Tipping Your Server: Better With Cash or Credit

Jay M. Allison
Missouri Western State College
December 5, 1995

Abstract

This study was conducted at the Cabbage Roll Deli at 27th and Lafayette in St. Joseph, Missouri. St. Joseph is located in the northwest corner of Missouri and has a population of about seventy-one thousand citizens. The Cabbage Roll Deli is St. Joseph's only German restaurant which made this experiment unique from other experiments done in chain restaurants. The variables of main interest were whether method of payment would affect the amount that was tipped to a server. Data was collected over a four week period, consuming weekend evenings during October and November 1995. What is meant by method of payment, is whether the patron paid with cash or credit card. What was found was that those who did pay with credit card did tip more than those who paid with cash. An Independent t-test revealed that t(62)= -2.74, p<.05. Another test showed that those who paid with cash did tip a slightly higher percentage of the bill, but an analysis showed this to be insignificant with t(62)=.17, p>.05. Other tests were run on the size on the dining party and on the average price of the bills.


Introduction

Michael Lynn and Bibb Latane (1984) state that tipping is said to have originated in the 18th-century pubs where customers would attach coins to notes to the waiter "To Insure Promptness" thus you get the shorter version (T.I.P.). Upon examination into the issue of tipping, or as it sometimes called "gratuities", I found many different studies pertaining to this concept. The two variables I took into consideration were the number of people dinning in a given party and the way in which the patron's in that party paid. That is, did they pay with cash or credit. I believe that I'm going to find that tip increased based upon the method of payment. I hope to find that in all cases the server got a larger amount of tip from those patrons who paid with a credit card, and a lower amount of tip from those who paid with cash. I also hope to find that the size of the dining party will also affect the amount that is tipped.

Upon examination into this area I found many studies that have been beneficial to my particular study. The other thing I discovered is that although many different experiments have been done, none looked at method of payment and party size. Perhaps the biggest examiner on this issue has been Michael Lynn from Ohio State University. Lynn has done many experiments that have proved to be very helpful with my research in this field. He's looked at things like whether or not server posture or introduction plays a role in tipping, does neuroticism effect tipping, and does alcohol consumption effect levels of tipping. What Lynn has found is that you have to take into consideration all of these and many other different factors before you look at the tip that was left for the server. Lynn (1983), said that their at least two potential explanations for alcohol's effects on tipping. First, according to Lynn(1983) alcohol improves people's moods and this positive affect may increase tipping. Second, Lynn (1983) believes alcohol decreases people's ability to process information and, in situations in which the most salient cues call for helping, this may reduce people's ability to process additional cues that would inhibit the behavior. However in Lynn's research he found that there was no correlation between tip and the amount of alcohol consumed. These findings were inconsistent with several other studies done by Crusco and Wetzel, (1984).

Lynn also examined the posture of the server and their introduction technique, and the effects that they may have on tipping. The results of this study were consistent with a study done by Kimberly Garrity and Douglas Degelman (1990) which also looked at the effects that sever introduction had on restaurant tipping. The major finding in this study was that having the server identify themselves by name resulted in a large statistically significant effect on the amount that was that tipped. This could be due partially do to the fact that this gives off cues as to the class of the restaurant, which in return creates an atmosphere that produces larger tips. Lynn (1993), found that although server posture didn't affect bill size, it did dramatically affect tipping.

The final variable that was examined by Lynn (1983), dealt with whether tipping did evolve as a means of reducing service workers' envy to their customers. Since service workers have some control over the well-being of their customers, it's in the customers' interest to share their good fortunes with servers and thereby avoid servers' envy. Tipping is a means whereby customers can do this. If this theory holds true, tipping should be prevalent in areas with a high amount of neuroticism and lower in areas with low levels of neuroticism.

Numerous other studies examined whether or not tipping was cheaper by the bunch. In a study done by Freeman, Walker, Borden, and Latane (1975) they found that the larger the dining party the smaller the percentage of the bill is left as a tip. Those dining alone tipped nearly nineteen percent of the bill where as groups of four to six only tipped thirteen and a half percent. Donald Elman (1976) further examined this study and felt that not only did you need to look at the number of individuals at the table but also the size of check. He proposed that there was a more reasonable explanation than the diffusion of responsibility notion suggested by Freeman, Walker, Borden, and Latane (1975). Elman felt that the best explanation as to why tipping seemed to be cheaper by the bunch was do to the fact that one or two people could be served easier than six or seven. Thus maybe the quality of service was lower with a large group. Melvin Snyder (1976) disagreed with Elman's notion that two people can be served easier than three or four. Upon interviewing ten waitresses and one waiter he found that servers preferred to wait on one party of four over four parties of one. They stated that the four together could be waited on easier than the four singles. If this is true, larger parties may tip a smaller per cent because what they perceive as fair compensation for the service rendered is smaller fraction of the bill.

Two separate studies done by April Crusco and Christopher Wetzel (1984) and Renee Stephen and Richard Zweigenhaft (1985) looked at whether interpersonal touched effected the amount that was tipped. What was found on both sides was that depending on the type of touch exhibited by the server the amount of tip increased slightly. Crusco and Wetzel (1984) discovered touching a patron on the shoulder did significantly reduce tips compared to the fleeting touch, but both types did increase tipping compared to the no touch condition. As for the sex main effect, males tended to tip more than females when touched. This study by the way was only done with females. The study by Stephen and Zweigenhaft (1985) looked at whether a touch on the palm of the hand, a shoulder touch, or no touch would increase the tip amount. Again only done with females the results were some what identical to the previous study. Although the type of touch played no major role as to the amount of tip left, those who weren't touched tended to tip less than those who were touched. Also another interesting result was that the females who were touched tipped more than the males who were touched.

What I propose is that servers will both get a higher tip from those individuals who pay by means of a credit card a lower tip from those individuals who pay with cash. I alos believe that the size of the dining party will alsi affect the amount that a server is tipped.


Methods

Subjects

The individuals that I used to collect my data from were individuals that dined at the Cabbage Roll Deli. The Cabbage Roll Deli is a German restaurant that is located at 27th and Lafayette in St. Joseph Missouri. St. Joseph is a town located in northwest Missouri, and has population of about seventy-one thousand citizens. Despite St.Joseph's rich German heritage, the Cabbage Roll Deli is St. Joseph's only German restaurant. We seat about eighty people on an extremely busy night. We will serve to anyone who is interested in a good German meal, and we have both dine in service as well as carry out. The menu is fairly reasonable with dinners ranging from $5.99 to $7.49, and we also have a good variety of sandwiches and delicious desserts. I collected data from sixty-four patrons of the restaurant on the weekend evenings between October sixth and November eleventh. We had a dinner buffet on those evenings that was all you can eat buffet and priced at $9.95. The buffet included five or six hot items, soup, salad, all drinks except beer and wine, and all the dessert the patron could eat. Since the majority of my data was obtained during Oktoberfest we were always busy. We had patrons ranging from ages of a month old to one hundred years old, so I always had a wide range of individuals to examine. We also had both males an females with varying educational backgrounds and different environmental backgrounds. I may note that we do get many tourists from all over the country that dine at the restaurant at all times of the year. Finally, many of the people that I collected data from were customers who were regular's in the restaurant and knew their server very well.

Materials

In order to collect my data at work I constructed a set a very rudimentary charts that allowed me to write down everything that was significant to my study and could have possibly had some bearing on my results. At the restaurant we have a total of twenty different tables, twelve being on the new side of the restaurant and eight being on the old side. We also have a beer garden, but at this time of the year we typically don't seat any one out there. I decided to note which side of the restaurant the patrons were seated on and whether or not they ordered buffet or menu service. I also took note of whether or not any alcohol was consumed at the table, and if the patrons used a coupon. As a side note coupons are usually only given to those customer's who are regulars at the restaurant. I took note of these variables because I felt they probably had some bearing on the amount that was tipped, and these were also factors examined in previous research. One other reason I decided to take down all of these variables was do to the fact I still wasn't sure what the second variable of my study was going to be. Again on the chart's that I constructed I was sure to write down the male to female ratio at the table, who paid the bill at the table, whether they paid with cash or credit, and the gender of the server who waited on them. In the end I devised a final chart in Word Perfect. It was divided into three columns with the final three variables being the number of people at the table, method of payment, and the amount left as a tip.

Procedure

The way that I went about collecting may data was done as follows. I constructed a very rough set of chart's that were broken down into the different servers we have at the restaurant. The sex of the server was only noted on the original charts. There is Jay, Heidi, and Andy, the owner of the Cabbage Roll Deli. On the charts I marked the number of people in the party, who paid the bill, whether the bill was paid by cash or credit, and how much was left as a tip. I also took note of the alcohol consumption in the party, location of the table, if the buffet was ordered or if menu service was requested, and finally if a coupon was used. Since I was working at the time most of my data was being collected my chart's were very hard to read and pretty jumbled. My one main area of concern was that I needed at least ten subjects for each condition, and I was afraid I wasn't going to be able to get twenty charges on weekend's alone. Yet in the end I arrived at a large enough sample size that I could analyze my data properly.

After examining the variables that I had recorded I decided to use three of the variables as my conditions: the number of people in the party, their method of payment, and the amount of tip left. I took all the material I had collected and condensed them into three conditions by placing them on Word Perfect. Once in Word Perfect I then broke them down into three separate columns, one column being the number of people in the party, the next being the method of payment, and the third and final column being the amount of tip that was left for the server. I then devised a simple scoring scale for my study. If the patron paid with cash I gave that a one. If the patron paid with a credit card I scored that a two. I had forty-three subjects who paid with cash and only twenty-one who paid with a credit card. I then saved that on a disk so an Independent t-test could be ran on my data at a later date.


Results

The statistical test that I applied to my data was an Independent t-test. The results I obtained after conducting my statistical test proved to be very interesting and they also supported my original hypothesis. One thing the t-test allowed my to do that I previously hadn't thought about, was to also examine the percentage of the bill that was left as a tip.

Like mentioned earlier, I had sixty-four subjects in my study. Forty-three of those subjects paid with cash and only twenty-one paid with a credit card. The average percentage of the bill that was left as a tip by those who paid with cash was 11.03%. The average percentage of the bill that was tipped by those who paid with a credit card was slightly lower, at 10.83%. This was found not to be a significant difference upon running the test with t(62)=.17,p>.05. My t-value was equal to .17 with 62 degrees of freedom and with p being .868 it was greater than the alpha level of .05 making it non-significant. In this instance the standard deviation for paying with cash was 4.74 and the standard deviation of paying with a credit card was 4.535

The next t-test I ran looked at the theory that the patrons who paid with a credit card still tipped more overall than those patrons who paid with cash. Upon looking at the previous results I found those who paid with cash did tip a slightly higher percentage of the bill than those who paid with a credit card. Still the analysis of that data proved the difference not to be significant.

The average amount of the patrons bill who paid with cash was twenty-six dollars and forty-seven cents with a standard deviation of 8.49. The average amount of the patrons bill who paid with a credit card was forty-six dollars and fifty-four cents with a standard deviation of 29.314. An individual who paid with a credit card had a bill that was almost twice that of an individual who used cash to pay their bill. When looking at the results obtained from the test I found that the difference in the bill prices was significant with t(62)=3.14,p<.05. The t value was equal to -3.14 with 62 degrees of freedom and p being .003 was less than the alpha level of .05 making it a significant difference. The difference in the bill prices was largely do the fact that the people who paid with cash average 2 people per party with a standard deviation of .798 and those who paid with credit cards had 4 people per party with a standard deviation of 2.755.

The original question I posed with my study was that those patrons who paid with a credit card would tip a higher amount than those patrons who paid with cash. I found that the average price of the bill's was significantly different from one another. Those who paid with credit card paid almost twice the price to eat than those who used cash as a method of payment. Because of this fact you find that overall the difference in bill price seemed to affect the overall amount that was tipped. The average tip of those who paid with cash was $2.83 with a standard deviation of 1.26. The average tip of those who paid with a credit card was $4.14 with a standard deviation of 2.55. Upon conducting a t-test I found that t(62)=-2.74,p<.05. This was a significant difference since p being equal to .008 made it less than the alpha level of .05.


Discussion

When I posed the theory that method of payment would affect tipping I had no idea that upon running the t-test I'd arrive at the answers I did. I originally felt that those who paid with cash would tend to tip less than those who paid with a credit card. A t-test supported my notion by showing a significant difference between these variables with t(62)=2.74, p<.05. Those who paid with cash did tend to tip less ($2.83) when compared to those who paid with a credit card ($4.14). The thing I found to be so interesting is that although the credit card individual tipped more overall, they did tip a slightly smaller percentage of the bill. However this difference proved not to be significant with t(62)=.17,p>.05. This conclusion could have been partially do to the fact that the average price of a cash person's bill was $26.47 and the credit card person's bill averaged $41.54, which was a significant difference. The size of checks could have been one of the reasons why the two methods of payment affected tip the way they did . The people who paid with a credit card averaged 4 people in their party where as those who paid by means of cash only averaged 2 people. Thus also could have caused the difference in tip amount.

Even though my results did support my original theory, my study was not without its limitations. The biggest problem I believe I faced was that I only conducted my study in one restaurant. If I would have had the time, I would have liked to have gone to about five or six restaurants and conducted my study. Another limitation with my study was that I only used three servers to collect all of my data from. I really feel that in order to do the study properly I'd need to go to a restaurant where there are many servers with many different working shifts. I also was limited to looking at the people who tipped during dinner time hours and who ate off a buffet. To get a good indication of the tipping difference I think I'd need to look at people who eat at restaurants during different times of the day. Another flaw is that by having a limited number of servers and only using one restaurant I lowered the amount of data I collected. I had only twenty-one subjects who paid with credit cards and forty-three who paid with cash. It would have been nice if I could have got an equal distribution of subjects for each payment condition. Perhaps the biggest limitation of my study is that I was unable to control for other possible variables that may have affected my study. There are so many things that affect what a patron of restaurant will tip a server, I couldn't possible control for all of them. Even though I made note of them, I still didn't conduct any test to see their effect on tipping.

My study was a simple one that didn't have many steps to it. For this reason and a number of others I also believe that my study has very good generality. This study, as well as the others I have used for background material, could be done in any culture or society where tipping of servers is a prevalent concept. Many of the studies I examined were not just located in one specific area of the United States. Many times they varied on what variables were being manipulated by the experimenter. In fact a study done by Lynn (1994) looked at tipping behavior as it occurred across the country. So my study is one that could be applied to many different societies and cultures where tipping is done on a regular basis.

Directions for future research is also another thing that I feel my study could be used for. My study, although it has its limitations, could be expanded on and done to the point where it is considered a superb experiment. One thing that would have to be done, would be to control for many of the extraneous variables that could have a bearing on tipping behavior. I took note of some of these variables on my charts. A few of these that I feel would have an effect on tipping would be alcohol consumption, gender of the server, gender of the individual that paid the bill, whether the patron was familiar with the server, and what meal they were eating, be it breakfast, lunch, or dinner. Most of these variables have been examined in numerous other studies, so I decided to look at method of payment, party size, and tip. If I myself or other interested researchers would like to expand on this study and take it a step further, it wouldn't be to difficult to do. There is so much research out there that an interested experimenter would never run out of ideas.

What I found after conducting my research is that many of my findings matched that of early research done in the same field. Now although I didn't investigate the same variables, I could see how other experimental findings matched mine. One such correlation that I really believed to be comparable was the finding of percentage of the bill that was tipped and the party size of those dining. Even though my findings here were not significant, there was a little difference between them, that was worth noting. My findings that people who pay with cash tipped a higher percentage of the bill matched those findings reported by Freeman, Walker, Borden, and Latane (1975). Elman (1976) followed up on the previous study by looking at the size of a party a server is waiting on. Elman (1976) believed a smaller group could be waited on much easier than a larger group, thus you find the difference in tip. Elman (1976) believed a smaller group would tip more than a larger group, based solely on the quality of service. However, my results were different than those of Elman (1976). I found the larger the group the higher the tip, and it was the larger groups that use the credit card to pay their bill. I however did not look at the quality of service.

The other experiments I found in my literature search, just served as basic background material for the study I conducted. They really didn't investigate the same variables that I did, but they helped to give my a starting point for my project. There are so many variables that affect tipping that you could conduct numerous experiments and still find something else that affected tipping.


References

Crusco, A.H. & Wetzel, C.G. (1984). The effects od interpersonal touch on restaurant tipping. Personality and Social Psychology Bulletin, 10, 512-517.
Elman, D. (1976). Why is tipping cheaper by the bunch: Diffusion or just deserts? Personality and Social Psychology Bulletin, 2, 307.
Freeman, S. & Walker, M.R. & Borden, R. & Latane, B. (1975). Diffusion of responsibility and restaurant tipping: cheaper by the bunch. Personality and Social Psychology Bulletin, 1, 584-587.
Garrity, K. & Gegelman D. (1990). Effects of server introduction on restaurant tipping. Journal of Applied Social Psychology, 20, 168-172.
Lynn, M. &. Latane, B. (1983). The psychology of restaurant tipping. Journal of Applied Social Psychology, 14, 549-561.
Lynn, M. (1988). The effects of alcohol on restaurant tipping. Personality and Social Psychology Bulletin, 14, 87-91.
Lynn, M. & Mynier, K. (1993). Effects of server posture on restaurant tipping. Journal of Applied Social Psychology, 23, 678-685.
Lynn, M. (1994). Neuroticism and the prevalence of tipping: A cross country study. Personality and Individual Differences, 17, 137-138.
Snyder, M.L. (1975) The inverse relationship between restaurant party size and tip percentage: Diffusion of responsibility or equity? Personality and Social Psychology Bulletin, 2, 308.
Stephen, Renee & Richard L. Zweigenhaft. The effects on tipping a waitress touching male and female customers. Journal of Social Psychology, 126, 141-142.

Back to MWSU Psychology Research Page
Back to PSY302 Research Page