What's in a tip? The creation and refinement of a restaurant-tipping motivations scale: A consumer perspective

International Journal of Hospitality Management 37 (2014) 121–130

Contents lists available at ScienceDirect

International Journal of Hospitality Management journal homepage: www.elsevier.com/locate/ijhosman

What’s in a tip? The creation and refinement of a restaurant-tipping motivations scale: A consumer perspective Jeremy E. Whaley a,∗ , Alecia C. Douglas b , Martin A. O’Neill b a b

College of Business, Department of Economic Development and Tourism, The University of Southern Mississippi, United States College of Human Sciences, Department of Nutrition, Dietetics, and Hospitality Management, Auburn University, United States

a r t i c l e Keywords: Consumer behavior Tipping Motivations Service Social norms

i n f o

a b s t r a c t According to Segrave (1998), since the late 1800s, the study of tipping has provoked debate in a range of abstract dimensions such as economics, sociology, and psychology. To date, the research on the topic has been largely qualitative in nature, while addressing motivating themes (service, social norm, and future service considerations) in isolation from one another. Following a thorough examination of the literature, there is a definite lack of research on the development and testing of a more holistic quantitative scale aimed at identifying the motivational Gestalt driving actual consumer tipping behavior. Therein lies the major theoretical contribution of this study, namely the development and testing of a Tipping Motivations Scale, which over two separate analyses, supports a number of drivers of consumer tipping motivation. In this study, exploratory and confirmatory factor analyses were conducted to test the empirical dimensions of consumer tipping motivations. The results obtained indicate a reasonable fit between the data and the proposed model across both analyses. This was repeated on two separate occasions and the results largely remained consistent. The findings point to the key role of service in driving the consumer’s decision to tip. Other important factors included social conformity, the issue of future visitation, and server actions. Further research is needed to explore whether these dimensions remain constant among other sample groups and across different tipped professions. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Motivations that drive restaurant tipping behavior are a much researched yet little understood phenomena in the hospitality domain as well as the wider service fields. The act of tipping is a commonly accepted and highly evolved custom in many countries and remains vital to the livelihood of service professionals. Tipping is the main source of income for millions and therefore is closely related to overall compensation in the restaurant industry (Azar, 2007a). In cultures such as the United States (US), most customers understand and conform to the custom of tipping. An annual estimate of tips earned by restaurant servers in the US accounted for $27 billion (Azar, 2007b). From a purely economic perspective, tipping is perceived as a rather strange or illogical economic action (Saunders and Lynn, 2010), given that consumers tip through choice, rather than because they are legally or ethically bound as patrons to tip. Further, the act of tipping is contradictory to normal economic exchange theory in that gratuities signify buyer determined values for services and voluntary expenditures that escalate the costs of amenities already received (Lynn et al., 1993). These

∗ Corresponding author. Tel.: +1 601 266 5100; fax: +1 601 266 6707. E-mail address: [email protected] (J.E. Whaley). 0278-4319/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhm.2013.11.005

factors imply that other less understood motivational influences are at play when it comes to the voluntary act of leaving a tip for service received. It is therefore not surprising that academicians are fascinated with tipping and, while there are no definitive answers regarding why consumers support the desired behavior, a body of previous research does provide plausible theories that might help explain these underlying factors driving the practice. For example, research suggests that consumers tip for a variety of reasons, such as social norms and pressures, social approval, equity in exchange relationships, and to acknowledge service received (Lynn and Grassman, 1990). A significant portion of the body of literature on tipping motivation research thus far has addressed motivational concepts in isolation, and mostly from a qualitative perspective. Further, the literature on tipping reviewed for this study used methods commonly employed in social and behavioral sciences research which includes the direct observations of server/guest interactions and self-reported questionnaires. The study described here reports on the creation and testing of the tipping motivation scale (TMS), a multi-dimensional, quantitative measure of tipping from the perspective of restaurant consumers. Using theoretical support from extant tipping literature, this scale was developed to represent a broader range of motivations that may potentially drive the consumer’s decision

122

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

to tip restaurant servers. By employing a series of qualitative and quantitative methods over four phases of the scale development process, the resultant TMS was achieved after the following phases: (1) item generation from the existent literature and item selection using the card sort technique in phase 1; (2) exploratory principal components analysis of TMS-1 items in phase 2; (3) focus group session and face-to-face interviews on the refined TMS-2 instrument in phase 3; and (4) scale purification and confirmatory analysis of TMS-2 in phase 4. 2. Literature review 2.1. What is a tip? Tipping is a custom that can be described in several different ways. For example, a tip can be defined as a gift, as a way of monetarily rewarding or punishing a service provider, or as an obligation for services received. As Shamir (1984) explained, “a price can be fixed on a hotel room, on a meal, or on a distance traveled by taxi or bus, but not on the smiles, the friendly gestures, the hospitable attitudes, etc.” (p. 62). When considering the evolution of tipping, there appear to be two dominant schools of thought: the first where the practice can be described as employer-driven whereas the second is consumer-led. From the employer-driven perspective, Segrave (1998) asserts that the practice of tipping has afforded more personal wealth to those socially irresponsible employers who cut labor costs and play on the emotions of customers who are sympathetic toward struggling low wage servers. The consumer-driven perspective on the other hand speaks to the exchange theory argument where tipping evolved as a system by which customers were freely able to monitor a company’s services and reward employees directly (or not) for the quality of service actually rendered. This perspective was further supported by Azar (2004) who suggested that the act of tipping servers was not only to reward good service but to promote better future service by giving employees “an incentive to do their best to satisfy the customer’s needs” (p. 761). While it is hard to say with any certainty when and where the custom originated, tipping as an economic activity and the motivations that drive it have frustrated fair wage and labor rights advocates, social scientists, economists, and reporters alike for more than a century (Segrave, 1998). A review of the extant literature does highlight a number of economic and psychological motives (Saunders and Lynn, 2010) such as rewarding good service delivery while at the same time guaranteeing future service delivery (Azar, 2005; Bodvarsson and Gibson, 1999; Lynn, 2003), pressure to conform to societal norms and a closely related need for social approval (Azar, 2004, 2005, 2006, 2007a,b; Boyles et al., 2006; Conlin et al., 2003; Lynn, 2001; Shamir, 1984), and out of empathy for service workers. Given these previously mentioned studies that support the overall complexity of tip-giving behavior, literary support provides the theoretical foundation for scale development in this study. Plausible tip-influencing themes found in this study are tipping and service, tipping and social norms, tipping and future service considerations, and tipping as the act relates to daily operations processes. 2.2. Tipping and service Perhaps the most widely researched motive is related to the provision of quality service and investment in future service (Azar, 2004; Strohmetz and Rind, 2001). Saunders and Lynn (2010) suggest that tipping serves as a quality control lever where consumers assume control over the determinants of good service. Poor service generally produces meager tips and results in little to no

income for the server, lending support to the belief that the main justification for tipping is that it promotes better service, by giving the workers an incentive to satisfy the guests’ needs (Azar, 2004). Paradoxically, service workers may strategically tailor the actual service provided according to the size of tip received from a previous service encounter. This view was also supported by Bodvarsson et al. (2003), who found that “service quality significantly affects tip size and when servers expect higher tips, customers rank service quality higher” (p. 1659). There are two underlying dimensions of service quality to consider; the technical and the functional aspects of service. The technical dimension of service refers to the more tangible and objectively measurable elements of the product or service being supplied. For example, in the context of a restaurant, technical elements of the service encounter might include the number of visits to the table (order taking, food delivery, and menu knowledge), the timeliness of the service delivery process, the accuracy of the menu description (promise) and whether it measures up in terms of what was served (delivery) – in other words, processes and/or systems the server utilizes in order to complete the task. The functional dimension, on the other hand, refers to the softer and/or more personal elements of the service exchange; or what might best be described as the more intangible, subjective, and/or relational elements of the service encounter – in other words, service delivery. Interestingly, research suggests that when a company fails in relation to the more technical aspects of service, the functional, more relational aspects of delivery may trump the breakdown in systems (O’Neill, 2000). Indeed evidence clearly suggests (Israeli and Barkan, 2004) that when a technical failure does occur the degree of satisfaction felt as a consequence of the recovery effort can sometimes lead to a greater degree of satisfaction and a larger tip than had the service been completed right first time. A number of studies support the proposition that the more functional aspects of service delivery permit real service differentiation and are highly correlated with actual tipping behavior. For example Speer (1997) and Jewell (2008) propose that a light touch, a warm and friendly smile, or direct eye contact enhance service delivery. Videbeck (2004) suggests that a server nonchalantly touching guests when returning change may increase the tip size, and that “squatting at the table, drawing a smiley face on the bill, forecasting good weather, telling a joke, and wearing a flower in your hair” (p. 40) can also influence tipping behavior. Further, Parret (2006) found that servers from the Netherlands who mimicked customers improved tips. The act of mimicking provided a way by which individual servers could build similarity and familiarity with the customers thereby breaking down any perceived formal service barriers. Whether mimicking facial expressions or attitudes, these actions were found to create an over-all influence with guests. 2.3. Tipping and social norms Although not as widely discussed in the literature as other motives, the desire to conform to social norms and/or acquire social approval while avoiding disapproval are increasingly being addressed by the research community as potential motives for tipping (Azar, 2004; Bodvarsson and Gibson, 1997; Fehr and Falk, 2002; Saunders and Lynn, 2010). Tipping can be described as a social norm driven practice. As Fehr and Falk (2002) explained, social condemnation was a primary contributor in the implementation of societal customs. Social norm theory suggests that many individuals tip out of pressure to conform, ultimately striving to avoid feelings of guilt or shame. Where tipping is required for services, consumers generally want to feel positive about giving rather than withholding a tip thereby avoiding those negative feelings associated with not tipping.

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

On the other hand, some believe tipping generously is a way to show off in front of others. This perspective is shared by Lynn and Grassman (1990) as well as Seiter (2007) who posited that consumers utilized impression management theory which shaped and/or molded individuals’ behaviors in order to be viewed favorably. Furthermore, when people desire to be seen as likable, one of the more common self-presentation and self-preservation strategies are ingratiation or mimicry before their peers. Azar (2004) lends further support to this belief by suggesting that “when a norm is costly to follow and people do not derive benefits from following it other than avoiding social disapproval, the norm erodes over time” (p. 749). While this may hold true for a lot of social norms, the practice of tipping has historically defied this assumption. It should be clear though that this is largely culture bound (Lynn, 2000) and the average consumer’s response to any service encounter will be informed by cultural background, identity and whether tipping is actually practiced as a norm in the person’s home country. Closely related to the social approval motivation theory is that pertaining to empathy and/or compassion for service employees. From a historical perspective, Saunders and Lynn (2010) propose that consumers have always tipped service workers out of compassion for what they do. Many consumers believe that the practice of tipping, regardless of service received, is necessary in order to make-up for a perceived wage shortfall on the part of traditionally low-income-earning service employees (Stephen and Zweigenhaft, 1986). In essence they have assumed a responsibility and/or duty of care toward service employees and derive a good feeling from the tips awarded. Whether out of compassion for their wage status or for what they do, it seems that a large proportion of the US population has accepted the fact that service workers need to be adequately compensated for what they do and have willingly embraced this responsibility. Empathy could therefore be a powerful motivating factor to tip especially when others can understand and relate to the disenchantment a server may feel when he or she is having a bad day. Many consumers, and in particular those who have worked in the tipped professions, tip above and beyond the requirement because “they know what it is like to live off of tips so they tend to tip accordingly” (Babcock, 2007, p. 44). Babcock (2007) further explains that from experience there is also a widespread belief among members of the service industry that you should take care of your own. Even in instances of sometimes-unbearable frustrations with poor service, many consumers will tip, no matter the service experience. According to Lynn and Graves (1996), this belief is consistent with Crespi’s (1947) seminal study which concluded that the public tipped more out of social disapproval for workers who received low wages and the fear of social disproval by their peers rather than reward for service. In contrast, equity theory values the balance in exchange relationships and implies that consumers should tip according to the perceived amount and quality of service received. Therefore, the output of a tip should match the input of the service (Videbeck, 2004). This notion firmly supports the common sense rationale that poor service results in menial tips, while good service generally produces large percentage tips. With the purchase of any service, the intangible aspect of the exchange produces the assumption of risk and anxiety. To reduce the amount of risk in the service experience, consumers are willing to tip, in order to receive the expected levels of service and ultimately to avoid anxiety. With few exceptions, most servers are willing to deliver on the promise of service, provided consumers are willing to tip for a waiter’s efforts. Within this interaction the server/guest relationship becomes equitable and one of trust for both participants. Provided both parties behave according to conventional norms this should not only ensure a memorable experience and appropriate reward at the time the service is rendered, but also during any return service encounters. In essence both

123

parties know the rules of the service game and play accordingly. Put simply, equity and trust set the tone for any future consumer/server relationship. 2.4. Tipping and operational processes Operational maintenance of systems is paramount to organizational success. Therefore, the overall experience of a restaurant setting can and sometimes will influence tipping behavior. For example, the overall cleanliness of an operation tends to signal to the guest a level and efficiency of service along with overall efficacy of the operation. Barber and Scarcelli (2009) found that those who visited a restaurant’s toiletry facilities tended to have a less than favorable outlook upon the service (that which was about to be received and/or the service which was momentarily rendered). In addition, a simple operational process such as the general expense of a restaurant can influence tipping behavior. For example, in a study conducted by Israeli and Barkan (2004), the authors considered why consumers tip as a percentage of the bill or a flatrate dollar amount: “this twofold consideration frequently leads to a situation in which tips for low bills exceed the percentage norm and tips for large bills fall short of the percentage norm” (p. 445). This is also known as the magnitude effect. Additionally, while monitoring service, these researchers posited that technical elements of service increase dollar tip amounts, whereas functional elements increased dollar amounts and percentage amounts. Furthermore, the quality of food is a driver of restaurant patronage (Raith, 2008; Tillotson, 2008), and servers who do not have the ability to serve hot-food-hot or cold-food-cold jeopardize the overall efficacy of the operation resulting in poor perceptions of the operation. Parsa et al. (2012) found that food quality may be important to some upscale restaurants; however, this was not the case in fast casual establishments. It is important to note, however, that in the U.S. restaurant market, there are countless fast food operations that do not utilize servers in the delivery of food products, and thus there is no reason to tip in these venues. 3. Methodology This study employed a mixed-methods approach to develop the Tipping Motivation Scale (TMS) and test its psychometric properties. The study included a series of qualitative and quantitative processes over four phases: (1) item generation and item selection using the card sort technique; (2) exploratory factor analysis of TMS-1 items; (3) focus group session and face-to-face interviews on the refined TMS-2 instrument; and (4) scale purification and confirmatory factor analysis of TMS-2. A likert-type scale was created based upon repetitive themes in the literature that may affect consumer-tipping behavior. 3.1. Phase 1 – TMS item generation and item selection Based on a review of literature on tipping, the researcher developed 60 statements on consumer tipping motivation. To establish face validity of the items, which were intended for the TMS, and to reduce the likelihood that the statements did not measure the intended tipping motivational constructs, the card sort technique was conducted with focus group participants. The card sort technique has “proven to be a viable tool to ascertain the individual subject’s perceptions regarding competencies” (Jahrami et al., 2009; p. 176). This phase used a convenience sample of 25 students from a southeastern U.S. university who were hospitality majors employed in a job related to the field of study and who were frequent restaurant customers. The use of a student sample was deemed appropriate, as these populations have been found to be effective in concept identification and construct analysis

124

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

(Emory, 1980). Moreover, the homogeneous nature of student samples allows for more exact theoretical predictions than possible with a heterogeneous group (Calder et al., 1981). Students were divided into three groups to participate in focus group sessions. Initially, all participants in Spring 2009 sorted the 60 statements generated from the literature review and those found to be related were grouped into like constructs for further evaluation in three distinct focus group sessions. The sessions were conducted to re-assess the grouping of the statements, confirm the names of the six constructs revealed and the respective items as well as to provide recommendations for item pool generation. Constructs identified as service, social compliance, server actions, future service concerns, peer pressure, and operational processes were treated in accordance with Spector’s (1992) recommendation that factors be clearly defined before selecting a pool of items to represent the meaning. It is important to note that the initial sixty possible scales items were reduced to thirty-two statements from the card sort technique. Some items seemed to overlap and describe the same construct; therefore, the reduction in items was a direct result of the collaboration with the focus groups. 3.2. TMS instrument design The 32-item scale was the third of three main sections in the TMS survey. Part 1 included demographic questions related to age, gender, ethnicity, college major, as well as employment status and years of experience in the hospitality industry. Based upon the tipping literature, whereas some individuals are thought to tip a certain portion of the bill, no matter the level of service, Part 2 sought information from respondents on current tipping practices and norms based upon percentage of the bill. In Part 3 of the survey, respondents were asked to indicate the level of agreement on the six aforementioned TMS constructs based on a five-point Likerttype scale anchored from strongly disagree (1) to strongly agree (5). The first construct, Service, measured both technical and functional elements of the service experience and included items such as: order taking, order delivery, tangible and intangible aspects, friendliness, personal service, and eye contact. The second construct, Social Compliance, included items related to feelings of shame, guilt, regret, equitable, relationships, and altruism. The third construct, Server Actions, required respondents to agree on how body language, eye contact, tableside manners associated with server squatting, stance and physical contact motivated them to tip. The fourth construct, Future Behavior, explored gratitude, gratuity, and frequency of visit, along with fellowship, social respect and esteem. Peer Pressure, the fifth construct, examined the respondent’s diffusion of responsibility or obligation to others for example, “I feel obligated to tip when dining with friends and/or family.” Lastly, the sixth construct – Operational Processes, examined the effect of the physical environment on tipping, for example, “overall restaurant cleanliness affects my tipping behavior.” 3.3. Phase 2 – Pilot study: exploratory TMS-1 analysis Having established face validity in Phase 1, an exploratory study was conducted on the refined TMS-1 instrument and piloted to a sample of 1000 students on the campus of the same southeastern U.S. University. Eight hundred and thirty-one instruments were completed, yielding a response rate of 83%. Respondents were predominantly female (63%) and Caucasian (83%). Only 13% were directly employed in hospitality and 10% held a tipped position. A majority (98%) indicated that they engaged in tipping behavior with varying norms reported; 41.5% indicated they normally tipped between 11% and 15%, 27.5%, between 16% and 10%, and 4.1%, above

20%, while 37.5% of all respondents tipped above 20% specifically for excellent service. Internal consistency of the 32-item scale was found to be acceptable (˛ = 0.79), exceeding the .70 threshold for exploratory research. Before proceeding with an exploratory factor analysis (EFA) where the goal was to reduce the number of items to a more manageable set, the data were determined to meet the assumptions of multivariate analysis, i.e. normally distributed with no skewness (<3), kurtosis (<10) or multicolinearity (<.85) violations (Kline, 2005). Using the principal components extraction technique with VARIMAX rotation, an EFA was performed in SPSS version 17 with the initial factor solution yielding a KMO statistic of .821 and Bartlett’s test for sphericity at 5819.690. The solution was significant at the .001 level with a Chi-Square value of .000 with 56% of the variance explained by nine components having eigenvalues of at least 1.000. Due to cross-loadings on multiple components, poor commonality statistics (<.400), and low factor loadings (<.400), a series of additional analyses were performed to remove offending items. The results of these tests rendered the data factorable with six components generated consistent with literature and the results of Phase 1 of the study. The resulting six-component structure with 23 items loading cleanly explained 54% of the variance; loadings ranged from .501 to .821. The corresponding reliability co-efficient for this 23-item scale was ˛ = .710 while the solution yielded a KMO of .775, Bartlett’s test for sphericity of 3956.839 and significant Chi-Square (p = .001). The six components were: “Social Pressure” (˛ = 51), “Process Related” (˛ = 51), “Server Action” (˛ = .63), “Future Service” (˛ = .68), “Social Compliance” (˛ = .73), and “Service Received” (˛ = .73). Overall, for an exploratory analysis, the scale reliabilities were satisfactory though scores below .60 are deemed problematic and thus warrant further exploration. 3.4. Phase 3 – TMS-2 refinement: focus group and face-to-face interviews To improve upon the psychometric properties of TMS-1, a refined TMS-2 survey instrument was developed. Based upon additional comments from survey participants of TMS-1, seven items found to have theoretical support as well as practical significance were added to the scale. These included the server’s gender, the timeliness of the service received, quality of food, meal expense, payment method, seating assignment, and the weather were introduced to the scale. After addition of the seven items, TMS-2 was thus subjected to assessment by a focus group comprised of four hospitality management graduate students from the same university; the findings of which were consistent with the previous results. Finally, in accordance with Noar’s (2003) recommendations, and for the establishment of content validity, four subject experts who were careered industry professionals with a minimum of 10 years of experience were interviewed concerning their perceptions on consumer tipping motivations and the TMS-2 instrument; their professions ranged from servers to general managers. Additionally, an exogenous construct was added to the scale to determine the influence of the tipping motivations being assessed on a patron’s overall tipping behavioral intention as well as the predictive power of each of the six constructs. From the review of extant literature, overall tipping behavioral intention can be categorized into six main norms such as: (1) the amount of the tip during the actual visit, (2) same server requests, (3) word-ofmouth recommendation, (4) increased check amount/expenditure, (5) increased tip amount on future visits, and (6) actual future visitation. A variety of academic and industry sources including two university professors and three food and beverage specialists were consulted (via one-on-one interviews) regarding the inclusion of

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

125

Table 1 Respondent profile for main study. Demographic

Characteristics

Frequency

Percent

Gender

Male Female

117 128

47.8 52.2

Age

19–24 25–39 40–59 60+

116 66 52 11

47.3 26.9 21.2 4.5

Annual income

<$15,000 $16,000–39,000 $40,000–69,000 $70,000–99,000 $100,000–129,000 $130,000+

82 57 31 22 15 26

33.5 23.3 12.7 9 6.1 10.6

Employment status

Employed Unemployed Student Retired

163 13 61 7

66.5 5.3 24.9 2.9

Employed in hospitality

Yes No

58 182

23.7 74.3

Hold a tip job

Yes No

36 208

14.7 84.9

these variables in the second pilot. As such, the refined instrument included a total of 30 variables (23 from TMS-1 and 7 new items) as well as 6 items representing tipping behavior intention. As with TMS-1, TMS-2 was anchored on a Likert type scale ranging from 1 (Strongly Disagree) through to 5 (Strongly Agree).

3.5. Phase 4 – Main study: purification and confirmation of the TMS-2 instrument During the months of October and November in 2009, tailgaters at a southeastern US university were administered the TMS-2 instrument. Unlike the pilot study, the targeted sample was comprised of a wider cross section of potential restaurant customers. Demographic questions related to gender, ethnicity, age, city and state of residence, employment status, hospitality industry experience, and household income. A total of 400 surveys were distributed with 245 responses received representing a response rate of just over 61%. Respondents were majority female (52%) and Caucasian (87%), with 25% directly employed in hospitality and 14% holding a tipped position per Table 1. When asked to indicate which percentages represented individuals’ tipping norms, 33% stated a tipping norm between 11% and 15% for normal service. As anticipated, more respondents increased the tipping norm as the quality of service received also increased; specifically, 31% normally tipped between 16% and 20% for good service while almost 40% normally tipped over 20% for excellent service. Approximately 65% tipped at least 16% for excellent service whereas 63% would tip 15% at most for regular service. Table 2 provides more a detailed breakdown of the respondents’ tipping norms.

4. Data analysis The data was analyzed in two stages. First, exploratory factor analyses using principal component method with VARIMAX rotation were conducted in SPSS v 21 on the 30 TMS-2 items and the 6 items representing behavioral intentions respectively to examine the dimensionalities and psychometric properties. The data were deemed appropriate for multivariate analysis with no violations of its assumptions. Next, the relationships of service received, social compliance, server actions, future service concerns, social pressure, operational processes and behavioral intentions were empirically tested using confirmatory factor analysis (CFA).

4.1. Empirical results Using guidelines established in phase 2, the initial EFA conducted on TMS-2 produced a KMO of .787, Bartlett’s test for sphericity of 2265.228 with Chi-square significant at the p = .001 level and nine components explaining 62.5% of the variance. A total of seven items were found to have cross loadings above the .400 minimum and were sequentially removed. This resulted in a 23-item scale with seven components that explained 62% of the variance. The KMO statistic was .768 while Bartlett’s test for sphericity was 1576.288 and the Chi-Square statistic significant at p = .001 level. TMS-2 factor loadings ranged from 0.466 to 0.858 while reliability coefficients for the seven components ranged from 0.58 to 0.81 as shown in Table 3. For the EFA on the behavioral intention items, the resultant factor structure yielded 2 components that explained 72% of the variance with a KMO of .741, Bartlett’s of 638.084 and a significant Chi-Square statistic at the p = .001 level.

Table 2 Respondent tipping norms. Tip percentage

1–5% 6–10% 11–15% 16–20% Over 20% Total

Tip norm

Good service

Excellent service

Freq.

Percent

Freq.

Percent

Freq.

Percent

31 43 80 63 20 237

12.7 17.6 32.7 25.7 8.2 96.7

19 37 66 77 39 238

7.8 15.1 26.9 31.4 15.9 97.1

8 23 51 63 92 237

3.4 9.4 20.8 25.7 38.8 96.7

126

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

Table 3 Tipping motivation scale 2 – final exploratory PCA solution. TMS-2 Survey items and components

Factor loading

SR: Service Received V4: The service received influences my tipping behavior V11: A server’s attitude influences my tipping behavior V10: Poor service influences my tipping behavior V3: Timeliness of service influences my tipping behavior

.812 .782 .774 .727

SC: Social Compliance V22: I feel regret if I do not leave a tip V16: I feel obligated to tip even when service is bad V24: I feel embarrassed when others in my party do not tip V14: I leave a larger tip when others I have dined with do not tip

.788 .711 .696 .572

SA: Server Attentiveness V25: Number of times a server visits my table V15: A server’s ability to sell the menu influences my tipping behavior V21: If a server is caring and empathetic this influences the amount I tip V12: A server’s menu knowledge affects my tipping behavior

.707 .581 .551 .516

FS: Future Service V30: I always consider future visitation when I tip V23: I always consider future service when I tip

.858 .752

SP: Social Pressure V1: A Server’s gender (male/female) influences my tipping behavior V13: I feel more obligated to tip when dining with friends and/or family V5: On occasion, I tip to impress

.778 .729 .624

SS: Server Actionsa V6: A server’s body language influences my tipping behavior V7: Direct eye contact with a server influences my tipping behavior V28: A server making physical contact with me affects my tipping behavior

.790 .750 .466

OP: Operational Processes V17: The general expense of a restaurant influences my tipping behavior V29: Overall restaurant cleanliness affects my tipping behavior V8: The quality of a restaurant’s food influences my tipping behavior

.718 .603 .571

a

Variance explained (%)

Cumulative variance (%)

Cronbach’s alpha

TMS-2

TMS-1

21.680

21.680

0.81

0.73

11.985

33.665

0.72

0.73

7.283

40.948

0.62

N/A

5.711

46.659

0.74

0.68

5.563

52.221

0.60

0.51

4.977

57.198

0.58

0.63

4.583

61.781

0.62

0.51

Server actions was omitted from further analysis due to the low reliability score.

Cronbach’s alpha scores for the two behavioral intention constructs were 0.59 and 0.87 per Table 4. Notably, the 23-item TMS-1 was outperformed by the refined 23-item TMS-2 as the reliability coefficient improved from ˛ = 0.72 to ˛ = 0.82. Additionally, based on the results provided in Table 3, the reliability scores of four components consistent across both scales showed significant improvement. TMS-2 components that were below the criterion of 0.60 based on Peterson’s (1994) recommendation for acceptable reliability were omitted from further analysis, namely Server Actions (˛ = 0.58) and Component 2 of the behavioral intention scale (˛ = 0.59). Service Received, Social Compliance, Server Attentiveness, Future Service, Social Pressure, Operational Processes and Behavioral Intentions were thereby retained for further analysis. An initial first-order CFA was performed on TMS-2 along with the behavioral intentions construct to ensure the data fit the a priori assumptions from the resultant EFA. As such, 27 items were

posited to seven latent constructs. Due to the continuous nature of the data, maximum likelihood estimation procedure along with the covariance matrix method was appropriate for latent structure analysis and convergent validity checks. The model tested in AMOS v. 21 was scrutinized against the fit indices criteria in Fig. 1 as recommended by Kline (2005) and Hair et al. (2010) to determine the adequacy of the model. Overall, the initial first-order CFA produced a poorly fitted model with GFI (0.866), TLI (0.848) and CFI (0.873) indices falling below the recommended 0.90 minimum though having favorable RMSEA (0.064) and SRMR (0.0631) statistics indicating an otherwise acceptable fitting model. Upon closer examination, the modification indices signaled error term correlations for items V3 and V4 posited to Service Received as well as V21 and V22 posited to Behavioral Intentions. Although the standardized residual covariances for all 27 items were within range (<|4|) (Hair et al., 2010) and had statistically significant loadings, items V16 ( = 0.481), V17

Table 4 Behavioral intentions scale – final PCA solution. BI: Behavioral intentions

Factor loading

Component 1 BI2: Recommend the server to friends and family BI3: Ask for the same server on future visits BI4: Spend more money on future visits BI5: Tip more money on future visits

.899 .835 .830 .764

Component 2 BI1: Tip above my norm BI6: Visit a restaurant more often

.852 .786

Variance explained (%)

Cumulative variance (%)

Cronbach’s alpha

53.174

53.174

0.87

19.079

72.246

0.59

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

127

Fig. 1. Initial and final tipping motivation scale. Confirmatory factor analysis models. Note: RMSEA – Root Mean Square Error of Approximation; SRMR – Standardized RMR; GFI – Goodness-of-Fit; TLI – Tucker-Lewis Index; CFI – Comparative Fit Index.

( = 0.465), and V25 ( = 0.400) were found to violate the recommended minimum factor loading score of 0.50 advised by Hair et al. (2010) and were candidates for deletion from the model. Given the poor loading for V25 ( = 0.400) posited to Server Attentiveness, the item was omitted from the model, however items V16 ( = 0.481) and V17 ( = 0.465) were retained due to the strong theoretical support for these items and their relative proximal distance to 0.50. The final first-order CFA model solution produced a better fitting model (2 = 300.776; df = 207; p < .001; 2 /df = 1.453; RMSEA = .043; SRMR = .0609; GFI = .908; TFI = .935; CFI = .947). Service related constructs where shown to have moderately strong correlations as in the case of Service Received and Server Attentiveness (r = 0.467) and Future Service and Server Attentiveness (r = 0.562). The highest correlation between the constructs occurred between Operational Processes and Server Attentiveness (r = 0.617). When examining the relationship between TMS-2 constructs and the latent Behavioral Intention, the strongest correlations were found with Server Attentiveness (r = 0.549) and Future Service (r = 0.530) indicating that each of these constructs, respectively, are potentially more influential than the others. 4.2. Construct validity Convergent validity, which explains how well observed variables posited to a latent construct converge or share a high proportion of variance, is supported by item reliabilities in Table 5, where ˛ values for the final first-order CFA constructs were greater than or equal to 0.60; excluding Server Attentiveness. Given the relaxed nature of Cronbach’s Alpha to test convergent validity, construct reliability (CR > 0.60) and average variance extracted

(AVE > 0.50) were also calculated following guidelines set forth by Fornell and Bookstein (1982) and Hair et al. (2010). As shown in Table 5, CR and AVE estimates ranged from 0.586 to 0.848 and 0.323 to 0.590, respectively. Again, with the exception of Server Attentiveness (0.586) and Social Pressure (0.597), CR values for the remaining constructs were satisfactory; however, AVE values for Server Attentiveness (0.323), Social Pressure (0.332), Operational Processes (0.354), Social Compliance (0.389) and Service Received (0.481) fell below the 0.50 minimum. Therefore, convergent validity of the TMS-2 instrument was marginally supported. Discriminant validity on the other hand provides evidence that each construct can capture its own unique information not obtained from other constructs in the model and where each observed variable is posited to only one construct. An acceptable measure of discriminant validity is the correlation between the constructs in the model where correlates greater than .850 (Kline, 2005) would indicate definitional overlap of the concepts. As stated earlier, the highest correlation was r = 0.617 between Operational Processes and Server Attentiveness. While there are no issues with discriminant validity, the sub-par AVE values warrant further investigation though item reliabilities and construct reliabilities were, for the most part, satisfactory. The presence of the Behavioral Intention construct in the model was to establish the nomological validity of the TMS-2 instrument. For nomological validity, the behavior of the latent constructs of interest is investigated on the basis of their theoretical relationships with each other (Netemeyer et al., 2003). To corroborate the existence of nomological validity, the constructs should possess distinct antecedent causes and consequential effects and/or modifying conditions (Netemeyer et al., 2003), support for which is

128

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

Table 5 Convergent validity of TMS-2. TMS-2 constructs

Items

FLa

Error

Cronbach’s alpha

CR

AVE

Service Received

V3 V4 V10 V11

0.575 0.653 0.714 0.811

0.669 0.574 0.490 0.342

0.81

0.785

0.481

Social Compliance

V14 V16 V22 V24

0.634 0.481 0.639 0.717

0.598 0.769 0.592 0.486

0.72

0.714

0.389

Social Pressure

V1 V5 V13

0.513 0.617 0.593

0.737 0.619 0.648

0.60

0.597

0.332

Future Service

V23 V30

0.754 0.773

0.431 0.402

0.74

0.780

0.543

Operational Processes

V8 V17 V29

0.680 0.465 0.662

0.538 0.784 0.562

0.62

0.682

0.354

Server Attentiveness

V12 V15 V21

0.548 0.647 0.501

0.700 0.581 0.749

0.58

0.586

0.323

Behavioral Intention

BI2 BI3 BI4 BI5

0.577 0.701 0.930 0.819

0.667 0.509 0.135 0.329

0.87

0.848

0.590

Note: FL – Factor Loadings; CR – Construct Reliability; AVE – Average Variance Extracted. a All standardized factor loadings are significant at the .001 level.

typically established using structural equation modeling (SEM). However, due to reservations with the AVE values, SEM analysis was foregone with a goal to improve on the constructs of concern in future research. Particularly, Server Attentiveness, which was shown to have a potentially strong influence on Behavioral Intention, should be further researched and more items with theoretical support added to improve on the construct’s psychometric properties. 5. Discussion and conclusions The main objective of this study was to develop and test a quantitative scale to measure tipping motivations from the perspective of restaurant consumers. The study makes a significant contribution to the extant literature on tipping as it advances the understanding of the nature of tipping motivations from a theoretical development perspective while also revealing managerial implications. Results from the exploratory analysis suggest that motivations to tip are multi-dimensional in nature with as many as seven underlying components confirmed in the analysis of the TMS-2. The structural components common across both TMS-1 and TMS-2 showed improved reliability coefficients with the exception of Server Actions which decreased and Social Compliance which remained stable. Results of the exploratory PCA across TMS-1 and TMS-2 provide empirical support for Server Actions as a motivating factor in the context of restaurant service. Server Actions such as physical contact with the customer or “touch” have been found to be a motivating factor of tipping behavior in other studies (Jewell, 2008). The results also show empirical support for Social Compliance and Social Pressure as dimensions of tipping motivations and are consistent with the extant literature that social norms drive tipping behavior (Azar, 2004, 2005, 2007a,b; Conlin et al., 2003; Lynn and Grassman, 1990). Similarly, items measuring Service Received produced steadily increasing reliability coefficients (where ˛ = .73 in TMS-1 and ˛ = .81 in TMS-2) and accounted for the largest variance explained supporting the notion that the quality of the service

performed is the most common motivator to tip (Azar, 2003, 2004, 2005; Babcock, 2007; Bodvarsson and Gibson, 1999; Bodvarsson et al., 2003; Lynn and Grassman, 1990). The implications of the TMS in practice are profound. It is important for all servers to recognize that by applying the fundamentals of technical service in their actions, their likelihood of receiving a tip will increase. These service fundamentals would include for example: the accuracy and delivery of the food and/or beverage order, the amount of visits to the guest’s table to ensure satisfaction, and the removal of used plateware. To further illustrate, guests in a fast-casual dining environment may require more table visits than those guests in the slower-paced, intimate fine dining atmosphere. Regardless of the setting, managers should train servers to perform authentic and attentive service as a majority of consumers genuinely appreciate and look forward to being served in this manner. From a managerial perspective, motivations to tip can significantly influence the performance, and ultimately livelihood, of restaurant servers. The frequency and value of tips received provides a baseline measure for restaurant servers on how to perform higher levels of service. For example, restaurant managers can utilize the TMS to better educate their servers on the perceived importance of Future Service and Server Attentiveness as factors restaurant customers take into consideration when deciding whether or not to tip. As it relates to the Operational Processes, restaurant customers may often times be motivated not to tip as a way to signal their disapproval of deficiencies in the service delivery system or with poorly managed facilities which can present significant hurdles for servers to overcome during a customer’s dining experience. Failures in the design of service delivery systems may be attributed to overwhelmed wait staff serving multiple tables which can lead to a reduction in server attentiveness and slow service. Moreover, the level of organization and competency of those working in the “back-of-the-house” can present major challenges to the efficiency with which restaurant customers are served. Operational failures may also extend to the appearance and upkeep of a restaurant’s

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

facilities such as un-manicured landscaping, soiled carpets, dirty restrooms, inadequate seating or lighting, disorganization and apathy of the wait staff, and/or failure to greet and seat the guest in a timely manner.

6. Research limitations and future research There are several limitations of the study related to the sampling techniques employed at various stages of the research and with the applicability of the results to other service settings thereby limiting generalizability. First, during the initial phases of the TMS development, survey administration, and data collection, the study incorporated the use of convenience samples of undergraduate and graduate students studying hospitality at a US university. Although not randomly selected, the student samples were deemed appropriate for concept identification, construct analysis, and theoretical predictions which were important to establishing reliability and validity of the scale. Moreover, students were employed in a job related to the field of study and were frequent restaurant customers. Interviews were also conducted with experienced hospitality industry professionals to establish content validity of the TMS items. Collectively, sample selection measures employed have minimized the effects of the limitations on the study’s inferences. Second, it should be noted that results from this study are limited in terms of their applicability beyond the restaurant industry. Although similarities exist between hospitality and other service environments where tipping is also practiced, the TMS was only administered to restaurant customers therefore the results can only be inferred in this context. The scale does however provide a starting point for investigating these motivations in other service settings which, if replicated successfully, would establish crossvalidity of the instrument and point to the universality of tipping motivations when applied to the services industry. Third, the study is further limited in its focus on the development of a tipping motivations scale and therefore only examined the underlying structure of those components deemed to be antecedents of restaurant tipping behavior. As such, the effects of sample demographics on the motivations to tip were not investigated nor were group differences examined. Finally, given the poor AVE scores for Social Pressure, Operational Processes, Social Compliance and Service Received, further testing of TMS-2 was limited to a confirmatory analysis. As convergent validity is of concern, suggesting the need for further development of the scale, motivational constructs were not tested for their predictive capabilities. These limiting factors do however present several opportunities for continued exploration of the tipping motivations scale in future research studies. Further refinement and testing with a randomly drawn sample across various service settings would improve the reliability and generalizability of the instrument as well as the convergent validity of the scale components. In this vein, special attention should be given when refining the battery of items posited to measure the TMS components. As scale components currently discriminate well against each other, a stronger and larger pool of items for those under performing components should be drawn from the literature and retested to establish overall construct validity. Additionally, testing the psychometric properties of TMS against previously validated external scales such as the Big Five factors of personality (Saucier and Goldberg, 1996) and the Empathy Scale (Carré et al., 2013) should be considered to further establish construct validity. A valid TMS instrument could therefore be used when investigating the effects of gender, ethnicity, socio-economic status, work experience in a service related job, and length of service experience on each component of tipping motivations. Furthermore, a split-sample analysis between respondents with and without work

129

experience in a tipped position could be performed to compare similarities and/or contrast differences within or across samples. Assuming that those with previous or current service experience would be more motivated to tip, future studies could also assess the extent to which tipping other service providers is practiced. The versatility of TMS could be further demonstrated by conducting a comparative analysis of the scale between restaurant customers versus hotel guests to determine if the motives to tip a waiter and the motives to tip a bellhop or housekeeper differ. A similar study could be done to compare tipping motivations between casual and fine dining restaurant customers. Ultimately, investigating tipping motivations are essential to understanding what drives customers to practice tipping as well as what factors influence these motivations. Personal characteristics of customers could indirectly drive tipping behavior above and beyond the direct influence of the service provided, social compliance, operational processes, and future service expectations. Consequently, TMS should be examined as a mediator variable where demographic characteristics (gender, experience, and ethnicity) and psychological factors (such as empathy, mood, and personality) are hypothesized to influence tipping motivations which in turn predicts tipping behavior. Mediation analysis could therefore confirm or disprove that tipping behavior is an emotionally driven practice influenced by the motives identified in this study.

Acknowledgement The authors would like to acknowledge the reviewers for their time spent and the quality of the comments provided which helped to improve this study on restaurant tipping motivations.

References Azar, O.H., 2004. The history of tipping – from sixteenth-century England to the United States in the 1910’s. Journal of Socio-Economics 33 (6), 745–764. Azar, O.H., 2005. Incentives and service quality in the restaurant industry: tippingservice puzzle, Retrieved online from http://mpra.ub.uni-muenchen.de/4457/ Azar, O.H., 2006. Who do we tip and why? An empirical investigation. Applied Economics 37 (16), 1871–1879. Azar, O.H., 2007a. Do people tip strategically, to improve future service? Theory and evidence. Canadian Journal of Economics 40 (2), 515–527. Azar, O.H., 2007b. Why pay extra? Tipping and the importance of social norms and feelings in economic theory. Journal of Socio-Economics 36 (2), 250–265. Babcock, M., 2007. The effects of personal background on tipping behavior, Retrieved online from http://www.deedspublishing.com/objects/ shortStories/Mark%20-%20Tipping%20Paper%20-%20short%20story.pdf Barber, N., Scarcelli, J.M., 2009. Clean restrooms: how important are they to restaurant consumers? Journal Of Foodservice 20 (6), 309–320. Bodvarsson, Ö.B., Gibson, W.A., 1997. Economics and restaurant gratuities: determining tip rates. American Journal Of Economics & Sociology 56 (2), 187–203. Bodvarsson, O., Gibson, W.A., 1999. An economic approach to tips and service quality: results of a survey. Social Science Journal 36 (1), 137. Bodvarsson, O.B., Luksetich, W.A., McDermott, S., 2003. Why do diners tip: rule-ofthumb or valuation of service? Applied Economics 35, 1659–1665. Boyles, W.J., Mounts, W.S., Sowell, C., 2006. Restaurant tipping: free riding, social acceptance, and gender differences. Journal of Applied Social Psychology 34 (12), 2616–2625. Calder, B.J., Phillips, L.W., Tybout, A., 1981. Designing research for application. Journal of Consumer Research 8 (1), 197–217. Carré, A., Stefaniak, N., D’Ambrosio, F., Bensalah, L., Besche-Richard, C., 2013. The Basic Empathy Scale in Adults (BES-A): factor structure of a revised form. Psychological Assessment 25 (3), 679–691. Conlin, M., Lynn, M., O’Donoghue, T., 2003. The norm of restaurant tipping. Journal of Economic Behavior and Organization 52, 297–321. Crespi, L.P., 1947. The implications of tipping in America. Public Opinion Quarterly 11, 424–435. Emory, C., 1980. Business Research Methods Revised Edition. Irwin-Dorsey Limited, Homewood, Il. Fehr, E., Falk, A., 2002. Psychological foundations of incentives. European Economic Review 46 (4/5), 687–724. Fornell, C., Bookstein, F.L., 1982. Two structural equation models: LISREL and PLS applied to consumer exit-voice theory. Journal of Marketing Research 19 (4), 440–452.

130

J.E. Whaley et al. / International Journal of Hospitality Management 37 (2014) 121–130

Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2010. Multivariate Data Analysis, 7th ed. Prentice Hall, Upper Saddle River, NJ. Israeli, A.A., Barkan, R., 2004. Developing a framework for rewards in combined production/service businesses: the case of tipping in the restaurant industry. International Journal of Service Industry Management 15 (5), 444–459. Jahrami, H., Marnoch, G., Gray, A.M., 2009. Use of card sort methodology in the testing of a clinical leadership competencies model. Health Services Management Resources 22 (4), 176–183. Jewell, C.N., 2008. Factors influencing tipping behavior in a restaurant. Psi Chi Journal of Undergraduate Research 13 (1), 38–48. Kline, R.B., 2005. Principles and practice of structural equation modeling, 2nd ed. The Guilford Press, Spring Street, NY. Lynn, M., 2000. National personality and tipping customs. Personality and Individual Differences 28 (2), 395–404. Lynn, M., 2001. Restaurant tipping and service quality: a tenuous relationship. Cornell Hotel and Restaurant Administration Quarterly 42 (14), 14–20. Lynn, M., 2003. Tip levels and service: an update, extension, and reconciliation. Cornell Hotel and Restaurant Administration Quarterly 44 (5/6), 139–148. Lynn, M., Grassman, A., 1990. Restaurant tipping: an examination of three rational explanations. Journal of Economic Psychology 11, 169–181. Lynn, M., Graves, J., 1996. Tipping: an incentive/reward for service? Journal of Hospitality and Tourism Research 20 (1), 1–14. Lynn, M., Zinkhan, G.M., Harris, J., 1993. Consumer tipping: a cross-country study. Journal of Consumer Research 20, 478–488. Netemeyer, R.G., Bearden, W.O., Sharma, S., 2003. Scaling Procedures: Issues and Applications. Sage Publications, Thousand Oaks, CA. Noar, S.M., 2003. The role of structural equation modeling in scale development. Structural Equation Modeling 10 (4), 622–647. O’Neill, M.A., 2000. The measurement of service quality in the hospitality industry. In: Kandampully, J., Sparks, B., Moks, C. (Eds.), Service Quality and Customer Satisfaction in the Hospitality Industry. John Wiley, New York, pp. 159–186.

Parret, M., 2006. An analysis of the determinants of tipping behavior: a laboratory experiment and evidence from restaurant tipping. Southern Economic Journal 73 (2), 489–514. Parsa, H.G., Gregory, A., Self, J.T., Dutta, K., 2012. Consumer behavior in restaurants: assessing the importance of restaurant attributes in consumer patronage and willingness to pay. Journal of Services Research 12 (2), 29–56. Peterson, R., 1994. A meta-analysis of Cronbach’s coefficient alpha. Journal of Consumer Research 21 (2), 381–391. Raith, M., 2008. Specific knowledge and performance measurement’. RAND Journal of Economics 39 (4), 1059–1079. Saucier, G., Goldberg, L.R., 1996. The language of personality: lexical perspectives on the five-factor model. In: Wiggins, J.S. (Ed.), The Five-Factor Model of Personality: Theoretical Perspectives. Guilford, New York, NY, pp. 21–50. Saunders, S.G., Lynn, M., 2010. Why tip? An empirical test of motivations for tipping car guards. Journal of Economic Psychology 31, 106–113. Segrave, K., 1998. Tipping: An American Social History of Gratuities. McFarland & Company, Jefferson, NC. Seiter, J.S., 2007. Ingratiation and Gratuity: the effect of complimenting customers on tipping behavior in restaurants. Journal of Applied Social Psychology 37 (3), 478–485. Shamir, B., 1984. Between gratitude and gratuity: an analysis of tipping. Annals of Tourism Research 11, 59–78. Spector, P.A., 1992. Summated Rating Scale Construction: An Introduction. Sage Publications, Newbury Park, CA. Speer, T.L., 1997. The give and take of tipping. American Demographics 19 (2), 51. Stephen, R., Zweigenhaft, R.L., 1986. The effect of tipping of waitress touching male and female customers. Journal of Social Psychology 126 (1), 141. Strohmetz, D.B., Rind, B., 2001. The impact of tipping recommendations on tip levels. Cornell Hotel & Restaurant Administration Quarterly 42 (3), 71. Tillotson, J.E., 2008. Fast food-Ray Kroc and the dawning of the age of McDonald’s. Nutrition Today 43 (3), 107–113. Videbeck, S., 2004. The economics and etiquette of tipping. Policy 20 (4), 38–41.