Each year is replete with occasions to give gifts. From

Transcription

1 Journal of Consumer Researc Advance Access publised December 29, 2016 Experiential Gifts Foster Stronger Social Relationsips tan Material Gifts CINDY CHAN CASSIE MOGILNER Eac year is replete wit occasions to give gifts. From birtdays to religious olidays, Valentine s Day to Fater s Day, tese occasions are fraugt wit te question: Wat to give?! Sould you give your dad a designer tie or golf lessons? Would giving your spouse a watc or concert tickets spark greater affection? Would a set of wine glasses or a wine tasting better cement your friendsip wit your favorite colleague? And, ultimately, wy Cindy Can (cindy.can@utoronto.ca) is an assistant professor of marketing at te UTSC Department of Management and te Rotman Scool of Management, University of Toronto, Toronto, ON M1C 1A4. Cassie Mogilner (cassie.olmes@anderson.ucla.edu) is an associate professor of marketing at te UCLA Anderson Scool of Management, University of California, Los Angeles, CA Tis article is based on a portion of te first autor s dissertation. Te autors greatly appreciate te support and facilities provided by te Warton Beavioral Lab and te financial support provided by Warton s Dean s Researc Fund, as well as te researc assistance provided by Cristina Gizzo and Karisma Antia. Te autors are also grateful to ave enjoyed te connecting power of experiential gifts in teir own relationsips, particularly wit teir usbands, Tom and Rob. Ann McGill and Darren Dal served as editors for tis article. Darren Dal also served as associate editor. Advance Access publication December 11, 2016 Interpersonal relationsips are essential to well-being, and gifts are often given to cultivate tese relationsips. To inform gift givers of wat to give and to gain insigt into te connecting function of gifts, tis researc investigates wat type of gift is better at strengtening relationsips according to gift recipients material gifts (objects for recipients to keep) or experiential gifts (events for recipients to live troug). Experiments examining actual gift excanges in real-life relationsips reveal tat experiential gifts produce greater improvements in relationsip strengt tan material gifts, regardless of weter te gift giver and recipient consume te gift togeter. Te relationsip improvements tat recipients derive from experiential gifts stem from te intensity of emotion tat is evoked wen tey consume te gifts, rater tan wen te gifts are received. Giving experiential gifts is tus identified as a igly effective form of prosocial spending. Keywords: gift giving, experiential purcases, material purcases, emotion, relationsips would one of tese gifts strengten te relationsip more tan te oter? Wit te average American ouseold spending almost 2% of its annual income on gifts (US Bureau of Labor Statistics 2013), and wit gift-giving occasions serving as great opportunities (and liabilities) for relationsip building, tese are surprisingly consequential decisions. Indeed, interpersonal relationsips are essential to well-being (Baumeister and Leary 1995; Clark and Lemay 2010; Mogilner 2010; Reis, Collins, and Bersceid 2000), and gifts serve as a means to foster tese important connections (Algoe, Haidt, and Gable 2008; Dunn et al. 2008b; Rut, Otnes, and Brunel 1999; Serry 1983). To elp inform gift givers of wat to give and to gain insigt into te interpersonal benefits of gifts, tis researc adopts te gift recipients perspective and experimentally investigates wic type of gift is more effective at strengtening teir relationsip wit teir gift giver material gifts (objects for te recipients to keep) or experiential gifts (events for te recipients to live troug). And wy? MATERIAL VERSUS EXPERIENTIAL GIFTS Borrowing Van Boven and Gilovic s (2003) definition of material and experiential purcases, we define material VC Te Autor Publised by Oxford University Press on bealf of Journal of Consumer Researc, Inc. All rigts reserved. For permissions, please journals.permissions@oup.com Vol. 00 November 2016 DOI: /jcr/ucw067 1

2 2 JOURNAL OF CONSUMER RESEARCH gifts as objects to be kept in te recipient s possession (e.g., jewelry or electronic gadgets) and experiential gifts as an event tat te recipient lives troug (e.g., concert tickets or a potograpy lesson). Te researc comparing material and experiential purcases to date as focused on te effects of making tese purcases for oneself, finding tat buying an experience is typically more personally beneficial tan buying a material good (Gilovic, Kumar, and Jampol 2015). Compared to possessions, experiences lead to greater satisfaction (Carter and Gilovic 2010), less regret (Rosenzweig and Gilovic 2012), and greater appiness (Van Boven and Gilovic 2003), especially wen te outcome of te experience is positive (Nicolao, Irwin, and Goodman 2009). Te benefits of acquiring an experience over a possession stem from te fact tat experiences are more likely to be sared wit oters (Caprariello and Reis 2013), contribute more to one s sense of self (Carter and Gilovic 2012), are more unique (Rosenzweig and Gilovic 2012), and are arder to compare against alternatives (Carter and Gilovic 2010). Altoug prior researc offers guidance on weter to buy experiences or material goods to improve one s own wellbeing, te question of wat to buy to strengten one s relationsips wit oters remains unanswered. Would giving someting to do or someting to keep forge a stronger social bond? It turns out tat people are more inclined to give material gifts. In a survey we conducted among 219 gift givers (66% female; ages 18 74, M ¼ 34.68), 78% reported aving most recently given a gift tat was material. Tis tendency is consistent wit te argument tat giving a gift tat is durable will leave a lasting impression (Ariely 2011). A pilot study we conducted around Fater s Day, owever, ints tat tis tendency to give material gifts migt be misguided. Recipients of Fater s Day gifts (N ¼ 42; ages 48 75; M ¼ 55.05) participated in a two-part survey: one completed te week before Fater s Day and one te week after. Bot before and after Fater s Day, dads rated te strengt of teir relationsip wit teir cild (1 ¼ feel extremely distant and disconnected, 9 ¼ feel extremely close and connected); any cange would reflect te impact of receiving te gift on te relationsip. Dads also rated to wat extent te gift tey received was material (1 ¼ not at all, 7 ¼ completely) and experiential (1 ¼ not at all, 7 ¼ completely). A multiple regression analysis predicting cange in relationsip strengt sowed tat gifts tat were more experiential strengtened dads relationsips wit teir cildren (b ¼ 0.16, SE ¼ 0.07, t(39) ¼ 2.21, p ¼.03, d ¼.71), wereas te material nature of te gift did not (b ¼ 0.03, SE ¼ 0.07, t(39) ¼ 0.39, p ¼.70, d ¼.12). It was not tat experiential gifts were more likely to be given in initially stronger relationsips, since te material and experiential gift ratings were unrelated to relationsip strengt before Fater s Day (p s >.43). Tese results were corroborated by a second pilot study conducted following Moter s Day among moms wo ad received a gift from teir cild (N ¼ 99; ages 38 64, M ¼ 51.9; 11 unspecified). In tis study, moms first reported te relational impact of teir gift on a subjective cange scale (1 ¼ felt more distant and less connected, 9 ¼ felt closer and more connected), and ten rated te experiential versus material nature of te gift on a bipolar scale (1 ¼ purely material, 9 ¼ purely experiential; Van Boven and Gilovic 2003). Like dads, moms wo received gifts tat were more experiential reported aving a stronger relationsip wit teir cild as a result of receiving te gift (b ¼ 0.21, SE ¼ 0.07, t(97) ¼ 2.96, p ¼.004, d ¼.60). Altoug tese results are correlational and based on small samples, tey provide preliminary evidence to suggest tat experiential gifts are more effective tan material gifts at strengtening relationsips between gift recipients and teir gift givers. RELATIONSHIPS AND THE ROLE OF EMOTION Recent experimental researc on gift giving as made great strides in understanding ow recipients evaluate different types of gifts (Flynn and Adams 2009; Gino and Flynn 2011; Waldfogel 1993; Zang and Epley 2012); owever, less is known about ow giver-recipient relationsips are best cultivated troug different types of gifts (Aknin and Human 2015). Tat is, muc of te work on gift giving as focused on ow muc recipients appreciate, value, or like particular gifts, rater tan te impact of tese gifts on te relationsip. For instance, prior giftgiving experiments ave sown tat despite gift givers beliefs tat expensive gifts will be more appreciated, recipients appreciate expensive and inexpensive gifts alike (Flynn and Adams 2009) and put a lower monetary value on a gift tan its actual cost (Waldfogel 1993). And altoug gift givers tink tat unsolicited gifts convey greater tougtfulness and serve as a stronger signal of relationsip value, recipients prefer receiving cas or gifts tat tey ad explicitly requested (Gino and Flynn 2011; Ward and Broniarczyk 2015). Additionally, wen buying for multiple recipients, gift givers select overly individuated gifts in an attempt to be tougtful and understanding of eac unique recipient, but recipients unfortunately tend not to recognize te tougt put into gifts tey like (Zang and Epley 2012), and gift givers efforts to convey tougtfulness can even result in selecting gifts tat recipients like less (Steffel and LeBoeuf 2014). Even te most well-intentioned gifts can go awry, as givers also tend to mispredict ow muc recipients will appreciate socially responsible gifts, suc as caritable donations given on teir bealf (Cavanaug, Gino, and Fitzsimons 2015). In ligt of tese findings tat gift givers are poor predictors of wat recipients will like, it is fortunate tat recipients

3 CHAN AND MOGILNER 3 can regift teir gifts witout offending te giver (Adams, Flynn, and Norton 2012)! Our researc adopts a different approac to assess a gift s value. Namely, we measure te gift s influence on relationsip strengt from te recipient s perspective, rater tan ow muc te recipient likes te gift. We looked to te literature on close relationsips to define relationsip strengt. Altoug tere is substantial variation among relationsip types (i.e., friendsips, romantic partners, and family members) wit respect to wat constitutes a strong relationsip, tere are principle indicators of relationsip strengt tat span across personal relationsips: te extent to wic partners feel close to eac oter (Algoe et al. 2008; Aron, Aron, and Smollan 1992; Dibble, Levine, and Park 2012; Kok et al. 2013; Kok and Fredrickson 2010) and connected to eac oter (Algoe et al. 2008; Dibble et al. 2012; Hutcerson, Seppala, and Gross 2008), as well as ow satisfied tey are wit teir relationsip (Rusbult, Martz, and Agnew 1998). Across our studies, we adopt tese indicators of relationsip strengt to measure ow muc te gift canges te gift recipient s perception of te strengt of is or er relationsip wit te gift giver from before to after receiving te gift. Tis perspective on te success of a gift is similar to tat taken in earlier qualitative researc exploring ow gift excanges produce relationsip cange. A series of dept interviews and surveys offer ric insigts into ow te context, rituals, meaning, and emotions tat surround a gift excange can lead to different relational outcomes ranging from relationsip strengtening to rare cases of relationsip severing (Rut et al. 1999; Rut, Brunel, and Otnes 2004). For instance, Rut et al. (1999) observed tat gift excanges tat involve igly personalized rituals can imbue te gift wit sared meaning and often lead to relationsip strengtening. More recently, experimental work as identified tat gifts reflecting te giver can promote relationsip closeness (Aknin and Human 2015). Te current work builds on tese insigts by specifically testing te relational impact of particular types of gifts tose tat are material versus experiential. It furter examines wy experiential and material gifts may differ in teir ability to forge a stronger relationsip between gift recipients and givers. A distinction between experiential and material purcases tat as yet to be explored is ow muc emotion tey evoke during consumption. Prior researc as focused on te appiness elicited by experiences and material possessions (Van Boven and Gilovic 2003; Weidman and Dunn 2016), but it is important to note tat experiences can stimulate a wide range of emotions (Battacarjee and Mogilner 2014; Can et al. 2014; Derbaix and Pam 1991; Halvena and Holbrook 1986; Mogilner, Aaker, and Kamvar 2012; Mogilner, Kamvar, and Aaker 2011; Ricins 1997). For instance, a safari adventure can elicit feelings of awe and fear; a rock concert can fuel excitement; a spa package can promote calmness; and an opera may move one to tears. Given te diversity of discrete emotions tat consuming bot experiences and material goods can evoke, we focus our investigation on te overall intensity of emotion felt during gift consumption, and propose tat te emotion felt by recipients wen consuming an experiential gift will be more intense tan wen consuming a material gift. Researc on relationsips igligts emotion as a key feature in relationsip development and maintenance. Emotions expressed and experienced witin te context of a relationsip yield positive interpersonal effects (Clark and Finkel 2004; Graam et al. 2008; Kubacka et al. 2011; Laurenceau, Barrett, and Pietromonaco 1998; Slatcer and Pennebaker 2006). For instance, disclosing one s emotions (vs. facts and information) makes people feel closer (Laurenceau et al. 1998), positive emotions suc as gratitude promote relationsip maintenance beaviors (Kubacka et al. 2011), and saring negative emotions promotes bonding (Graam et al. 2008). Te relationsipstrengtening effect of emotions extends to situations in wic te emotions are not sared. Prior researc as sown tat partners wo write about teir feelings witin a relationsip are more likely to stay togeter, even wen tey don t sare wat tey wrote (Slatcer and Pennebaker 2006). More generally, greater emotional intensity as been found to reduce perceived psycological distance (Van Boven et al. 2010). Taking tese findings togeter, we assert tat a gift tat evokes greater emotion sould be more effective at strengtening relationsips tan a gift tat elicits a weaker emotional response, and tus experiential gifts sould foster stronger relationsips tan material gifts. Furtermore, in te case of gifts, te mere fact tat te experience was given by te relationsip partner psycologically places te experience and te resulting emotion witin te context of te relationsip. We terefore propose tat te experiential gift need not be sared between te gift giver and recipient for it to evoke greater emotion and tereby improve te relationsip. Notably, te emotion evoked during te gift consumption is distinct from te emotion evoked during te gift excange. In is teoretical model delineating te impact of gifts on relationsips, Serry (1983) igligts te importance of focusing beyond te gift excange to te consumption of te gift, during wic te gift becomes te veicle by wic te relationsip of te donor and te recipient is realigned (165). Indeed, it is te emotion evoked wile te recipient is consuming te gift tat we propose drives te difference between experiential and material gifts on relationsip cange. Wereas material and experiential gifts are bot likely to elicit emotion during a gift excange (e.g., a recipient could feel grateful weter given a wallet or tickets to a comedy sow), experiential gifts sould elicit more intense emotion during gift consumption as te recipient lives troug an event (e.g., a recipient may feel

4 4 JOURNAL OF CONSUMER RESEARCH mild enjoyment wile using a wallet, yet will likely feel intensely amused and deligted wile attending a comedy sow; Weidman and Dunn 2016). Additionally, altoug Rut et al. (1999, 2004) found tat te valence of te emotion during a gift excange mattered more tan te intensity of emotion in predicting canges in te relationsip (peraps because te gift giver is likely te source and target of te emotions evoked during a gift excange), we propose tat it is te intensity of emotion evoked during gift consumption tat is responsible for te greater power of experiential gifts to strengten relationsips. Altogeter, we predict tat experiential gifts will improve relationsips more tan material gifts, and tat tis is driven by te greater emotional intensity evoked from consuming an experience rater tan a possession. More formally, we predict: H1: From te recipient s perspective, experiential gifts strengten relationsips more tan material gifts, irrespective of weter te gift is consumed wit te gift giver. H2: Consuming experiential gifts evokes more intense emotion tan consuming material gifts, and tis greater emotionality drives te effect of gift type on cange in relationsip strengt. To test tese ypoteses, we conducted experiments involving actual gift excanges in te context of existing personal relationsips. In study 1, gift givers were provided wit $15 to purcase eiter an experiential or material gift to give to someone tey know; in study 2, gift givers were provided wit a coffee mug, framed as experiential or not, to give to someone tey know; and in studies 3 and 4, participants recalled experiential or material gifts tey ad received from someone tey know. Across te studies, te experiential versus material nature of te gift was manipulated to test ow gift type influenced relationsip strengt from te recipient s perspective. To examine te underlying role of emotion, study 3 measured and study 4 manipulated te emotion evoked during gift consumption. Togeter, tese studies seek to contribute a better understanding of ow te type of gift received can differentially affect relationsips. Across all four studies, sample size was determined prior to eac study wit an effort to collect as many participants as resources would permit in te allotted timeframe, and all data exclusions and experimental conditions are reported. STUDY 1: A $15 GIFT Study 1 used a two-part design measuring gift recipients reports of pre- and post-gift relationsip strengt to test our primary ypotesis tat experiential gifts strengten relationsips more tan material gifts. Participants were recruited wit a friend, and in eac participant pair, one was randomly assigned to be te gift giver and te oter to be te gift recipient. Gift givers were provided wit $15 and instructed to purcase eiter an experiential gift or material gift for teir friend, wic tey were not to consume wit teir friend. We specifically examined gifts tat were not consumed togeter to counter te explanation tat experiential gifts strengten relationsips solely because saring in te experience involves te giver and recipient spending more time togeter (Battacarjee and Mogilner 2014; Hersfield, Mogilner, and Barnea 2016; Mogilner and Aaker 2009). Metod Fifty-nine pairs of friends (118 participants; 57% female, 1% unspecified; ages 18 27; M ¼ 20.63) were recruited troug a university laboratory to participate in a gift-giving study in excange for $10. Upon arriving to te laboratory, participants in eac friend pair were randomly assigned to te role of gift giver or recipient. Gift givers were provided wit an additional $15 along wit instructions for ow to spend tis money. Gift Type Manipulation. Gift givers were randomly assigned to purcase eiter an experiential or material gift for teir friend using definitions adapted from Van Boven and Gilovic (2003). Gift givers in te experiential gift condition were instructed, Purcase a gift tat is an experience...experiential gifts are experiences intended for te recipient to do or live troug. Gift givers in te material condition were instructed, Purcase a gift tat is a material good... Material gifts are tangible items for te recipient to ave and keep for im/erself. All gift givers were furter instructed to give a gift tat teir friend would consume witout tem and witin te next week, to spend as close to te $15 as possible on te gift, to give teir friend te gift witin te next tree days, and not to tell teir friend our instructions regarding te type of gift tey were to purcase. Gift givers left te laboratory wit te $15 and a printout of te gift instructions corresponding to teir assigned condition. Cange in Relationsip Strengt. To serve as te baseline measure of relationsip strengt, gift recipients rated teir relationsip wit teir friend on four items. Te first measure was te inclusion of oter in self (IOS) scale adapted from Aron et al. (1992). Prior researc as effectively visually portrayed and measured te sense of interconnection central to relationsip strengt troug te degree of overlap between two circles tat represent eac partner s self-concept (Aknin and Human 2015; Aron et al. 1991; Aron et al. 1992; Brown et al. 2009). We terefore presented gift recipients wit a set of nine circle pairs, in wic one of te circles was labeled self and te oter circle was labeled oter. Tese pairs ranged in teir degree of overlap to represent te strengt of te recipient s relationsip wit te gift giver. Gift recipients were asked

5 CHAN AND MOGILNER 5 to coose te set of circles tat best described teir relationsip wit teir gift giver (see appendix A). Next, recipients were asked to rate teir relationsip wit teir friend on tree Likert scales measuring closeness (1 ¼ extremely distant, 9 ¼ extremely close), connection (1 ¼ extremely disconnected, 9 ¼ extremely connected), and relationsip strengt (1 ¼ extremely weak, 9 ¼ extremely strong). Te average of tese four items served as our measure for pre-gift relationsip strengt (a ¼.84). Recipients were ten told tat tey would be receiving a gift from teir friend witin te next tree days and a link to an online follow-up survey from us in one week. Tey were instructed to consume te gift tey receive once witin te next week (before completing te follow-up survey), and to not consume te gift wit teir friend. Recipients left te laboratory wit a printout of teir gift instructions. One week later, gift recipients received an inviting tem to complete te online follow-up survey in excange for a $5 Amazon.com gift card. Forty-four gift recipients responded (n experiential ¼ 20, n material ¼ 24; 57% female; ages 18 25, M ¼ 20.5). After describing te gift tey ad received, recipients reported te strengt of teir relationsip wit teir friend using te same four items as before. Tese responses were averaged to serve as te post-gift relationsip strengt measure (a ¼.93). Te post-gift relationsip strengt score was subtracted from te pre-gift score for our measure of cange in relationsip strengt. One extreme outlier was excluded from furter analyses (greater tan tree standard deviations from te mean, studentized residual ¼ 4.72, Cook s D ¼ 0.59). Tougtfulness and Liking. Because muc of te experimental researc on gift giving as focused on ow muc recipients like te gift and ow tougtful tey perceive te gift to be (Flynn and Adams 2009; Gino and Flynn 2011), we also measured tougtfulness and liking to assess weter material and experiential gifts differ on tese dimensions. Recipients rated te tougtfulness of teir gift on four items adapted from Flynn and Adams (2009) and Gino and Flynn (2011): te extent to wic te gift was tougtful, considerate, took teir needs into account, and took wat tey really wanted into account (1 ¼ not at all, 7 ¼ to a great extent; a ¼.86). Recipients rated ow muc tey liked te gift on tree items: ow muc tey liked te gift, ow satisfied tey were wit te gift, and cost aside, ow desirable te gift would be to an average oter person (tird item adapted from Rosenzweig and Gilovic 2012; 1¼ not at all, 7 ¼ to a great extent; a ¼.85). Manipulation Cecks. As a ceck for weter gift givers ad followed teir instructions, we asked recipients to 1) rate to wat extent te gift tey received was material or experiential (1 ¼ purely material, 5 ¼ equally material and experiential, 9 ¼ purely experiential), 2) report weter tey ad sared in te consumption of teir gift wit teir gift giver, and 3) estimate te price of te gift. Participants also reported ow muc time tey ad spent wit teir gift giver during te gift excange and ow muc time tey ad spent consuming te gift. Results and Discussion Gifts Received. Experiential gifts included a pass to a barre class and movie tickets. Material gifts included a sirt, a poster, and a wine aerator. Te manipulation ceck confirmed tat recipients in te experiential gift condition received gifts tat were more experiential (M ¼ 4.89, SD ¼ 2.38) tan tose in te material gift condition (M ¼ 3.17, SD ¼ 2.24; t(41) ¼ 2.45, p ¼.02, d ¼.71). Also, te majority of recipients (86%) ad not consumed teir gift wit teir gift giver, tere was no significant difference in estimated price between recipients of experiential gifts (M ¼ $14.01, SD ¼ 4.19) and material gifts (M ¼ $13.10, SD ¼ 5.53; t(41) ¼ 0.59, p ¼.56, d ¼.18), and tere were no significant differences in ow muc time recipients ad spent wit teir gift giver during te gift excange (p >.99) or ow muc time tey ad spent consuming teir gift (p ¼.17). Cange in Relationsip Strengt. Te pre-gift relationsip measures confirmed tat tere were no differences in baseline levels of relationsip strengt among participants in te experiential condition (M ¼ 6.71, SD ¼ 2.12) and material condition (M ¼ 7.10, SD ¼ 2.12; t(41) ¼ 0.95, p ¼.35, d ¼.29). In support of our first ypotesis, te relationsip cange measure revealed tat recipients of an experiential gift (M ¼ 0.08, SD ¼ 0.79) sowed a more positive cange in relationsip strengt tan recipients of a material gift (M ¼ 0.54, SD ¼ 1.10; t(41) ¼ 2.06, p ¼.05, d ¼.61). Because some participants rated teir pre-gift relationsip using te extreme ends of te scales, we conducted a robustness ceck by trimming te data of any participants wo reported a pre-gift relationsip score greater tan 8 (n ¼ 11) or less tan 2 (n ¼ 0). Omitting tese participants strengtened te effect of gift type on relationsip cange, wit recipients of experiential gifts reporting greater relationsip improvements tan recipients of material gifts (t(32) ¼ 2.41, p ¼.02, d ¼.78). Tougtfulness and Liking. Altoug experiential gifts were directionally perceived as more tougtful and better liked, te effects were not significant (tougtfulness: M exp ¼ 5.49, SD ¼ 2.32 vs. M mat ¼ 5.07, SD ¼ 2.32; t(41) ¼ 0.94, p ¼.35, d ¼.29; liking: M exp ¼ 5.68, SD ¼ 0.96 vs. M mat ¼ 5.07, SD ¼ 1.43; t(41) ¼ 1.61, p ¼.12, d ¼.43). Terefore, te ability of experiential gifts to strengten relationsips does not appear to be driven by perceived tougtfulness or liking.

6 6 JOURNAL OF CONSUMER RESEARCH Te results of study 1 sowed tat receiving an experiential gift improves te strengt of recipients relationsips wit teir gift giver, compared to receiving a material gift. Study 2 tests weter igligting te experiential aspect of a material gift can similarly improve relationsip strengt. STUDY 2: A COFFEE MUG GIFT Across a variety of gifts tat were individually selected by eac gift giver, study 1 demonstrated te relationsipstrengtening benefits of experiential gifts over material gifts. Tis next study provided an even more rigorous test for te connecting power of experiential gifts by olding te gift itself constant and varying only te experiential versus material framing of tat gift. Indeed, many material gifts ave experiential components. For example, a stereo is a material object tat is kept in one s possession for years, yet it also provides te experience of listening to music. Similarly, a bottle of wine as a tangible, pysical presence tat can contribute to a collection, but it can also provide a very pleasurable multisensory experience wen consumed wit a perfectly paired ceese. Study 2 took advantage of te malleable distinction between material and experiential gifts and tested weter a material gift (a coffee mug) could be framed as more experiential (by igligting te experience of drinking coffee) to furter strengten gift givers and recipients relationsips. Metod Gift Type Manipulation. Two undred gift givers were recruited troug a university laboratory (57% female; ages 18 39, M ¼ 20.6) and provided wit a gift-wrapped coffee mug to give as a gift to someone tey know. Participants were randomly assigned to eiter give a mug tat igligted te experience of drinking coffee (wit te words my coffee time inscribed on it) or give a mug identified as a material possession (wit te words my coffee mug inscribed on it; see appendix B). A betweensubjects pre-test validated te manipulation: participants (N ¼ 68; 56% female; ages 18 29, M ¼ 20.94) were presented wit one of te two mugs and asked to rate te mug on a nine-point scale (1 ¼ purely material, 9 ¼ purely experiential). Participants rated te my coffee time mug as more experiential (M ¼ 3.69, SD ¼ 2.20) tan te my coffee mug mug (M ¼ 2.63, SD ¼ 1.83; t(67) ¼ 2.13, p ¼.04, d ¼.50). Te mugs did not differ in rated desirability, positivity, or favorability (a ¼.90; t(67) ¼ 0.06, p ¼.95, d ¼.01). Cange in Relationsip Strengt. A survey link was provided on a voucer tat was inside te gift-wrapped coffee mug. Te voucer was for $5 at a local coffee sop, and it would become valid if te gift recipient completed a brief online survey. Eac mug condition ad a separate survey link, allowing us to know te type of mug received. One undred nine recipients completed te survey (64% female; ages 16 58, M ¼ 21.5; n material ¼ 64; n experiential ¼ 45). Te survey asked gift recipients to rate ow receiving te gift canged te strengt of teir relationsip wit te person wo gave tem te gift (1 ¼ felt more disconnected, 9 ¼ felt more connected), and ow muc tey liked te gift (1 ¼ ate it, 9 ¼ love it). Results and Discussion Recipients of te more experiential gift (M ¼ 7.47, SD ¼ 1.50) reported greater strengtening of teir relationsip wit teir gift giver tan recipients of te more material gift (M ¼ 6.92, SD ¼ 1.34; t(107) ¼ 1.99, p ¼.05, d ¼.37). Again, tis effect appears to be independent of ow muc recipients liked te gift, because recipients reported no difference in ow muc tey liked teir mug (M exp ¼ 7.33, SD ¼ 1.41 vs. M mat ¼ 7.25, SD ¼ 1.50; t(107) ¼ 0.29, p ¼.77, d ¼.06). Study 2 provided additional support for our main ypotesis (ypotesis 1) using a igly conservative and controlled test for te effect of gift type on cange in relationsip strengt. Holding all features of te gift constant except for te extent to wic te giving of an experience was igligted, tis study sowed tat receiving a more experiential gift is better at strengtening relationsips tan receiving a more material gift. Indeed, even a material gift (a coffee mug) could be made more connecting by reminding te recipient of te experience it offers (te time spent drinking coffee). Many gifts ave bot experiential and material elements, and tese results demonstrate tat gift givers can enjoy some of te relational benefit of experiential gifts by merely igligting te experience te gift provides. Te next study explored a mecanism for te effect, testing te underlying role of emotion from gift consumption. STUDY 3: THE MEDIATING ROLE OF EMOTION FROM GIFT CONSUMPTION Studies 1 and 2 provided evidence for experiential gifts being more effective at strengtening relationsips tan material gifts. In tis next study, we explored te underlying role of emotion. Participants in study 3 were asked to recall eiter an experiential or material gift tey ad received and ten to rate ow te gift impacted teir relationsip wit te gift giver. To examine te mecanism, tis study additionally measured te intensity of te emotion evoked from gift consumption separately from te emotion evoked from te gift excange. Qualitative researc observed tat a gift excange can be igly emotional, and te combination of negative and positive emotions felt during a gift excange, as well as te AQ1

7 CHAN AND MOGILNER 7 recipient s reaction to te emotions expressed by te gift giver, contribute to relationsip cange (Belk and Coon 1993; Rut et al. 1999, 2004). We predict tat wile a gift excange can be igly emotional for bot material and experiential gifts, consuming an experiential gift will elicit a more intense emotional response tan consuming a material gift (Weidman and Dunn 2016). For example, attending a teatre performance or going on a vacation is likely to be more emotional tan wearing a new pair of boots or driving a car. Furtermore, it is tis emotion evoked from consuming experiential gifts tat we propose is responsible for teir positive impact on relationsip strengt (ypotesis 2). Toug study 1 sowed te more positive effect of receiving an experiential gift versus a material gift even wen participants were instructed to not consume te gift togeter, tis study furter examined te role of saring te gift troug a 2 (gift type: material vs. experiential) 2 (consumption: sared vs. nonsared) between-subjects design. Metod Gift Type Manipulation. Six undred adults were recruited troug Amazon Mecanical Turk to participate in tis study in excange for $0.75. Participants were randomly assigned to recall a particular type of gift tey ad received: sared experiential gift, nonsared experiential gift, sared material gift, or nonsared material gift. Participants in te experiential gift conditions were instructed, Please recall and describe an experiential gift tat you ave received at some point in your life from anoter person. Participants in te material gift conditions were instructed, Please recall and describe a material gift tat you ave received at some point in your life from anoter person. Participants were provided wit a definition of material or experiential gifts adapted from Van Boven and Gilovic (2003). Tose in te sared consumption conditions were furter instructed, Tis sould be [a material/ an experiential] gift tat you consumed wit te person wo gave it to you (i.e., you sared te gift wit your gift giver). Tose in te nonsared consumption conditions were furter instructed, Tis sould be [a material/an experiential] gift tat you consumed on your own (i.e., you did not sare te gift wit your gift giver). Participants wo could not recall a gift (n ¼ 7), did not complete te survey (n ¼ 41), or did not follow te gift recall instructions (i.e., described a gift tey ad given, n ¼ 1; described a gift received from multiple people, n ¼ 15; described a gift of cas, n ¼ 1; described multiple gifts, n ¼ 1) were eliminated from te analysis. Tis left 534 gift recipients in te analyzed dataset (59% female; ages 18 78, M ¼ 33.1). Cange in Relationsip Strengt. Using te measures from study 1, participants rated te strengt of teir relationsip to te gift giver before and after receiving te gift. Participants first cose two pairs of overlapping circles: one to represent teir relationsip before receiving te gift and one to represent teir relationsip after receiving te gift (see appendix A; adapted from Aron et al. 1992). Participants also rated teir relationsip bot before (a ¼.92) and after (a ¼.91) receiving te gift in terms of closeness (1 ¼ extremely distant, 9 ¼ extremely close), connection (1 ¼ extremely disconnected, 9 ¼ extremely connected), and relationsip strengt (1 ¼ extremely weak, 9 ¼ extremely strong). Te difference between eac of te before and after ratings on te four relationsip measures was calculated, and tese values were averaged to form an overall indicator of cange in relationsip strengt. Emotion. Recipients reported ow emotional tey felt from te gift excange separately from ow emotional tey felt during gift consumption. Tey were specifically instructed, Tink about te emotions you felt from receiving te gift. Focus on te moment wen you felt te most emotional from receiving te gift, and rate ow intensely you felt tat emotion (1 ¼ did not feel emotional at all from receiving te gift, 7 ¼ felt extremely emotional from receiving te gift); and Tink about te emotions you felt from consuming te gift. Focus on te moment wen you felt te most emotional from consuming te gift, and rate ow intensely you felt tat emotion (1 ¼ did not feel emotional at all from consuming te gift, 7 ¼ felt extremely emotional from consuming te gift). We asked participants to focus on te moment tey felt most emotional to remove te influence of edonic adaptation tat is more likely to ave occurred for te more durable material gifts (Nicolao et al. 2009). To account for tis difference in durability, we also asked participants to estimate te total amount of time tey ad spent consuming te gift. To explore te specific emotions evoked by teir gifts, we ten asked participants to select one primary emotion from a list of 30 randomly ordered discrete emotions tat tey were feeling at te moment tey felt most emotional (see appendix C). Tis list was followed by a text box, in case te emotion tey felt was not provided. Te listed emotions were drawn from te Positive and Negative Affective Scedule Expanded Form (PANAS-X; Watson and Clark 1994), including te two general dimension scales (10 positive and 10 negative emotions), along wit eigt additional basic emotions (four positive and four negative). Given our interest in te social aspects of a gift excange and consumption, we also added two emotions (embarrassed and grateful) tat serve important social functions (Fiscer and Manstead 2008; Tooby and Cosmides 2008).

8 8 JOURNAL OF CONSUMER RESEARCH Tougtfulness and Liking. Tougtfulness and liking of te gift were measured using te same items as in study 1. Perceived tougtfulness was measured using four items (a ¼.84), and liking was measured using tree items (a ¼.73). Oter Features of te Gift. To account for te likely variation among te gifts received, we asked recipients to estimate te price of te gift, to report wen tey ad received te gift, and to indicate ow tey were related to teir gift giver (spouse or significant oter, cild or grandcild, parent, anoter family member, close friend, acquaintance, colleague, or oter). Lastly, participants responded to manipulation cecks by rating te extent to wic te gift tey received was material or experiential (1 ¼ purely material, 5 ¼ equally material and experiential, 9 ¼ purely experiential), and by indicating weter tey ad consumed te gift wit teir gift giver (yes, no). Results Gifts Received. Sared experiential gifts included vacations, meals, and tickets to concerts or sporting events. Nonsared experiential gifts included music or dance lessons, spa services, vacations, meals, and tickets for events tat were not attended wit te gift giver. Sared material gifts included coffee makers, game consoles, televisions, tablet computers, and cars; and nonsared material gifts included jewelry, cloting, computers, portable music players, and digital cameras. Manipulation cecks confirmed tat participants in te experiential gift conditions recalled gifts tat were more experiential (M ¼ 7.55, SE ¼ 0.13) tan participants in te material gift conditions (M ¼ 2.90, SE ¼ 0.13; t(532) ¼ 25.49, p <.001, d ¼ 1.48); and most participants in te sared gift conditions (93%) consumed teir gifts wit teir gift giver (vs. 3% in te nonsared gift conditions; v 2 (1) ¼ , p <.001). Participants in te experiential gift conditions also consumed teir gift over a sorter period of time (M ¼ 3.41 days, SE ¼ 12.56) tan participants in te material gift conditions (M ¼ days, SE ¼ 12.24; t(532) ¼ 6.59, p <.001, d ¼.55), consistent wit te more durable nature of material gifts. Te estimated price of te gifts ranged from $1 (a magnet) to $19,000 (a car). Te majority of gifts (60%) were received witin te past year, but te oldest gift was received in Most gifts were received from a spouse or significant oter (37%), parent (19%), anoter family member (16%), or a close friend (19%). Cange in Relationsip Strengt. A22ANOVA conducted on cange in relationsip strengt revealed only a main effect of gift type, wit experiential gifts (M ¼ 0.66, SE ¼ 0.05) strengtening relationsips more tan material gifts (M ¼ 0.40, SE ¼ 0.05; F(1, 530) ¼ 11.81, p <.001, d ¼.30; see figure 1). Neiter te sared consumption main effect (p ¼.50) nor te gift type by sared consumption interaction (p ¼.81) was significant, suggesting tat te relationsip-strengtening effect of receiving an experiential gift occurred regardless of weter recipients consumed te gift wit teir gift giver (see web appendix A for robustness ceck). Given te wide range of gifts, we also conducted a 2 (gift type) 2 (sared) ANCOVA on cange in relationsip strengt, controlling for estimated price, date of receipt, and ow te recipient was related to te gift giver (dummy coded). Results again sowed tat receiving an experiential gift strengtened relationsips more tan receiving a material gift (F(1, 520) ¼ 6.83, p ¼.009, d ¼.23). Including te covariates did not affect te significance levels of te sared consumption main effect (p ¼.72) or te interaction effect (p ¼.32). Emotion from Consumption. To examine te role of emotion, we first conducted a 2 (gift type) 2 (sared) ANOVA on te extent to wic consuming te gift made recipients feel emotional. Te results revealed only a main effect of gift type, wit experiential gifts (M ¼ 5.14, SE ¼ 0.09) evoking greater emotion tan material gifts (M ¼ 4.70, SE ¼ 0.09; F(1, 530) ¼ 11.08, p <.001, d ¼.29). Tere was a nonsignificant effect of saring (p ¼.14) and a nonsignificant interaction effect (p ¼.50). Tese effects eld wen te covariates were included in te analysis. Te results again revealed only a main effect of gift type, wit experiential gifts evoking greater emotion tan material gifts (F(1, 520) ¼ 15.55, p <.001, d ¼.34), and a nonsignificant effect of saring (p ¼.92) and interaction effect (p ¼.90). Accounting for factors suc as te type of relationsip and te time tat as passed since te gift excange, tese results suggest tat consuming an experiential gift evokes greater emotion tan consuming a material gift, regardless of weter recipients consume te gift wit teir gift giver. Te vast majority of te specific Relationsip cange FIGURE 1 STUDY 3: RELATIONSHIPS IMPROVED MORE AMONG RECIPIENTS OF EXPERIENTIAL (VS. MATERIAL) GIFTS Material Nonsared Experiential Sared

9 AQ6 CHAN AND MOGILNER 9 emotions participants felt most intensely wile consuming te gift were positive (97.6%; see table 1). We next conducted a mediation analysis to test our teory tat te positive effect of receiving an experiential gift (vs. material gift) on relationsip strengt is driven by te greater emotion evoked during gift consumption. We entered recipients ratings of ow emotional consuming te gift was as te mediator, controlling for estimated price, date of receipt, and ow te recipient was related to te gift giver. As our previous analyses sowed, experiential gifts strengtened relationsips more tan material gifts (b ¼ 0.10, SE ¼ 0.04, t(522) ¼ 2.70, p ¼.007, d ¼.23). In addition, gifts tat were more emotional were more effective at improving relationsip strengt (b ¼ 0.14, SE ¼ 0.02, t(522) ¼ 33.95, p <.001, d ¼.49). Wen bot gift type and emotion were entered into te model predicting cange in relationsip strengt, te effect of consumption emotion remained significant (b ¼ 0.13, SE ¼ 0.02, Emotion TABLE 1 STUDY 2: EMOTIONS FELT MOST INTENSELY DURING GIFT CONSUMPTION AND GIFT RECEIPT (FIVE MOST COMMONLY REPORTED) Gift consumption %of participants Emotion Gift receipt %of participants Happy 29.0% Grateful 20.0% Deligted/ceerful 15.9% Deligted/ceerful 17.6% Grateful 13.1% Excited 17.4% Excited 10.3% Happy 13.7% Entusiastic 6.9% Surprised 13.5% FIGURE 2 t(521) ¼ 5.44, p <.001, d ¼.46), wereas te effect of gift type was no longer significant (b ¼ 0.07, SE ¼ 0.04, t(521) ¼ 1.80, p ¼.07, d ¼.15). Corroborating evidence was obtained in a bootstrap analysis, wic generated a confidence interval of te indirect effect tat did not cross zero (95% CI ¼ [.03,.12]; Hayes 2012; Zao, Lync, and Cen 2010; see figure 2). In sum, experiential gifts tend to be more emotional to consume, and gifts tat are more emotional to consume lead recipients to ave a stronger relationsip wit teir gift giver, tus supporting our ypotesis tat experiential gifts strengten relationsips more tan material gifts because tey evoke greater emotional intensity during consumption (ypotesis 2). Emotion from Gift Excange. Having identified te significant role of te emotion evoked during gift consumption, we next examined te emotion evoked during te gift excange. A 2 2 ANOVA conducted on ow emotional recipients felt upon receiving te gift sowed a nonsignificant main effect of gift type. Material and experiential gifts did not differ in ow emotional it was to receive te gift (p ¼.88). Te main effect of sared consumption (p >.99) and te gift type sared interaction (p ¼.63) were also not significant. Like te emotions evoked during gift consumption, te vast majority of te specific emotions participants felt most during te gift excange were positive (96.8%; see table 1). Tese findings are consistent wit our teorizing tat experiential and material gifts are similarly emotional wen received, and tus it is te emotion felt from gift consumption, rater tan te gift excange, tat is responsible for te greater relationsip-strengtening effect of experiential gifts. STUDY 3: EXPERIENTIAL GIFTS WERE MORE EMOTIONAL TO CONSUME AND THEREFORE MORE CONNECTING Emotion from gift consumption a =.26(.07)*** b =.13(.02)*** Experiential gift vs. material gift c =.10(.04)** c' =.07(.04) Relationsip cange Note: ** p <.01, *** p <.001, two-tailed test. Parameter estimates are listed wit standard errors in parenteses, wit estimated price of gift, date of gift receipt, and ow te recipient was related to te gift giver (dummy coded) as model covariates.

10 10 JOURNAL OF CONSUMER RESEARCH Tougtfulness and Liking. We again found no differences in tougtfulness and liking across conditions. A 2 2 ANOVA predicting perceived tougtfulness of te gift revealed no significant effects for gift type (p ¼.77), sared consumption (p ¼.16), or te interaction (p ¼.08). Similarly, a 2 2 ANOVA predicting liking of te gift revealed no significant effects for gift type (p ¼.19), sared consumption (p ¼.42), or te interaction (p ¼.75). Discussion Examining a wide range of real-world gifts across a variety of relationsips, tis study provided robust evidence tat experiential gifts strengten relationsips more tan material gifts, regardless of weter gift recipients and givers consume te gift togeter (ypotesis 1). Furtermore, te mecanism underlying tis effect is te intensity of emotion evoked during gift consumption, wic is distinct from te emotion evoked during te gift excange. Specifically, consuming experiential gifts evokes greater emotion tan consuming material gifts, and it is tis emotional intensity tat strengtens recipients relationsips wit teir gift givers (ypotesis 2). Because te vast majority of participants in study 3 reported te emotion tey felt most intensely wile consuming teir gift to be positive, tere was not sufficient data to assess weter te effect of emotion on relationsip strengt would generalize to negative emotions felt during gift consumption. For example, would an intense feeling of sadness wile watcing a performance of Madame Butterfly or an intense feeling of fear wile watcing Silence of te Lambs strengten te giver-recipient relationsip? To explore te role of emotional valence, we conducted a similar study in wic we asked participants (N ¼ 523; 46% female, tree unspecified; ages 18 66, M ¼ 32.0, one unspecified) to recall a significant material or experiential gift tey ad received. Participants rated ow muc teir relationsip wit te gift giver ad strengtened as a result of te gift, as well as ow intensely tey felt eac of 30 discrete emotions wile consuming teir gift (15 positive emotions and 15 negative emotions; see appendix C). Ratings for all 30 emotions were averaged to create an index of overall emotion. In addition, te ratings for te positive and te negative emotions were averaged separately. Te results sowed tat recipients of experiential (vs. material) gifts felt more emotional overall (M exp ¼ 3.29, SE ¼ 0.07 vs. M mat ¼ 3.02, SE ¼ 0.07; F(1, 510) ¼ 20.02, p <.001, d ¼.39), and tis effect eld for purely positive emotions (M exp ¼ 3.73, SE ¼ 0.06 vs. M mat ¼ 3.52, SE ¼ 0.07; F(1, 510) ¼ 12.96, p <.001, d ¼.30), and purely negative emotions (M exp ¼ 2.22, SE ¼ 0.07 vs. M mat ¼ 2.08, SE ¼ 0.07; F(1, 510) ¼ 5.09, p ¼.03, d ¼.20). Furtermore, significant indirect effects were observed wen we used te average of all 30 discrete emotions (95% CI ¼ [.05,.15]), as well as just te 15 positive emotions (95% CI ¼ [.04,.14]), and just te 15 negative emotions (95% CI ¼ [.003,.08]) as mediators for te effect of gift type on cange in relationsip strengt. Tis offers preliminary evidence suggesting tat strong negative emotions evoked troug gift consumption can also strengten giver-recipient relationsips. STUDY 4: THE MODERATING ROLE OF EMOTION FROM GIFT CONSUMPTION Building on te mediation evidence from study 3, study 4 sougt furter evidence for te underlying role of emotion from gift consumption troug a test of moderation. Tis study followed a 2 (gift type: material vs. experiential) 2 (emotion: control vs. emotion) between-subjects design. Participants were asked to recall eiter an experiential gift or material gift tey ad received, and alf were more specifically instructed to recall a gift tat ad evoked intense emotion during consumption. Additionally, tis study design allowed us to identify a boundary condition for te benefit of receiving experiential gifts over material gifts. Altoug we argue tat consuming experiential gifts tends to evoke greater emotion tan consuming material gifts, tere surely are some material gifts tat elicit a great deal of emotion wen consumed. For example, wearing one s engagement ring opefully makes one feel incredibly loved and loving, and looking at a potograp tat captures a meaningful moment sould stir emotion. Because we argue tat experiential gifts strengten giver-recipient relationsips by eliciting more intense emotion during consumption, material gifts tat evoke intense emotion sould similarly strengten relationsips. Metod One tousand forty-two participants were recruited troug Amazon Mecanical Turk to participate in tis study in excange for $0.50. Participants wo did not complete te survey (n ¼ 26) or wrote tat tey could not follow te instructions or could not tink of a gift (n ¼ 21) were eliminated from te analysis. Tis left 995 participants in te analyzed dataset (45% female, two unspecified; ages 18 77, M ¼ 33.2). Gift Manipulations. Participants were randomly assigned to one of te four conditions comprising te 2 (gift type: material vs. experiential) 2 (emotion: control vs. emotion) design. Participants in te control conditions were asked to Please recall and describe [a material/an experiential] gift you ve received. Participants in te emotion conditions were asked to Please recall and describe [a material/an experiential] gift you ve received tat makes or made you feel emotional wile consuming it. Participants were provided wit te definition of material or

11 CHAN AND MOGILNER 11 experiential gifts from studies 1 and 3. All participants were also instructed tat te gift sould be one tey received from someone tey know and one tey did not consume wit te gift giver. Cange in Relationsip Strengt. Tis study used yet anoter measure for cange in relationsip strengt, wic was adapted from te Relationsip Satisfaction Level index of Rusbult et al. s (1998) Investment Model Scale. Participants rated teir agreement (1 ¼ don t agree at all, 5 ¼ somewat agree, 9 ¼ agree completely) wit five statements (a ¼.95): I feel more satisfied wit our relationsip as a result of te gift ; Our relationsip is closer to ideal as a result of te gift ; Our relationsip is muc better tan oters relationsips as a result of te gift ; Our relationsip makes me very appy as a result of te gift ; and Our relationsip does a good job of fulfilling my needs for intimacy, companionsip, etc. as a result of te gift. Tougtfulness and Liking. Tougtfulness and liking of te gift were measured using te same items as in studies 1 and 3. Again, perceived tougtfulness of te gift was measured using four items (a ¼.78), and liking was measured using tree items (a ¼.56). Oter Features of te Gift. To account for te variation among te gifts received, we asked recipients to estimate te price of te gift, to report wen tey ad received te gift, and to indicate ow tey were related to teir gift giver (spouse or significant oter, cild or grandcild, parent, anoter family member, close friend, acquaintance, colleague, or oter) and for wat occasion tey ad received te gift (birtday, wedding, anniversary, Cristmas, Valentine s Day, Moter s Day, Fater s Day, graduation, for no special occasion, or oter). Lastly, participants responded to manipulation cecks by rating te extent to wic te gift tey received was material or experiential (1 ¼ purely material, 5 ¼ equally material and experiential, 9 ¼ purely experiential), and ow emotional tey ad felt wile consuming te gift (1 ¼ did not feel emotional at all from consuming te gift, 9 ¼ felt extremely emotional from consuming te gift). Results Gifts Received. As in study 3, te experiential gifts recalled included vacation, meals, and tickets to performances and events. Te recalled material gifts included cloting, electronics, musical instruments, and jewelry. A22ANOVA conducted on te material-experiential manipulation ceck confirmed a main effect of gift type (F(1, 991) ¼ , p <.001, d ¼ 1.45), along wit a main effect of emotion condition (F(1, 991) ¼ 11.72, p <.001, d ¼.15) and an interaction (F(1, 991) ¼ 34.88, p <.001, d ¼.25). Te experiential gifts were rated as more experiential tan te material gifts in te control conditions (M exp ¼ 7.72, SE ¼.13 vs. M mat ¼ 2.42, SE ¼.13; t(991) ¼ 28.85, p <.001, d ¼ 1.70) and emotion conditions (M exp ¼ 7.39, SE ¼.13 vs. M mat ¼ 3.66, SE ¼.14; t(991) ¼ 19.48, p <.001, d ¼ 1.19). A 2 2 ANOVA conducted on te emotion manipulation ceck confirmed a main effect of emotion condition (F(1, 991) ¼ , p <.001, d ¼.62), as well as a main effect of gift type (F(1, 991) ¼ 11.83, p <.001, d ¼.21), and an interaction (F(1, 991) ¼ 12.84, p <.001, d ¼.22). Te recalled emotional gifts were rated as more emotional tan te control gifts in te experiential conditions (M emot ¼ 5.63, SE ¼.09 vs. M cont ¼ 5.02, SE ¼.09; t(991) ¼ 4.84, p <.001, d ¼.40) and material conditions (M emot ¼ 5.64, SE ¼.10 vs. M cont ¼ 4.38, SE ¼.09; t(991) ¼ 9.62, p <.001, d ¼.83). Consistent wit our teorizing tat experiential gifts tend to evoke more intense emotion during consumption tan material gifts, tese results sow tat in te control conditions, experiential gifts were more emotional tan te material gifts (M exp ¼ 5.02, SE ¼.09 vs. M mat ¼ 4.38, SE ¼.09; t(991) ¼ 5.07, p <.001, d ¼.42). Employing yet anoter approac to ceck te emotion manipulation, we conducted a textual analysis on participants written description of teir gift using te Linguistic Inquiry and Word Count (LIWC; Pennebaker, Boot, and Francis 2007), wic is an effective measure of te amount of emotion expressed (Kan et al. 2007). Te LIWC enumerated te percentage of emotion words written by eac participant. Tis additional ceck confirmed tat, overall, participants in te emotion conditions expressed greater emotion wen writing about teir gift, and tat witin te control conditions, participants in te experiential condition also expressed greater emotion wen writing about teir gift. Specifically, participants in te experiential emotion (M ¼ 6.89, SE ¼ 0.31), material emotion (M ¼ 6.42, SE ¼ 0.33), and experiential control (M ¼ 5.08, SE ¼ 0.31) conditions eac wrote a significantly iger percentage of emotion words tan did participants in te material control condition (M ¼ 3.83, SE ¼ 0.31; p s <.005). Again, tis is consistent wit our teorizing tat experiential gifts tend to evoke greater emotion tan material gifts. Te estimated price of te gifts ranged from $0 (e.g., a anded-down sirt) to $25,000 (a wedding ring). Fifty percent of gifts were received witin te past year, and te oldest gift was received in Most gifts were received from a spouse or significant oter (32%), parent (24%), anoter family member (17%), or a close friend (18%), and were received for a birtday (34%), Cristmas (25%), or no special occasion (26%). Wedding and anniversary gifts combined constituted less tan 6% of te gifts. Cange in Relationsip Strengt. A 2 (gift type) 2 (emotion) ANOVA conducted on cange in relationsip strengt revealed a main effect of gift type (F(1, 991) ¼ 21.33, p <.001, d ¼.29), a main effect of emotion (F(1, 991) ¼ 17.81, p <.001, d ¼.26), and te predicted

12 12 JOURNAL OF CONSUMER RESEARCH interaction (F(1, 991) ¼ 10.62, p ¼.001, d ¼.21). Consistent wit our previous findings, in te control conditions, experiential gifts (M ¼ 5.78, SE ¼ 0.13) resulted in greater improvements in relationsip strengt tan material gifts (M ¼ 4.71, SE ¼ 0.13; t(991) ¼ 5.69, p <.001, d ¼.49). Furtermore, in support of our proposed underlying role of emotion, material gifts tat evoked intense emotion during consumption (M ¼ 5.73, SE ¼ 0.15) resulted in similar improvements in relationsip strengt as experiential gifts tat evoked emotion (M ¼ 5.92, SE ¼ 0.13; t(991) ¼ 0.94, p ¼.35, d ¼.08; figure 3). Controlling for te estimated price of te gift, date of receipt, ow te recipient was related to te gift giver (dummy coded), and te gift occasion (dummy coded) did not affect te significance of te main effect of gift type (F(1, 972) ¼ 15.72, p <.001, FIGURE 3 STUDY 4: RELATIONSHIPS IMPROVED MORE AMONG RECIPIENTS WHO RECEIVED EXPERIENTIAL (VS. MATERIAL) GIFTS OR GIFTS THAT WERE EMOTIONAL TO CONSUME Relationsip cange Control Material Experiential Emotion TABLE 2 d ¼.26), emotion (F(1, 972) ¼ 15.20, p <.001, d ¼.24), or teir interaction (F(1, 972) ¼ 10.10, p ¼.002, d ¼.20). We observed tat many of te emotional material gifts were pieces of jewelry commemorating a meaningful life event (e.g., engagement or wedding, birt of a cild, graduation) or passed down as eirlooms. Oters included potograps and religious items (e.g., bible, rosary). Toug te predicted effects of gift type and emotion on cange in relationsip eld wen we controlled for gift occasion and oter features of te gifts (i.e., price, date of receipt, and relationsip to te gift giver), it is still possible tat te emotion manipulation elicited gifts in participants minds tat differed in ways oter tan teir material or experiential distinction. To assess tis, we asked two researc assistants wo were blind to study conditions and ypoteses to categorize te gifts participants ad received. First, te researc assistants jointly determined purcase categories tat encompassed te full range of gifts (e.g., travel; food and drink; cloting, soes, and accessories). Te researc assistants ten independently assigned eac gift to one of te 10 purcase categories (84% agreement, and disagreements were resolved troug discussion). An examination of te coded gifts revealed some differences among te material gifts between te control and emotion conditions (see table 2). For example, jewelry was more likely to be mentioned as an emotional material gift, and toug electronics were also frequently mentioned as an emotional material gift, tey were more likely to be mentioned in te control condition. More importantly, owever, te purcase category of te gifts received did not affect te primary outcome of relationsip cange. Controlling for purcase category (dummy coded), a 2 (gift type) 2 (emotion) ANCOVA conducted on cange in relationsip strengt still revealed a main effect of gift type (F(1, 959) ¼ 3.89, p <.05, d ¼.20), a main effect of emotion STUDY 4: PERCENTAGE OF GIFTS RECEIVED (BY CONDITION) AND MEAN CHANGE IN RELATIONSHIP STRENGTH IN EACH PURCHASE CATEGORY Material Experiential Mean cange in relationsip strengt Control Emotion Control Emotion (n ¼ 259) (n ¼ 220) (n ¼ 258) (n ¼ 258) Travel 0% 0.9% 24.4%* 16.3%* 6.18 Recreation and leisure 1.5% 2.7% 21.3% 16.7% 6.10 Food and drink 0%* 2.3%* 19.4% 19.4% 5.67 Beauty and wellness 2.7% 2.3% 8.1% 6.2% 5.60 Arts, music, and literature 5.8% 9.6% 20.9% 25.2% 5.58 Jewelry 10.8%* 19.6%* 0% 1.6% 5.54 Electronics and tecnology 36.7%* 19.6%* 0%* 4.3%* 5.27 Home and garden 8.1% 13.2% 0% 0.8% 5.21 Oter 7.0%* 12.7%* 3.9% 5.4% 5.19 Cloting, soes, and accessories 27.4%* 17.3%* 1.2% 4.3% 4.86 Note: Witin eac material or experiential gift condition, * represents a statistical difference between te control and emotion conditions for eac purcase category at p <.05.

13 CHAN AND MOGILNER 13 (F(1, 959) ¼ 19.39, p <.001, d ¼.28), and te predicted interaction (F(1, 959) ¼ 8.42, p ¼.004, d ¼.18); te effect of purcase category was not significant (F(1, 959) ¼ 1.43, p ¼.17, d ¼.22), and none of te individual purcase categories significantly predicted te cange in relationsip strengt. Tis suggests tat te specific purcases generated in eac condition cannot explain te effect of consumption emotion on relationsip strengt for experiential versus material gifts. Moreover, an examination of te average cange in relationsip strengt across te various purcase categories reflects a pattern consistent wit te primary ypotesis: recipients of gifts in more experiential categories reported greater improvements in relationsip strengt tan tose wo received gifts in more material categories (see table 2). Tougtfulness and Liking. A22 ANOVA predicting tougtfulness revealed only a main effect of emotion, wit emotional gifts (M ¼ 6.23, SE ¼.04) being considered more tougtful tan control gifts (M ¼ 6.12, SE ¼ 0.04; F(1, 991) ¼ 3.85, p ¼.05, d ¼.12). Te effects of gift type and te interaction were not significant (p s >.16). A 2 2 ANOVA predicting liking revealed a main effect for gift type, wit experiential gifts (M ¼ 6.18, SE ¼.04) being better liked tan material gifts (M ¼ 6.04, SE ¼ 0.04; F(1, 991) ¼ 6.31, p ¼.01, d ¼.16). Te effects of te emotion manipulation and te interaction were not significant (p s >.50). Wen tougtfulness and liking were included as covariates in te analysis of cange in relationsip strengt, te results eld: we still observed a main effect of gift type (F(1, 989) ¼ 19.71, p <.001, d ¼.27), of emotion (F(1, 989) ¼ 15.68, p <.001, d ¼.24), and teir interaction (F(1, 989) ¼ 10.20, p <.001, d ¼.19). Again, tis suggests tat perceived tougtfulness and liking of te gift are not responsible for te greater ability of experiential gifts to strengten relationsips. Discussion Using yet anoter measure of cange in relationsip strengt, te control conditions in study 4 replicated te beneficial effect of receiving experiential gifts over material gifts observed in te previous studies. Tis effect was robust even wen we controlled for many oter features of te gift. Moreover, tis study used a test of moderation to provide additional evidence for te underlying role of consumption emotion and to identify an important boundary condition for te effect: material gifts tat made recipients feel emotional wile consuming tem were as effective at strengtening te relationsip as experiential gifts. GENERAL DISCUSSION Consumers spend a lot of money on oters (Americans spend approximately $130 billion on gifts per year; Unity Marketing 2015), and spending money on oters as been proven to increase one s own appiness (Dunn, Aknin, and Norton 2008a). Te current researc explores te more farreacing effect on relationsips between people, finding tat not all prosocial expenditures are equally beneficial. Despite gift givers tendencies to give material possessions, our findings sow tat material gifts are less effective tan experiential gifts at strengtening relationsips between gift givers and teir recipients. Tis researc provides guidance for gift givers on wat to give and offers insigt into te relational function of gifts. Taking te recipients perspective to assess te success of gifts, we conducted experiments involving a variety of real-life gift excanges and ways of measuring relationsip cange, and we consistently found tat experiential gifts strengten relationsips more tan material gifts (studies 1 4). Tis effect also emerged wen te very same gift was framed as being relatively more experiential (study 2). A driving factor underlying tis effect is te greater level of emotion elicited wen recipients consume experiential gifts versus material gifts, wic we identified troug tests of mediation (study 3) and moderation (study 4). Even toug tere was no difference in te intensity of emotion recipients felt upon receiving experiential and material gifts, recipients felt more emotional wen consuming experiential (vs. material) gifts, wic served to strengten teir relationsip wit te gift giver. From tis, we learn tat gift givers seeking to foster closer relationsips wit teir recipients are likely to acieve greater success by giving experiential gifts, rater tan material gifts. Teoretical Contributions To build on te now-establised body of work tat as demonstrated tat purcasing experiences (vs. material goods) for oneself positively affects one s personal wellbeing (Carter and Gilovic 2010; Gilovic et al. 2015; Nicolao et al. 2009; Rosenzweig and Gilovic 2012; Van Boven and Gilovic 2003), researc as begun investigating factors tat precede experiential and material purcasing (Dai, Can, and Mogilner 2016; Kumar and Gilovic 2016; Kumar, Killingswort, and Gilovic 2014; Tully, Hersfield, and Meyvis 2015). Findings suggest tat te benefit of acquiring experiences for te purcaser can be largely explained by te typically more social nature of experiences (Battacarjee and Mogilner 2014; Caprariello and Reis 2013). Our findings furter contribute to tis burgeoning stream of researc by being te first to sow te interpersonal consequences of experiential versus material purcases. In addition, we identify a novel advantage of experiential purcases: consuming an experience evokes greater emotion tan consuming a material possession. Our finding tat te emotion felt during gift consumption is responsible for strengtening relationsips is consistent wit past work on interpersonal relationsips tat AQ2

14 14 JOURNAL OF CONSUMER RESEARCH AQ3 as igligted te importance of emotion in close relationsips (Aron et al. 2000; Bazzini et al. 2007; Clark and Finkel 2004; Laurenceau et al. 1998; Nummenmaa et al. 2012; Peters and Kasima 2007; Ragunatan and Corfman 2006; Ramanatan and McGill 2007; Slatcer and Pennebaker 2006). Our researc builds on tis literature by sowing tat te gift of an emotional experience can strengten relationsips, even wen relationsip partners do not consume te gift togeter. Our researc also contributes to gift-giving researc by testing ow different types of gifts impact relationsips and by examining te emotion evoked from gift consumption. Te bulk of te existing experimental work examining recipients responses to gifts as focused on identifying gifts tat are better liked and appreciated (Flynn and Adams 2009; Gino and Flynn 2011), rater tan on understanding ow gifts can cange te relationsip between te gift giver and recipient (Aknin and Human 2015). Altoug our findings indicate tat te extent to wic te recipient likes te gift is positively related to cange in relationsip strengt, liking does not appear to drive our effect, as we did not find consistent differences in ow muc recipients liked experiential and material gifts. Instead, te gift s emotionality is wat seems to make experiential gifts better tan material gifts at strengtening relationsips. By examining emotions evoked during gift consumption, our researc also complements qualitative work examining te emotion tat arises during te gift excange (Belk and Coon 1993; Rut et al. 1999, 2004). By studying te emotion evoked from gift consumption and testing its impact on relationsips, our researc tus provides new insigts into gift giving. Future Researc and Marketing Implications Altoug experiences tend to be more emotional, our researc sows tat emotionally evocative material gifts can also strengten relationsips. Future work sould explore ow possessions become associated wit emotion, and wat types of possessions are most meaningful. For example, recent researc as found tat items associated wit loved ones and special events carry sentimental value (Yang and Galak 2015). Terefore, would material gifts tat commemorate experiences be more prone to evoke emotion so as to effectively strengten relationsips? Gift giving is a ripe context for suc investigations into meaningful possessions, in ligt of te underlying role of emotion and te focus on interpersonal relationsips. Future researc sould also delve furter into te patways troug wic te emotions evoked from gifts can affect relationsips. Our findings sowed tat emotional gifts tat were not jointly consumed by te gift giver and recipient were equally effective at strengtening relationsips. In tese instances, is te recipient linking tese consumption emotions to te relationsip (e.g., vicariously saring te gift consumption wit te giver and tereby feeling closer), or is te process more indirect (e.g., feeling closer to a giver wo as emotionally enriced te recipient s life)? A related question is weter tere are particular emotions tat are more connecting tan oters. For instance, are gifts given out of gratitude versus guilt differentially connecting (Can, Mogilner, and Van Boven 2016)? And wat are te circumstances in wic negative emotions are connecting? For example, to deepen our understanding of ow experiential gifts can affect relationsips, future researc sould contrast te effects of negative emotions tat are intentional (e.g., fear from watcing a scary movie) versus unintentional (e.g., frustration due to bad service at a restaurant), or te effects of negative emotions directed at te experience (e.g., sadness over a tragic play) versus at te relationsip partner (e.g., anger because te partner arrived late for te sow). Prior work as sown tat te benefits of purcasing experiences over material goods for oneself are attenuated and sometimes reversed wen te purcase outcome is negative (Nicolao et al. 2009); terefore, it is quite possible tat te effects of unintended negative consumption emotions due to failed experiential gifts could be particularly detrimental for relationsips. Wen examining te effect of gifts on relationsips, one must also consider ow to measure relationsip cange. Bot study 1 and te Fater s Day pilot study measured relationsip strengt at two time periods before and after receiving te gift to directly assess te gift s impact on te relationsip. In subsequent studies, participants retrospectively evaluated ow a gift ad affected teir relationsip wit te giver. Toug te latter approac allowed us to more feasibly manipulate gift type for actual gifts received in real relationsips for real gift occasions, it also increased te possibility of participants reporting teir lay beliefs about gifts, rater tan teir true reactions. Tis concern is alleviated, owever, because te predicted pattern of results was consistent across studies tat used te two-stage approac and te retrospective approac. In addition, te retrospective studies sowed effects tat lay teory would be unlikely to predict: experiential gifts improve relationsips even wen not consumed wit te gift giver. Furtermore, te use of between-subject designs lessened te potential for demand effects. Still, future researc on gift giving sould strive to employ longitudinal designs tat measure relationsip strengt bot before and after gift excanges to establis te role gifts play in ongoing relationsips. Longitudinal designs could also be used to conduct a longer-term examination of te effects of gifts on relationsips to furter contribute to te gift-giving literature. Across our studies, we focused on te sort-term effects of receiving a single gift. However, a gift could ave a longer-lasting effect on a relationsip (Algoe et al. 2008), and migt influence future gift-giving interactions. Furter, altoug we did not observe a significant effect of saring

15 CHAN AND MOGILNER 15 AQ4 in te consumption of te gift, it may be tat te benefits of saring in experiential gift consumption could emerge later on as tose cerised sared memories gain greater interpersonal value. More generally, tis researc direction would substantiate te relationsip-strengtening effects of gifts over time and more broadly inform te impact of gift giving on relationsips. Across our studies, we examined te effects of different types of gifts on relationsips contingent upon a gift being given. However, te act of gift giving in itself as social value (Ward and Can 2015), and it would be wortwile to examine its role in maintaining relationsips over time, compared to relationsips in wic no gifts are given. Te current researc empasizes te interpersonal benefits of experiential gifts. Future work could explore potential intrapersonal benefits of giving experiences. Engaging in relationsip maintenance beaviors as been found to increase individual well-being wen tese efforts are successful in improving relationsip quality, but to decrease well-being wen tese efforts are unsuccessful (Baker et al. 2012), and researc as documented ow personal appiness can be gained from prosocial spending (Dunn et al. 2008a; Hersfield et al. 2016). Because giving experiential gifts is more effective at strengtening relationsips, gift givers migt derive greater edonic benefits from giving an experience tan a material good. Furtermore, gift givers migt reap personal benefits from saring in te experience wit te recipient since giving one s own time can lead to greater feelings of interpersonal connection and self-efficacy (Mogilner, Cance, and Norton 2012). Future researc could also examine weter te relational benefits observed in tis researc extend to consumer-brand relationsips. For example, rater tan promoting mercandise rewards, te Starwood Hotels & Resorts Starwood Preferred Guest loyalty program encourages its members to redeem teir Starpoints for incredible experiences and unforgettable events. We see tat retailers, suc as Sepora, Nordstrom, and Saks Fift Avenue, give private parties and events for teir loyal customers in addition to more material gifts, suc as free cosmetic items. Follow-up work sould test weter experiential rewards are more effective at strengtening consumer-brand connections tan material rewards. Finally, companies tat sell experiences, suc as tose in te travel or entertainment industry, sould create opportunities for consumers to easily purcase experiences to give as gifts. For example, Travelers Joy is a service tat enables engaged couples to build an experiential gift registry for teir oneymoon, so tat family and friends can give part of teir oneymoon (e.g., a surf lesson, dinner, adventure tour) as a wedding gift. Given tat gift recipients prefer receiving gifts from teir registry over individually selected gifts (Gino and Flynn 2011), our researc implies tat suc experiential gift registries will benefit gift givers, recipients, and te companies tat provide experiences. Conclusion Consumers frequently struggle wit te callenge of coosing wat to give. Most gift-giving occasions are terefore accompanied by a flurry of advice columns and top 10 lists of gift ideas, as media and marketers try to elp consumers make coices tat will improve teir relationsips. Tis researc offers simple guidance: to make your friend, spouse, or family member feel closer to you, give an experience. DATA COLLECTION INFORMATION Te first autor supervised te collection of data for studies 1 and 2 by researc assistants at te University of Pennsylvania s Warton Beavioral Lab in summer 2013 and winter 2011, respectively. Te first autor managed te data collection for studies 3 and 4 on Amazon s Mecanical Turk in summer 2013 and fall 2015, respectively. Te data were analyzed by te first autor in collaboration wit te second autor.

16 16 JOURNAL OF CONSUMER RESEARCH APPENDIX A STUDIES 1 AND 3: INCLUSION OF OTHER SCALE ADAPTED FROM Aron et al. (1992) APPENDIX B STUDY 2: COFFEE MUGS GIVEN AS GIFTS AQ5 Relationsip before receiving gift Relationsip after receiving gift APPENDIX C STUDY 3: 30 DISCRETE EMOTIONS MEASURED PANAS-X General Dimension Scales Positive affect: active, alert, attentive, determined, entusiastic, excited, inspired, interested, proud, strong Negative affect: afraid, scared, nervous, jittery, irritable, ostile, guilty, asamed, upset, distressed Oter positive: appy, deligted/ceerful, calm, surprised, grateful Oter negative: sad, lonely, angry, disgusted, embarrassed REFERENCES Adams, Gabrielle S., Francis J. Flynn, and Micael I. Norton (2012), Te Gifts We Keep on Giving: Documenting and Destigmatizing te Regifting Taboo, Psycological Science, 23 (10), Aknin, Lara B. and Lauren J. Human (2015), Give a Piece of You: Gifts tat Reflect Givers Promote Closeness, Journal of Experimental Social Psycology, 60, 8 16.