require less effort and time toprocess

Numerous retailers manufacturers and nongovernmen-tal organizations have implemented a wide variety of FOPlabeling programs such as Walmartrsquos ldquoGreat for Yourdquo ini-tiative Wingmanrsquos Wellness Keys and Unileverrsquos ChoicesProgrammed However these promotional programs varygreatly in terms of both information content and formatSome provide consumers with summarized objective nutri-tion information taken from the NFP others offer evalua-tive nutrition information that provides an interpretation ofa productrsquos overall healthfulness Although the stated pur-pose of FOP nutrition labeling is to ldquoeducate consumersand help them make healthier food choicesrdquo (FederalRegister 2010 see also Food and Drug Administration[FDA] 2015) the types of FOP labeling programs that bestengage and ultimately influence consumersrsquo product evalu-ations and decision-making processes are still poorly un-derstood (Andrews Burton and Kees 2011 Institute ofMedicine [IOM] 2010) Moreover we are not aware of anyprior nutrition labeling research that has specifically con-sidered how the effects of these different systems mightvary according to the particular type of processing task en-countered by consumers (ie noncomparative vs compar-ative evaluative tasks) (Hieke and Taylor 2012)

Therefore across two online experiments and two retaillaboratory experiments we utilize an objectiveevaluativecue framework and a comparativenoncomparative pro-cessing context framework (Oakley et al 2008 Olsen2002) to assess the interactive effects of FOP cue type andprocessing context on consumersrsquo evaluations and inten-tions to purchase healthy food products We also expandon the processing fluency literature (Jacoby and Dallas1981 Lee and Labroo 2004 Schwarz 2004 Whittlesea1993) to explain the processes underlying these effects Inthe following section we first present prior literature andbackground information on fluency different types of FOPcues (objective vs evaluative) and different types of pro-cessing contexts (comparative vs noncomparative) Wethen utilize resource matching theory (RMT) (Anand andSternthal 1987 Keller and Block 1997 Meyers-Levy andPeracchio 1995) to integrate these concepts and to guidethe development of our specific hypotheses

THEORETICAL FOUNDATION

Processing Fluency and Cue Type

Consumers view a product as an array of cues that canbe processed with varying degrees of effort speed andaccuracy (Novemsky et al 2007) Prior research has dem-onstrated that consumers base their judgments not only onavailable cue information but also on the ease with whichthey can process that information and generate relatedthoughts (Schwarz 2004) More specifically processingfluency is defined as the subjective feelings of ease or dif-ficulty with which external information can be processed(Schwarz 2004) Fluency can arise from either the pro-cessing of physical characteristics of a stimulus such asmodality or shape (perceptual fluency) or from the mean-ing of a stimulus (conceptual fluency) (Lee and Labroo2004 Tulving and Schacter 1990 Whittlesea 1993)These two types of fluency represent distinct constructsand have unique antecedents and consequences (Cabezaand Ohta 1993 Lee 2002) For example Labroo and Lee(2006) note that a perceptually fluent brand can be easilyrecognized and identified by consumers whereas a con-ceptually fluent brand is one whose meaning and associa-tions come to mind more easily Thus conceptual fluencyrelates to higher order reasoning and interpretive pro-cesses (Jacoby Kelley and Dywan 1989 Winkielmanet al 2012) Consumers often place more weight on cueinformation that feels easier to process when formingjudgments and making decisions (Shah and Oppenheimer2007)

In this research we propose that alternative types ofFOP nutrition cues lead to varying levels of conceptualfluency across different processing contexts and as a re-sult they have divergent effects on consumersrsquo food evalu-ations intentions and choices To better understand

ownloaded from

HYPOTHESES

NEWMAN ET AL 751

ownloaded from

FIGURE 1

FOP Cue Type

Processing Context

Healthfulness

Evaluations and

Purchase Intentions

NEWMAN ET AL 753

ownloaded from

PILOT STUDY

ownloaded from

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

HYPOTHESES

NEWMAN ET AL 751

ownloaded from

FIGURE 1

FOP Cue Type

Processing Context

Healthfulness

Evaluations and

Purchase Intentions

NEWMAN ET AL 753

ownloaded from

PILOT STUDY

ownloaded from

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

ownloaded from

FIGURE 1

FOP Cue Type

Processing Context

Healthfulness

Evaluations and

Purchase Intentions

NEWMAN ET AL 753

ownloaded from

PILOT STUDY

ownloaded from

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

FIGURE 1

FOP Cue Type

Processing Context

Healthfulness

Evaluations and

Purchase Intentions

NEWMAN ET AL 753

ownloaded from

PILOT STUDY

ownloaded from

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

PILOT STUDY

ownloaded from

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

STUDY 1A

Methods

Dependent Measures

RESULTS

Manipulation Check

FIGURE 2

Level F

luency

NEWMAN ET AL 755

ownloaded from

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

DISCUSSION

STUDY 1B

Methods

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

DISCUSSION

STUDY 2

Methods

FIGURE 3

NEWMAN ET AL 757

ownloaded from

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

Dependent Measures

RESULTS

Manipulation Check

ownloaded from

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

FIGURE 4

Level F

luency

FIGURE 5

Level F

luency

NEWMAN ET AL 759

ownloaded from

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

DISCUSSION

FIGURE 6

Healthfu

ations

TABLE 1

ownloaded from

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

STUDY 3

Methods

Dependent Measures

TABLE 2

NEWMAN ET AL 761

ownloaded from

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

RESULTS

GENERAL DISCUSSION

ownloaded from

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

NEWMAN ET AL 763

ownloaded from

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

REFERENCES

ownloaded from

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

NEWMAN ET AL 765

ownloaded from

ucv050-TF1

ucv050-TF2

ownloaded from

ucv050-TF1

ucv050-TF2

Documents

Effects of Objective and Evaluative Front-of-Package Cues on ......([email protected]) is professor of marketing, Sam M. Walton College of Business, University of Arkansas,