This study’s independent variable was a helpful message on the back of the dining check regarding the upcoming specials. The dependent variable was the percentage of the bill left as a tip to the waiter. This study operationalizes the restaurant tipping culture by focusing on the tip percentage. The potential confounding variables of this experiment may be the food quality, the server’s gender, the dining party’s size, and whether the customers were regulars.
The study’s strengths include design transparency, implementation simplicity, and potential replication in different settings. However, the message’s content varied depending on the special, appealing differently to customers. Furthermore, the study was conducted only in one restaurant, which renders extrapolation to different establishment classes difficult. These factors could impact the tipping outcomes and, therefore, the study results.
Furthermore, it is unclear whether the mere presence of a personalized message or the message content itself drove the outcome. To evaluate this possibility, the authors could have asked waiters to write different messages, some with no useful content, such as wishing a good day or drawing a smiley face. Moreover, the researchers could have used alternative arrangements, assigning the waiters randomly instead of assigning dining parties randomly.
Several aspects of this study could impact its internal validity. The researchers have instructed servers to keep their behavior consistent, thus limiting potential confounding variables. Varying attitudes may affect the final tip, thus undermining the causal relationship. Furthermore, a few hypothetical aspects could limit the research’s external validity. For instance, if the waiter were a young female adult, applying the same conclusions to male waiters and older servers would be challenging, given the potential gender factor. If the study was conducted at a private country club, extrapolating the result onto the high-class restaurant and casual diners alike would be unreliable. If the meal was buffet-style, the overall tipping percentage might have been lower due to the perceived lesser effort of a waiter, thus impacting its wider applicability.
Lastly, the study only used two groups, servers, and dining parties, with no size specification. The authors could have employed a third group of solo dining parties, thus separating groups from those dining by themselves. Such an addition would allow addressing the question of the groups’ size effect and individual, interpersonal interactions. Lastly, implementing the design with solo dining parties would address the effect of gender on the outcome.