941

2008 by JOURNAL OF CONSUMER RESEARCH, Inc. Vol. 35 April 2009All rights reserved. 0093-5301/2009/3506-0001$10.00. DOI: 10.1086/593692

Variety, Vice, and Virtue: How Assortment SizeInfluences Option Choice

ANER SELAJONAH BERGERWENDY LIU*

Assortment size has been shown to influence whether consumers make a choice,but could it also influence what they choose? Five studies demonstrate that be-cause choosing from larger assortments is often more difficult, it leads people toselect options that are easier to justify. Virtues and utilitarian necessities are gen-erally easier to justify than indulgences; consequently, choosing from larger as-sortments often shifts choice from vices to virtues and from hedonic to utilitarianoptions. These effects reverse, however, when situational factors provide acces-sible reasons to indulge, underscoring the role of justification. Implications forchoice difficulty and justification processes are discussed.

When going out to eat, consumers often get mired inone of lifes eternal strugglesthe choice betweenvice and virtue. Restaurant menus frequently include deli-cious but unhealthy options (e.g., pizza) as well as whole-some but less tasty alternatives (e.g., salads). Though suchdecisions can be resolved in different ways, one interestingquestion is whether menu size could influence what peoplechoose. That is, could the number of options on the menuinfluence the type of option consumers select? For example,could longer menus lead people to choose more salads?

In recent years, a resurgence of interest in how varietyinfluences consumer choice likelihood and satisfaction hasoffered an important correction to the notion that morechoice is always better (Chernev 2003a, 2003b; Iyengar andLepper 2000; Schwartz 2004). This stream of research dem-onstrates that too many options can lead consumers to notchoose at all and to feel less satisfaction and more regretabout the options they choose. But while this work has

*Aner Sela is a doctoral candidate in marketing at the Graduate Schoolof Business, Stanford University, 518 Memorial Way, Stanford, CA 94305(sela_aner@gsb.stanford.edu). Jonah Berger is assistant professor of mar-keting at the Wharton School, University of Pennsylvania, 700 Jon M.Huntsman Hall, 3730 Walnut Street, Philadelphia, PA 19104 ( jberger@wharton.upenn.edu). Wendy Liu is assistant professor of marketing at theAnderson School of Management, University of California, Los Angeles,110 Westwood Plaza, Los Angeles, CA 90095 (wendy.liu@anderson.ucla.edu). The first two authors contributed equally to this research. The authorswish to thank Itamar Simonson, Uzma Khan, the editor, associate editor,and reviewers for helpful comments that greatly strengthened the article.

John Deighton served as editor and Stephen Nowlis served as associateeditor for this article.

Electronically published October 15, 2008

provided insight into how the number of available optionsaffects whether consumers make a choice (and their accom-panying feelings), much less attention has been paid to howvariety might influence what consumers choose.

This article examines how assortment size influences thechoice between vice and virtue or hedonic and utilitarianoptions. When choosing whether to watch a slapstick com-edy versus an award-wining drama, eat an ice cream barversus a granola bar, or read a world news magazine versusa celebrity magazine, the number of options consumerschoose from can vary greatly. They may choose TV showsfrom hundreds of cable channels or from only the basicnetworks, select food from a tiny hotel minibar or from alarge vending machine, and pick a magazine from the fewoptions available at the dentists office or from the dozensavailable at the nearby Barnes and Noble. We argue thatassortment size can have an important influence on optionchoice. Specifically, we integrate two research streamsassortment size and justificationto suggest that becausechoosing from larger sets is often more difficult, it will leadpeople to select more justifiable options. Further, becausecertain types of options are typically easier to justify (e.g.,Kivetz and Simonson 2002b), increased variety should typ-ically increase the choice share of such options. Five studiesexamine this possibility and test our underlying conceptualframework, exploring the hypothesized roles of choice dif-ficulty and justification in these effects. We then discuss thetheoretical and practical implications of our findings anddevelop directions for future research.

VARIETY AND DECISION MAKINGThe number of available options can influence consumer

choice in multiple ways. Decades of research suggest that

942 JOURNAL OF CONSUMER RESEARCH

choice increases satisfaction (e.g., Langer and Rodin 1976)and that larger assortments increase the likelihood that con-sumers will find an option that matches their preferences(Baumol and Ide 1956; Lancaster 1990). People activelyseek variety (see Kahn and Ratner [2005] for a review),whether to satisfy the need for stimulation (Berlyne 1960)or for other reasons, and larger assortments help consumerssatisfy these needs.

However, while much of the earlier work on variety eithersuggested or assumed that more options were better, recentresearch has highlighted the downsides of too muchchoice. Choice can be overwhelming, and choosing fromlarger assortments increases choice difficulty and regret(Berger, Draganska, and Simonson 2007; Iyengar and Lep-per 2000). Consequently, when faced with too many options,consumers may defer choice or not choose at all (Dhar 1997;Iyengar, Huberman, and Jiang 2004; Iyengar and Lepper2000; Tversky and Shafir 1992). Consumers were less likelyto purchase jam, for example, when they were given anextensive array of 30 flavors to choose from as opposed toa more limited array of six flavors (Iyengar and Lepper2000). Similarly, choosing from larger assortments can in-fluence consumers postdecision satisfaction with their cho-sen options, leading to weaker preferences for the itemsselected and to greater regret (Chernev 2003b; Iyengar andLepper 2000; Schwartz 2004).

Nevertheless, while the above research focuses on howvariety influences choice difficulty, likelihood, and satisfac-tion, much less attention has been paid to how assortmentsize might influence the type of option consumers select.Although too many options may make choice difficult andlead consumers to defer choice or to not choose at all, whathappens in the preponderance of situations in which con-sumers do make a choice? Could the number of alternativesinfluence the process by which people choose and, conse-quently, what they select?

DECISION DIFFICULTYAND JUSTIFICATION

When faced with difficult decisions, consumers oftensearch for reasons to justify their choices. While normativetheories focus on the utility associated with various aspectsof the decision, other research notes that decision makingis often driven by a reason-based analysis (Shafir, Simonson,and Tversky 1993). This framework suggests that consumersseek reasons to resolve conflict and justify the options theyselect. In such instances, consumers often focus their de-cision processes on the choice of good reasons rather thanon the choice of good options (Simonson and Nowlis 2000).As the conflict and uncertainty associated with choice in-creases, the focus of consumers reasoning processes mayshift from desirability to justifiability. Options that provideconvincing, readily justifiable arguments are therefore morelikely to be chosen (Shafir et al. 1993; Simonson 1989,1992).

Certain types of options are easier to justify than others.

Broadly speaking, it is easier to justify the choice of utili-tarian necessities and virtues as opposed to hedonic indul-gences and vices (Bazerman, Tenbrunsel, and Wade-Benzoni1998; Kivetz 1999; Kivetz and Keinan 2006; Kivetz andSimonson 2002a). Hedonic consumption is harder to justify,in part, because it is often associated with guilt (Kivetz andSimonson 2002b; OCurry and Strahilevitz 2001; Prelecand Loewenstein 1998; Strahilevitz and Myers 1998). TheProtestant ethic emphasizes frugality (Weber 1958), andconsumers sometimes feel guilty about pleasurable con-sumption (Lascu 1991). In addition, it is often easier toconstruct reasons to justify utilitarian consumption. Thebenefit of hedonic goods lies in experiential enjoyment,which is harder to quantify than the more concrete benefitsthat utilitarian goods often provide. Quantifiable reasons aremore easily justified (Hsee 1995; Shafir et al. 1993), andthe search for reasons may increase utilitarian choice.

Consequently, situations that increase reliance on avail-able justifications can affect the types of options consumersselect. For example, choosing for public rather than privateconsumption generally favors options that are higher onutilitarian aspects (Bohm and Pfister 1996). People wereless likely to prefer an elegant motor scooter as a prize overmore virtuous options when they would supposedly receiveit on a public TV show as opposed to in a private maillottery. Importantly, however, the same mechanism can po-tentially make people more likely to prefer hedonic optionsif situational factors provide accessible reasons to indulge.For example, a task involving much personal effort mightlead people to select a hedonic reward because they feelthey have earned the right to indulge and thus can justifysuch a choice (Kivetz and Simonson 2002a; Kivetz andZheng 2006).

THE CURRENT RESEARCHBuilding on these findings, we argue that assortment size

will influence the type of options consumers select. Further,we develop a justification-based framework to account forthese effects. Because choosing from larger assortmentstends to be more difficult, we argue that it will often leadconsumers to select options that are easier to justify, suchas healthy food or highbrow television. Consequently, peo-ple faced with a larger menu might be more likely to takethe garden salad over the pepperoni pizza or the reducedfat strawberry ice cream over the double chocolate mochacrunch. Similarly, choosing from a larger array of channelsmight lead people to choose award-winning dramas andinsightful news over slapstick comedies and shoot-em-upaction movies.

However, while healthier or more virtuous options tendto be easier to justify in general, situational factors canprovide accessible justifications to indulge. Consequently,we propose that the effect of variety on choice will be mod-erated by accessible justifications. Under neutral or low-justifiability conditions, choosing from larger assortmentswill increase the share of utilitarian or virtuous options.However, when situational factors provide accessible rea-

HOW ASSORTMENT SIZE INFLUENCES OPTION CHOICE 943

sons to indulge, choosing from larger assortments will havethe opposite effect, increasing the share of vices or hedonicchoices.

Five studies test these hypotheses. First, they examinewhether assortment size can influence the type of optionsconsumers select. They demonstrate that choosing from alarger assortment can lead people to select reduced fat overregular ice cream (experiment 1A), fruit over cookies (ex-periment 1B), and utilitarian over hedonic consumer prod-ucts (experiments 25). Second, the studies test the under-lying mechanism behind these effects. They examine themediating role of choice difficulty (experiment 3) and themoderating role of accessible justifications (experiments 4and 5).

EXPERIMENT 1: VARIETY ANDOPTION CHOICE

Experiment 1A: Ice Cream

Experiment 1A provides a preliminary investigation intowhether assortment size influences the type of options con-sumers select. Participants were shown either a smaller ora larger array of ice cream options and were asked to choosethe option they liked best. Each flavor had both a regularversion and a reduced fat version. We compared the like-lihood of selecting a regular versus a reduced fat optionacross the two conditions. Prior research suggests it is oftenharder to justify indulgences (Kivetz and Simonson 2002b).Thus, we predict that people choosing from a larger assort-ment will be more likely to select reduced fat ice cream.

Method

Participants. One hundred and twenty-one respondentscompleted a group of studies in exchange for a chance towin a $25 gift certificate.

Materials and Procedure. Participants were shownpictures of ice cream and were asked to select their preferredflavor. For each flavor there was both a regular version (e.g.,vanilla) and a reduced fat version (e.g., reduced fat vanilla).Both versions used the same picture but had different labels.The key difference between conditions was the number ofavailable options. In the low-variety condition there weretwo options available, one regular and one reduced fat. Inthe high-variety condition there were 10 options available,five in each category.

Results

As predicted, assortment size influenced the type of op-tion people chose. Participants who chose from a largerassortment were more likely to select a reduced fat flavor( , ). While only 20% of participants2x (1) p 4.08 p ! .05who chose from the smaller assortment selected reduced fatice cream, 37% who chose from the larger assortment se-lected reduced fat.

Experiment 1B: Fruit or Cookies?

Experiment 1B used a real choice situation to investigatehow assortment size influences option choice in a field set-ting. A tray containing fruits and cookies was placed at eachof two building entrances, and a sign indicated that passersbycould help themselves to an item. We manipulated whetherthe tray contained a smaller or a larger array of options.Consistent with our conceptualization, we predicted thatpeople choosing from larger assortments would be morelikely to select fruit over cookies.

Method

Participants and Procedure. Seventy-five people atStanford University completed the study over the course ofa day. Trays were mounted on stands and placed at twodoors located on opposite sides of a campus building. At-tached to each tray was a printed sign indicating, Pleasehelp yourself to one item! One tray contained six types offruit (banana, red apple, pear, green apple, tangerine, andpeach) and six types of cookies and cakes (chocolate chip,oatmeal raisin, white chocolate chip, and M&M cookies;minicroissants; and banana nut muffins). The other tray con-tained only two options from each category. The exact con-tent of the smaller assortment was constantly counterbal-anced as options from the tray were consumed. The trayswere swapped between locations once during the study. Anexperimenter refilled the trays as the food was selected andinconspicuously noted the options consumed from each tray.Our key dependent variable was the type of option (fruit orcookie) chosen.

Results

As predicted, participants were more likely to selecthealthy fruit over tempting cookies when choosing from alarger assortment ( , ). While 55% of2x (1) p 3.90 p ! .05participants chose fruit over cookies from the smaller as-sortment, 76% of participants who chose from a larger as-sortment selected fruit.

Discussion

Experiments 1A and 1B provide preliminary support forour theorizing. Choosing from a larger set led consumersto select options that should have been easier to justify.People were more likely to select reduced fat over regularice cream (experiment 1A) or fruit over cookies for a reallunchtime snack (experiment 1B) when they were faced witha larger array of options.

One might question, however, whether our results are dueto the extra options available in the larger assortment con-ditions. Increasing the available options in a category shouldlead consumers to be more likely to find an option that fitstheir preferences. Participants in experiment 1A, for ex-ample, might have preferred reduced fat ice cream overallbut chose regular ice cream in the smaller variety condition

944 JOURNAL OF CONSUMER RESEARCH

FIGURE 1

INFLUENCE OF VARIETY ON OPTION CHOICE(EXPERIMENT 2)

because, within the vanilla flavor, they happened to likeregular more than reduced fat. Alternatively, one could arguethat because each choice involves a trade-off between twosalient goals (e.g., pleasure and good health), people bal-anced those goals within option choice (cf. Dhar and Si-monson 1999). Participants might have settled for a lesstasty ice cream (i.e., reduced fat) when variety ensured thatthey could get a preferred flavor but might have selectedthe tastier option (i.e., regular fat) from the smaller assort-ment when they could not get the flavor they preferred.

Experiment 1B partially addressed this issue by varyingthe specific content of the tray in the smaller variety con-dition throughout the experiment. To more directly rule itout, however, the next study examines whether increasingthe number of either hedonic or utilitarian options increasesthe choice share of utilitarian items.

EXPERIMENT 2: VARYING HEDONICAND UTILITARIAN SET

SIZE SEPARATELY

Experiment 2 tests whether the effect of assortment sizecan occur independently of whether there is an increasednumber of hedonic or utilitarian options. Participants saw aset consisting of printers (smaller or larger variety) and MP3players (smaller or larger variety) and were asked to indicatewhich of the available options they preferred. This allowedus to examine how both the overall assortment size and therelative number of options in each category affect the typeof options consumers select.

We used these particular options because a pretest indi-cated that printers were perceived as more utilitarian,whereas MP3 players were perceived as more pleasurable.Specifically, pretest participants rated the extent to whicheach of these products provides practical benefits and plea-sure. The results indicated that printers were perceived asproviding more practical benefits than pleasure (M pPrac.

vs. ; , ),5.45 M p 3.20 F(1, 206) p 287.02 p ! .001Pleas.whereas the opposite was true for MP3 players (M pPract.

vs. ; , ).4.06 M p 5.68 F(1, 206) p 287.02 p ! .001Pleas.Printers were also rated as providing more practical bene-fits than MP3 players ( vs. ;M p 5.45 M p 4.06Print. MP3

, ), whereas MP3 players wereF(1, 206) p 76.71 p ! .001rated as providing more pleasure than printers (M pMP3

vs. ; , ).5.68 M p 3.20 F(1, 206) p 303.95 p ! .001Print.Because utilitarian choices (i.e., printers) are generally

easier to justify (Kivetz and Simonson 2002b), we predictedthat increasing the number of available options would leadpeople to be more likely to select a printer. Importantly,however, our theoretical framework entails that increasingthe number of either MP3 players or printers may lead toincreased selection of printers in this instance, even whenthe number of products in the opposite category remainssmall.

Method

Participants. One hundred and one students at the Uni-versity of California, Los Angeles (UCLA), completed thestudy as part of a larger group of unrelated experiments.They were randomly assigned to a condition and received$20 for their time.

Materials and Procedure. Participants were shown anarray of printers and MP3 players and were asked which ofthe available options they preferred. Options were organizedin a two-row table with the upper row stating the optionnumber and category (e.g., Option 1: printer) and thelower row containing the product name (e.g., Epson StylusC88 printer). The options for each category appeared in arow, and category order was randomized across participants.Each category featured prominent brand names, and no in-formation about specific product attributes was provided.Thus, importantly, the level of similarity between optionswas kept identical across the two categories. Pretesting alsoindicated that none of the options dominated others withinits category.

The key difference between conditions was the numberof available options in each category. All participants wereshown both printers and MP3 players, but how many ofeach appeared (two or six) varied independently in a 2(printer variety: smaller vs. larger) # 2 (MP3 player variety:smaller vs. larger) between-subjects design. The key depen-dent variable was whether participants selected a printer oran MP3 player.

Results

As predicted, assortment size influenced the type of optionselected (fig. 1). Focusing first on the symmetrical choicesets: whereas only 11% of participants chose a printer inthe smallest assortment condition (two printers and two MP3

HOW ASSORTMENT SIZE INFLUENCES OPTION CHOICE 945

players), the percentage increased to 50% in the largestassortment condition (six printers and six MP3 players;

, ). Analysis of the asymmetric choice2x (1) p 9.03 p ! .005sets provides further insights. Compared with the smallestassortment condition, the number of people who chose aprinter jumped to 32% when the number of MP3 playerswas increased ( , ) and to 54% when the2x (1) p 3.15 p ! .08number of printers was increased ( , ).2x (1) p 10.90 p ! .001This suggests that increasing the number of either printersor MP3 player options was sufficient to lead participants tochoose more printers.

Discussion

Experiment 2 further supports our suggestion that as-sortment size can influence option choice. Larger assort-ments led consumers to select options that should have beeneasier to justify. Further, results from the asymmetricalchoice sets cast doubt on the alternative account discussedpreviously. Even when the number of available printers re-mained constant, adding more MP3 players to the choiceset increased the selection of printers. This underscores thenotion that our findings are not driven by the particularoptions available in the larger set. This result also demon-strates another violation of the regularity condition (Shafiret al. 1993), whereby the tendency to select a particularoption increases with the addition of alternatives.

The first two experiments consistently demonstrate thatwhen faced with larger assortments, consumers tend to selectoptions that are easier to justify. The next three experimentswere designed to provide deeper insight into the mechanismunderlying these effects. We theorize that increasing thenumber of options should increase the difficulty associatedwith choice, which, in turn, should drive people to selectmore justifiable options. We directly test this mediation hy-pothesis in experiment 3.

EXPERIMENT 3: THE MEDIATING ROLEOF CHOICE DIFFICULTY

Experiment 3 examines the role of choice difficulty in theeffect of assortment size on option choice. As in experiment2, participants chose between different consumer electronicsoptions (printers and MP3 players), from either a larger ora smaller choice set. After making their selections, partici-pants rated the difficulty of the choice task. We predictedthat difficulty would mediate the effect of assortment sizeon option choice. A larger choice set should increase choicedifficulty, which, in turn, should lead consumers to selectoptions that are easier to justify.

Method

Participants. One hundred and fifty-six respondentscompleted this experiment as part of a group of unrelatedstudies. They were randomly assigned to a condition andwere given a chance to win a $25 gift certificate in exchangefor their participation.

Materials and Procedure. The study employed a sim-ilar paradigm as experiment 2 but with only two varietyconditions: participants chose from either a smaller assort-ment (two printers and two MP3 players) or a larger as-sortment (six printers and six MP3 players). They wereshown an array of printers and MP3 players and were askedwhich option they preferred. After making their choices,participants were asked to rate the extent to which they foundit difficult to make a decision about which option to pick( ; ).1 p not at all difficult 7 p very difficult

Results

Effect of Assortment Size on Option Choice. Repli-cating the results of our prior experiments, assortment sizeinfluenced option choice; participants were more likely tochoose a printer when choosing from a larger (51%) ascompared to smaller (34%) set ( , ).2x (1) p 6.27 p ! .01

Mediation Analysis. We also examined whether choicedifficulty mediated the effect of assortment size on optionchoice (Baron and Kenny 1986). First, a logistic regressionwith option choice (printer vs. MP3 player) as the dependentvariable and assortment size (smaller vs. larger) as the in-dependent variable revalidated the effect of assortment sizeon choice ( , ). Next, a linear re-Wald(1) p 6.18 p ! .01gression with choice difficulty as the dependent variable andassortment size as the independent variable indicated thatincreasing assortment size significantly increased choice dif-ficulty ( , ; , ). Fi-B p .90 SE p .30 t(154) p 3.03 p ! .005nally, a logistic regression model with choice as the depen-dent variable and assortment size and difficulty as theindependent variables revealed a significant effect of diffi-culty on option choice ( , ;B p .25 SE p .09 Wald(1) p

, ), while the effect of assortment size on choice7.15 p ! .01decreased ( , ). A Sobel (1982) testWald(1) p 3.41 p ! .07revealed that this decrease was significant ( ,z p 2.01 p !

), indicating that the effect of assortment size on choice.05was mediated by choice difficulty.

Discussion

Experiment 3 again demonstrates that assortment size caninfluence option choice while also illuminating the processunderlying these effects. Choosing from a larger assortmentwas more difficult, and this difficulty led people to selectoptions that were easier to justify.

Consistent with research suggesting that utilitarian choicesare generally easier to justify (Kivetz and Keinan 2006;Kivetz and Simonson 2002b), the first three studies haveillustrated that larger assortments can lead consumers toselect healthier food or more utilitarian consumer goods.But while such choices are typically easier to justify, specificsituational factors (e.g., exerting effort or committing tovirtuous choices) can provide justifications to indulge (Khanand Dhar 2006; Kivetz and Zheng 2006). Furthermore, thefact that accessible justifications can vary allows us to di-rectly test our proposed mechanism. If ease of justification

946 JOURNAL OF CONSUMER RESEARCH

underlies the effect of assortment size on choice, as wesuggest, then providing people with accessible justificationsfor hedonic choice should reverse the effect. We test thispossibility in the final two studies.

EXPERIMENT 4: MANIPULATING EASEOF JUSTIFICATION THROUGH

PRIOR CHOICEIf larger assortments lead people to resort to accessible

justifications, as we suggest, then when justifications forindulging are made accessible, increased assortment sizeshould lead people to choose more hedonic options. Priorresearch indicates that commitment to a virtuous act, suchas volunteering, can activate a positive self-concept thatsubsequently licenses the choice of a self-indulgent option(Khan and Dhar 2006). We use such a licensing manipu-lation. Before the focal choice task, half our participants(high license condition) made a hypothetical choice betweentwo volunteer activities. The other half (low license con-dition) made a hypothetical choice between two fun activ-ities. All participants were then presented with either asmaller or a larger array of laptops and were asked to choosetheir preferred option.

We also used a framing manipulation to vary whetherparticular options were seen as hedonic or utilitarian withoutactually changing the options themselves. The laptops weregrouped into two sets (A and B). For half the participants,options in one set were described as more fun oriented, whilethose in the other were described as more work oriented(see the appendix). For the other half, these labels werereversed. This manipulation bolsters construct validity be-cause it allows us to vary option justifiability without con-founding it with what is actually being chosen.

We hypothesized that the licensing manipulation wouldmoderate the effect of assortment size on option choice.Specifically, the low licensing condition should replicate theresults of our prior experiments. When faced with moreoptions, participants should be more likely to choose util-itarian options (i.e., a laptop built for work). In the highlicensing condition, however, assortment size should havethe opposite effect. When faced with a larger set of options,these participants should be more likely to choose optionsthat provide pleasure (i.e., a laptop built for fun). Further,the increased choice likelihood should depend on whetherthe laptops were framed as more hedonic or utilitarian, ratherthan on the actual options themselves.

Method

Participants. One hundred and seventy-one individualsparticipated in two ostensibly unrelated studies as part of alarger group of experiments. They were randomly assignedto conditions in a 2 (assortment size: smaller vs. larger) #2 (licensing: low vs. high) between-subjects design in ex-change for the chance to win a $25 gift certificate.

Licensing Task. Following Khan and Dhar (2006), we

told participants that the first task was a survey about dif-ferent decisions people face in everyday life. They wereasked to imagine that they decided to dedicate 3 hours everyweek to a certain goal and then were asked to select a specificactivity related to that goal. Participants in the high licensingcondition were asked to imagine that they had volunteeredto spend 3 hours a week doing community service. Theywere asked to choose between two activities, teaching chil-dren in a homeless shelter and improving the environment.In contrast, participants in the low licensing condition wereasked to imagine that they had decided to spend 3 hours aweek on fun activities. They were asked to choose betweenplaying video games and watching television. All partici-pants were also asked to state reasons for their respectivechoices, to ensure involvement in the task.

Choice Task. Participants then completed an ostensiblyunrelated study in which they chose among different lap-tops. They were shown an array of laptops (e.g., LenovoThinkPad T60) and were asked to select the laptop theypreferred. Laptops were grouped in two sets. Those in oneset were described as built for fun. They feature high-endgraphics cards, top of the line speakers, and extra music andvideo software. Laptops in the other set were described asbuilt for work. They are lightweight, have extended batterylife, high-end security, and extra business and presentationsoftware. Importantly, these descriptions were counterbal-anced across participants. For half the participants, laptopsin set A were described as fun and laptops in set B as workoriented, and for the other half the descriptions were re-versed. This design allowed us to examine whether the ef-fects of assortment size on choice persist even when theoptions themselves remain the same.

In the smaller assortment condition, participants choseamong four laptops (two fun oriented and two work ori-ented). In the larger assortment condition, they chose among12 laptops (six of each). After selecting their preferred op-tion, participants were probed for suspicion and were askedto indicate what they thought might have affected their de-cision in the choice task. None suspected that the licensingtask was related to their subsequent choice.

Results

Analysis revealed that the effect of assortment size onoption choice was moderated by accessible justifications(fig. 2). Results in the low licensing condition mirrored thoseof our prior experiments. Whereas only 30% of participantsin the smaller assortment condition selected a work laptop,52% chose a work laptop in the larger assortment condition( , ). In contrast, as predicted by our2x (1) p 3.84 p ! .05justification framework, the pattern in the high licensingcondition was reversed. While 63% of participants in thesmaller assortment condition chose a work laptop, this num-ber was reduced to 39% in the larger assortment condition( , ). Labeling order did not have an ef-2x (1) p 5.11 p ! .05fect on choice in either condition.

HOW ASSORTMENT SIZE INFLUENCES OPTION CHOICE 947

FIGURE 2

INFLUENCE OF VARIETY AND LICENSING ON OPTIONCHOICE (EXPERIMENT 4)

Discussion

The results of experiment 4 support our suggestion thataccessible justifications moderate the effect of variety onoption choice. Consistent with our previous studies, whenpleasurable options were harder to justify (i.e., when a li-censing manipulation suggested that participants did notearn the right to indulge), larger assortments led to increasedselection of utilitarian options. In contrast, when participantshad a readily accessible excuse for rewarding themselves(i.e., when the licensing manipulation suggested they hadearned the right to indulge), greater assortment increasedthe likelihood of choosing pleasurable options. Further, bykeeping the options themselves the same but manipulatingthe labeling, we underscore the hypothesized role of justi-fication in these effects: increased assortment size led par-ticipants to choose options labeled in a way that made themeasier to justify.

EXPERIMENT 5: MANIPULATING EASEOF JUSTIFICATION THROUGH EFFORT

Experiment 5 uses a different manipulation to provideconvergent evidence for the role of justification in theseeffects. This allows us to capture two qualitatively differentinstantiations of justifiability, thus broadening the concep-tual basis of our justification construct and ruling out ad-ditional rival accounts. We borrowed a relative-effort ma-nipulation from prior work on justification that found thatexpending high effort provides a compelling justificationto indulge (Kivetz and Zheng 2006, 574). Before our mainassortment-size manipulation, participants were given an os-tensibly unrelated survey that asked them to complete a setof simple problems. All participants completed the samenumber of problems, but those in the high relative-effortcondition were led to believe that they completed more prob-lems than other participants, thus providing them with an

accessible excuse to reward themselves. Participants thencompleted the same choice task as in experiment 3, in whichthey were asked to select an option from either a smalleror larger assortment of consumer electronics (i.e., printersand MP3 players).

We predicted that the low relative-effort condition wouldreplicate the results of our other experiments; when facedwith a larger set of options, participants will be more likelyto choose utilitarian necessities (i.e., a printer). In contrast,participants in the high-effort condition should show theopposite effect. When faced with a larger set of options,these participants should choose options that provide plea-sure (e.g., an MP3 player).

The design also allows us to test our justification-basedexplanation against an alternative account. One might sug-gest that, rather than justification, the effect of assortmentsize on choice was driven by the activation of differentmodes of thinking (Hsee and Rottenstreich 2004; Small,Loewenstein, and Slovic 2007). Having more options tochoose from is associated with greater cognitive effort,which may put consumers in a cognitive mode of thinking.This, in turn, might benefit more cognitive (e.g., utilitarian)and less emotional (e.g., hedonic) options because suchchoices fit with the choice process being used. We testedthis alternative account by modifying the relative-effort taskused in prior work (Kivetz and Zheng 2006). Rather thancomplete sentences, participants solved math problems,which has been shown to activate a cognitive-based modeof thought (Hsee and Rottenstreich 2004). Participantssolved the same number of problems in both high and lowrelative-effort conditions, and thus all participants shouldhave had an equally cognitive mind-set activated. Further,a cognitive mode account would suggest that because varietyinduces greater cognitive trade-offs, it should always leadpeople to choose more utilitarian options, which are con-sistent with a more cognitive mode of thought. However, ifgreater assortment leads to choice of more hedonic optionsin the high relative-effort condition, as we predict, then itwill contradict this account.

Method

Participants. One hundred and sixty-eight undergrad-uates at UCLA participated in two ostensibly unrelated stud-ies as part of a larger group of experiments. They wererandomly assigned to a condition and received $20 for theirtime.

Relative-Effort Manipulation. The effort manipula-tion required participants to solve simple calculation prob-lemsfor example, If an object travels at five feet perminute, how many feet will it travel in 360 seconds? Theywere informed that the experimenters were interested in howdifferent levels of effort influence peoples performance onsimple tasks. All participants were told that they would berandomly assigned to either a low- or a high-effort task.They were then given the same nine calculation problems,but we manipulated their perceived effort relative to other

948 JOURNAL OF CONSUMER RESEARCH

FIGURE 3

INFLUENCE OF VARIETY AND RELATIVE EFFORT ON OPTIONCHOICE (EXPERIMENT 5)

participants. In the low- (high-) effort condition, participantswere told that each participant is asked to solve anywherebetween 8 to 16 [2 to 10] calculation problems. You havebeen assigned to the low [high] effort task. All participantsthen went on to solve the same nine problems. Thus, allparticipants expended the same amount of actual effort, butwe varied the amount of effort participants felt they hadexerted relative to other participants.

Procedure. First, participants completed the relative-effort task. Then they moved on to a supposedly unrelatedchoice study. This task was printed using different font andlayout than the effort manipulation task and consisted of thesame choice task used in experiment 3. Participants chosefrom either a smaller set (two printers and two MP3 players)or a larger set (six printers and six MP3 players). Afterselecting their preferred option, they were probed for sus-picion and were asked to indicate what they thought mighthave affected their decision in the choice task. None sus-pected that the effort task was related to their subsequentchoice.

Results

As predicted, the analysis revealed that the effect of as-sortment size on option choice was moderated by perceivedrelative effort (fig. 3). Specifically, in the low relative-effortcondition, the results mirrored the findings from the firstthree experiments. Whereas only 37% of participants in thesmaller assortment condition selected a printer, this in-creased to 61% in the larger assortment condition ( 2x (1) p

, ). In contrast, as predicted by our justification4.60 p ! .05framework, the pattern in the high relative-effort conditionwas reversed. While 48% of participants in the smaller as-sortment condition selected a printer, the percentage droppedto 26% in the larger assortment condition ( ,2x (1) p 4.71

). Participants who exerted high relative effort se-p ! .05lected more hedonic options when choosing from a largerchoice set. To look at the data another way, the relative-effort manipulation influenced choice for participants facedwith a larger assortment ( , ) but not2x (1) p 10.74 p ! .001for participants faced with a smaller assortment ( 2x (1) p

, ), suggesting that the justification manipulation1.03 p 1 .3played a greater role when the choice set called for greaterreliance on justifications.

Discussion

The results of experiment 5 underscore our suggestionthat variety influences option choice through greater relianceon accessible justifications; manipulating the ease of justi-fying indulgences moderated the effect of assortment sizeon option choice. Consistent with our previous studies, whenpleasurable options were harder to justify (i.e., when a ma-nipulation suggested that participants expended less effortthan their peers), a larger assortment led to increased choiceof utilitarian options. In contrast, when participants had areadily accessible excuse for rewarding themselves (i.e.,

when the manipulation suggested they had exerted moreeffort than their peers), greater assortment had the oppositeeffect. For these participants, greater assortment increasedthe likelihood of choosing pleasurable options.

The results also cast doubt on an alternative account basedon a cognitive mode of thought. Such an account has dif-ficulty explaining why the results would be moderated bya justification-based manipulation. Further, a cognitive modeaccount would suggest that increased assortment should al-ways lead to increased choice of utilitarian options and thusis less useful in explaining why participants in the highrelative-effort condition select more hedonic options whenchoosing from a larger set.

GENERAL DISCUSSION

This article examines how assortment size influences op-tion choice. Prior research has investigated how too muchchoice can reduce choice likelihood. In contrast, the presentresearch demonstrates that the number of options consumersare faced with can affect not only whether they choose butalso what they choose. Choosing from larger assortmentstends to increase choice difficulty and, consequently, cancause consumers to rely more on accessible justificationswhen making their choice. As a result, we argue that choos-ing from a larger assortment should lead consumers to selectoptions that are easier to justify.

Five experiments support our perspective. First, they dem-onstrate that when faced with larger choice sets, people tendto select more virtuous options. They selected reduced fatover regular ice cream (experiment 1A), fruit over cookies(experiment 1B), and utilitarian over hedonic consumerproducts (experiments 25).

Second, the studies demonstrate how choice difficulty andoption justifiability drive these effects. Choice difficulty me-diated the effect of assortment size on option choice (ex-

HOW ASSORTMENT SIZE INFLUENCES OPTION CHOICE 949

periment 3). Increasing the number of options increasedchoice difficulty, which, in turn, made people more likelyto select options that were easier to justify. Further, theeffects were moderated by accessible justifications (exper-iments 4 and 5). Supporting our framework, the effect foundin the first three experiments was reversed when participantshad an accessible justification for choosing indulgences. Un-der these conditions, larger assortments led people to choosehedonic options. Thus, assortment size influences optionchoice through increased difficulty, but the type of optionpeople will choose will depend on accessible justifications.

The findings also cast doubt on a number of alternativeexplanations for our results. Although one could argue thatthe attractiveness of the items in the larger and smaller as-sortments differed, this cannot explain why assortmentwould influence choice even when the same options werelabeled differently (experiment 4) or why manipulating rel-ative effort or licensing would reverse the results (experi-ments 4 and 5). An account based on activated cognitivemind-set also has trouble explaining this moderation and,further, does not suggest why increased assortment sizemight sometimes lead to increased choice of hedonic options(e.g., in the high relative-effort condition in experiment 5).Alternatively, one could argue that increasing the numberof options impairs ones ability to think vividly about anysingle option, thus hurting the attractiveness of hedonic op-tions in particular (which, arguably, depends more on theability to imagine the consumption experience). Again, how-ever, this explanation has difficulty explaining the moder-ation observed in experiments 4 and 5. Assortment sizeshould have the same effect on option vividness regardlessof relative effort or licensing.

An important boundary condition for these findings is theextent to which larger assortments actually increase choicedifficulty. If choosing from larger assortments is not moredifficult (e.g., when one option dominates or when consum-ers have well-defined preferences), then larger assortmentsshould be less likely to influence the type of options con-sumers choose. It is also possible that choice difficulty canbe greater when the number of options is smaller. For ex-ample, increasing the choice set from two to three or moreoptions can give rise to context effects that, in some cases,may facilitate easier and more confident decision making(cf. Dhar and Simonson 2003; Yoon and Simonson 2008).

Assortment, Choice Difficulty, and Justification

While recent work has focused on difficulty and regretas drivers of choice deferral (Iyengar and Lepper 2000), ourfindings take this insight a step further by linking it to earlierwork on justification (Shafir et al. 1993). Most choice-over-load studies examined situations where little justificationexists for choosing one option versus another. Few peoplecan likely come up with a strong reason for choosing milkchocolate praline over fudge delight ice cream, assumingthat the two are equally tempting. Part of the reason thesedifficult situations may be demotivating, then, is the lack ofaccessible reasons or justifications to simplify choice.

In situations in which accessible reasons exist, however,choosing from enlarged assortments may be less likely tolead to deferral. Indeed, research suggests that ideal pointavailability plays an important role in moderating the influ-ence of assortment on consumer preferences (Chernev2003b). Having a clear idea of what you want provides anobvious reason for selecting certain options, but making anyreason accessible, even those that are less related to utilitymaximization, may help consumers deal with enlarged as-sortments. Even seemingly irrelevant information (e.g., stat-ing that the previous consumer chose this option) couldprovide consumers with an accessible reason that may fa-cilitate choice.

Extensions and Practical Implications

Our theoretical framework leads to a number of broaderpredictions that can be tested in future research. Researchon choice overload suggests that enlarged assortments maylead consumers to be less satisfied with the chosen option,in part because of increased regret. The present research,however, suggests that assortment size may also influencesatisfaction with the chosen option by influencing the waythat option is evaluated. Increased set size should lead con-sumers to rely more on justifications for choice; conse-quently, when they do make a choice, their satisfaction withthe chosen option should depend a lot more on how it relatesto the available justifications. For example, participants whochose a more utilitarian option from a larger set might ac-tually be more (or less) satisfied with their choice. Similarly,although the increased assortment led many participants tochoose more utilitarian options, some participants still se-lected more hedonic alternatives. As a result of increasedreliance on reasons for choice, however, these participantsmay actually have been less satisfied with their chosen al-ternative than if they had selected it from a smaller set.Overall, then, the increased justification induced by assort-ment size should influence not only option choice but alsosatisfaction with the chosen option.

This research has focused on the role of assortment sizein option choice, but other factors that influence choice dif-ficulty may also have similar effects on decision making.One example is attributes. Trading off between more attri-butes can sometimes increase decision difficulty and mayinduce greater reliance on justifications for choice. Prelim-inary evidence indicates that presenting consumers with anenlarged list of attributes can influence choice between util-itarian and hedoic options (Sela and Berger 2009), althoughthe exact mechanism and moderating factors underlyingthese processes may differ slightly.

Similarly, although we focused on choices between viceand virtue or hedonic and utilitarian options, future researchmight examine how such justification-based processes in-fluence choice in other important domains. For example,how might a recruiter choosing among job candidates beinfluenced by the number of applications, their lengths, andthe degree of similarity between the individual resumes? Tothe extent that these factors are associated with greater

950 JOURNAL OF CONSUMER RESEARCH

choice difficulty, they are likely to bias choice outcome infavor of choices that are easier to justify. This could poten-tially lead the recruiter to select participants of particularraces, genders, or background characteristics.

Finally, our findings have important managerial and pol-icy implications. For example, manufacturers of healthysnacks would be better off pursuing venues with larger se-lections. Similarly, an award-winning (but nonentertaining)movie might be better off shown in a 21-screen multiplextheater rather than a small two-screen theater when rivalingshoot-em-up action flicks. From a policy perspective, whilerecent assortment research has illustrated how too muchchoice can impair social welfare (see Botti and Iyengar [2006]

for a review), our studies suggest that assortment can alsohave positive welfare effects by driving consumers to makemore virtuous choices. Although choosing from a largermenu may be demotivating, our findings suggest that whenconsumers do choose, the option they select is likely to behealthier. This unanticipated benefit of assortment can po-tentially be used to improve consumer welfarebut withcaution. Giving consumers more options should increasetheir reliance on justifications for choice, but this will onlyimprove their welfare in cases in which those reasons pointthem to better options. Whether the overall combination ofthese factors has positive or negative welfare implicationsremains to be seen.

APPENDIX

EXPERIMENT 4

Set AThese laptops are built for fun. They feature high-end graph-

ics cards, top of the line speakers, and extra music andvideo software.

These laptops are built for work. They are lightweight, have ex-tended battery life, high-end security, and extra business andpresentation software.

Set BThese laptops are built for work. They are lightweight, have

extended battery life, high-end security, and extra businessand presentation software.

These laptops are built for fun. They feature high-end graphicscards, top of the line speakers, and extra music and videosoftware.

Options*Acer Aspire 3690 Toshiba Qosimo Lenovo ThinkPad T60 Acer Aspire 4570 Toshiba Satellite P75 Lenovo ThinkPad T45HP Pavilion 6245 Compaq Evo N400c Fujitsu Lifebook P7630 HP Pavilion 6285c Compaq Presario 725 Fujitsu Lifebook P7230

*Extra options in the larger assortment condition are underlined.

REFERENCES

Baron, Reuben M. and David A. Kenny (1986), The Moderator-Mediator Variable Distinction in Social Psychological Re-search: Conceptual, Strategic and Statistical Considerations,Journal of Personality and Social Psychology, 51 (6), 117382.

Baumol, William J. and Edward A. Ide (1956), Variety in Re-tailing, Management Science, 3 (1), 93101.

Bazerman, Max H., Ann E. Tenbrunsel, and Kimberly Wade-Ben-zoni (1998), Negotiating with Yourself and Losing: MakingDecisions with Competing Internal Preferences, Academy ofManagement Review, 23 (2), 22541.

Berger, Jonah, Michaela Draganska, and Itamar Simonson (2007),The Influence of Product Variety on Brand Perception andChoice, Marketing Science, 26 (4), 46072.

Berlyne, Daniel E. (1960), Conflict, Arousal, and Curiosity, NewYork: McGraw-Hill.

Bohm, Gisela and Hans-Rudiger Pfister (1996), Instrumental orEmotional Evaluations: What Determines Preferences, ActaPsychologica, 93 (13), 13548.

Botti, Simona and Sheena S. Iyengar (2006), The Dark Side ofChoice: When Choice Impairs Social Welfare, Journal ofPublic Policy and Marketing, 25 (1), 2438.

Chernev, Alexander (2003a), Product Assortment and IndividualDecision Processes, Journal of Personality and Social Psy-chology, 85 (1), 15162.

(2003b), When More Is Less and Less Is More: The Roleof Ideal Point Availability and Assortment in ConsumerChoice, Journal of Consumer Research, 30 (September),17083.

Dhar, Ravi (1997), Consumer Preference for a No-Choice OptionSource, Journal of Consumer Research, 24 (September),21531.

Dhar, Ravi and Itamar Simonson (1999), Making ComplementaryChoices in Consumption Episodes: Highlighting versus Bal-ancing, Journal of Marketing Research, 36 (February), 2944.

(2003), The Effect of Forced Choice on Choice, Journalof Marketing Research, 40 (May), 14660.

Hsee, Christopher K. (1995), Elastic Justification: How Temptingbut Task-Irrelevant Factors Influence Decisions, Organiza-tional Behavior and Human Decision Process, 62 (3), 33037.

Hsee, Christopher K. and Yuval Rottenstreich (2004), Music, Pan-das, and Muggers: On the Affective Psychology of Value,Journal of Experimental Psychology, 133 (1), 2330.

Iyengar, Sheena S., Gur Huberman, and Wei Jiang (2004), HowMuch Choice Is Too Much? Contributions to 401(k) Retire-ment Plans, in Pension Design and Structure: New Lessonsfrom Behavioral Finance, ed. Olivia S. Mitchell and StephenP. Utkus, Oxford: Oxford University Press, 8395.

Iyengar, Sheena S. and Mark R. Lepper (2000), When ChoiceIs Demotivating: Can One Desire Too Much of a GoodThing? Journal of Personality and Social Psychology, 79(6), 9951006.

Kahn, Barbara E. and Rebecca Ratner (2005), Variety for theSake of Variety? Diversification Motives in ConsumerChoice, in Inside Consumption: Frontiers of Research onConsumer Motives, Goals, and Desires, ed. Srinivasan Rat-neshwar and David Glen Mick, London: Routledge, 10221.

HOW ASSORTMENT SIZE INFLUENCES OPTION CHOICE 951

Khan, Uzma and Ravi Dhar (2006), Licensing Effect in ConsumerChoice, Journal of Marketing Research, 43 (May), 25966.

Kivetz, Ran (1999), Advances in Research on Mental Accountingand Reason-Based Choice, Marketing Letters, 10 (3),24966.

Kivetz, Ran and Anat Keinan (2006), Repenting Hyperopia: AnAnalysis of Self-Control Regrets, Journal of Consumer Re-search, 33 (September), 27382.

Kivetz, Ran and Itamar Simonson (2002a), Earning the Right toIndulge: Effort as a Determinant of Customer Preferencestoward Frequency Program Rewards, Journal of MarketingResearch, 39 (May), 15570.

(2002b), Self-Control for the Righteous: Toward a Theoryof Precommitment to Indulgence, Journal of Consumer Re-search, 29 (September), 199217.

Kivetz, Ran and Yuhuang Zheng (2006), Determinants of Justi-fication and Self-Control, Journal of Experimental Psychol-ogy: General, 135 (4), 57287.

Lancaster, Kelvin J. (1990), The Economics of Product Variety:A Survey, Marketing Science, 9 (3), 189206.

Langer, Ellen J. and Judith Rodin (1976), The Effects of Choiceand Enhanced Personal Responsibility for the Aged: A FieldExperiment in an Institutional Setting, Journal of Personalityand Social Psychology, 34 (2), 19198.

Lascu, Dana N. (1991), Consumer Guilt: Examining the Potentialof a New Marketing Construct, in Advances in ConsumerResearch, Vol. 18, ed. Rebecca Holman and Michael Solo-mon, Ann Arbor, MI: Association for Consumer Research,29095.

OCurry, Suzanne and Michal Strahilevitz (2001), Probability andMode of Acquisition Effects on Choices between Hedonicand Utilitarian Options, Marketing Letters, 12 (1), 3749.

Prelec, Drazen and George Loewenstein (1998), The Red and theBlack: Mental Accounting of Savings and Debt, MarketingScience, 17 (1), 428.

Schwartz, Barry (2004), The Paradox of Choice: Why More IsLess, New York: Ecco.

Sela, Aner and Jonah Berger (2009), The Dual Effects of ProductAttributes on Choice: Creating Difficulty or Providing Ex-cuses, working paper, Marketing Department, Stanford Grad-uate School of Business, Stanford, CA.

Shafir, Eldar, Itamar Simonson, and Amos Tversky (1993), Rea-son-Based Choice, Cognition, 49 (12), 1136.

Simonson, Itamar (1989), Choice Based on Reasons: The Caseof Attraction and Compromise Effects Source, Journal ofConsumer Research, 16 (September), 15874.

(1992), The Influence of Anticipating Regret and Re-sponsibility on Purchase Decisions, Journal of ConsumerResearch, 19 (June), 10518.

Simonson, Itamar and Stephen M. Nowlis (2000), The Role ofExplanations and Need for Uniqueness in Consumer DecisionMaking: Unconventional Choices Based on Reasons, Jour-nal of Consumer Research, 27 (June), 4968.

Small, Deborah, George Loewenstein, and Paul Slovic (2007),Sympathy and Callousness: The Impact of DeliberativeThought on Donations to Identifiable and Statistical Victims,Organizational Behavior and Human Decision Processes, 102(2), 14353.

Sobel, Michael E. (1982), Asymptotic Confidence Intervals forIndirect Effects in Structural Equation Models, SociologicalMethodology, 13, 290312.

Strahilevitz, Michal and John G. Myers (1998), Donations toCharity as Purchase Incentives: How Well They Work MayDepend on What You Are Trying to Sell, Journal of Con-sumer Research, 24 (March), 43446.

Tversky, Amos and Eldar Shafir (1992), Choice under Conflict:The Dynamics of Deferred Decision, Psychological Science,3 (6), 35861.

Weber, Max (1958), The Protestant Ethic and the Spirit of Cap-italism, New York: Scribner.

Yoon, Song-Oh and Itamar Simonson (2008), Choice Set Con-figuration as a Determinant of Preference Attribution andStrength, Journal of Consumer Research, 35 (August),32436.