Increasing Vegetarian Food Consumption at the AUB Cafeteria

Executive Summary

Nudge Lebanon, in collaboration with students from the American University of Beirut (AUB), replicated an intervention designed by Gregg Sparkman and Gregory M. Walton^[1] to increase the consumption of vegetarian food at the AUB cafeteria using static and dynamic social norms. The intervention did not provide evidence indicating that either static or dynamic social norms yielded the expected behavioral change among customers lunching at the cafeteria. Customers receiving the dynamic social norms treatment reduced their consumption of vegetarian dishes by 6.2 percentage points relative to a control group (p-value = 0.296). Despite that, the intervention offers valuable insights to consider when replicating this study in different contexts.

Policy Challenge

Health concerns and the adverse impact of meat consumption on the environment have triggered a global shift towards plant-based diets, with sales of vegetarian and vegan products rising worldwide. ^[2] In fact, it is estimated that the demand for vegetarian and vegan products increased by 62% between 2013 and 2017.^[3] This shift is particularly important for preserving the environment, reducing greenhouse gas emissions, one of the leading factors of global warming.

In Lebanon, a survey of 3307 students from public and private universities, revealed a growing trend towards the consumption of plant-based food, with a majority of the study participants either adopting a vegetarian/low-calorie diet (23.8%, N = 788) or embracing a diet with a high concentration of plant-based foods (31.5%, N = 1042).^[4]

To further encourage the adoption of plant-based foods, Nudge Lebanon, in collaboration with students from the AUB, designed an intervention to increase consumption of vegetarian food at a the AUB cafeteria using an innovative technique known as dynamic social norms – the provision of information about how other people’s behavior is changing over time^[5]

Intervention Design

The intervention was implemented during lunchtime on weekdays at the AUB Cafeteria, and lasted for five days. A total of 321 customers waiting in the line to be served lunch were randomly assigned to one of three surveys on consumers’ opinion towards a trending issue, in exchange for a $1 voucher which can be redeemed at the cafeteria.

• Survey 1 (dynamic social norms group, N = 109): “Recent research has shown that Lebanese consumers are now starting to limit their meat consumption. In fact, meat consumption in Lebanon decreased by 30% in recent years. This means people have changed their behavior and started to eat a third less meat than they otherwise would. What do you think?”.
• Survey 2 (static social norms group, N = 106): “Recent research has shown that Lebanese consumers limit their meat consumption. In fact, meat consumption in Lebanon decreased by 30% in recent years. This means people eat a third less meat than they otherwise would. What do you think?”.
• Survey 3 (control group, N = 106): “Recent research has shown that more Lebanese University students are now starting to enroll in exchange programs abroad. In fact, some exchange programs have grown up by 400% since 2010. This means that university students have changed their behavior and have started to spend more time studying abroad than they otherwise would. What do you think?”.

The purpose of the Dynamic Social Norms and Static Norms surveys was to trigger thoughts among the participants about the decreasing trend in the consumption of meat in Lebanon, and therefore prime them to order vegetarian-based dishes.

Results

Customer receipts were collected by the team on a daily basis. Tracking of receipts was facilitated using color coded vouchers that identified the three different conditions. The cafeteria staff were instructed to staple the voucher to the relevant receipt. The outcome of interest is a binary variable that takes the value of 1 if a vegetarian dish was ordered and 0 otherwise.

Overall, the results revealed a slight increase of 0.5 percentage points in the consumption of vegetarian dishes among the static norms groups compared to the control group [27.4% in the S.S.N group vs 26.9% in the control group consumed at least one vegetarian dish]. However, this difference was not statistically significant (p-value = 0.944).

Interestingly and contrary to our expectations, the consumption of vegetarian dishes among the dynamic social norms group decreased by 6.2 percentage points relative to the control group [20.8% in the D.S.N group vs 26.9% in the control group consumed at least one vegetarian dish]. This difference, however, was also statistically insignificant (p-value = 0.296).

Figure 1 – Proportion of customers ordering at least one vegetarian dish (N = 321)

One consideration is that some individuals may have misunderstood the definition of meat-based product as solely red meat. Therefore, a series of subsidiary analyses took place whereby the dependent variable was replaced with combinations of vegetarian, white meat, and fish products. All analysis yielded statistically insignificant results for both static and dynamic social norms relative to the control.

Conclusion

While the intervention did not yield significant effects, it still offers valuable insights to consider when replicating this study in different contexts. For one, it may be that the choice of vegetarian options at the AUB cafeteria were not appealing enough (and were very limited) compared to the non-vegetarian options. A similar experiment implemented at a local restaurant known for serving a variety of vegetarian options among other things yielded positive and significant results in favor of the dynamic social norms condition.

Future replications should take into account the availability of appealing vegetarian options, as well as the type of clientele (e.g. health cautious or environmentally-friendly customers) participating in the study.

[1] The intervention which was implemented at the Stanford University cafeteria used dynamic and static social norms (D.S.N: Americans are starting to limit their meat consumption vs S.S.N: Americans are starting to limit their meat consumption) resulting in a reduction of meat consumption by 34% among the D.S.N group and 17% among the S.S.N group compared to a control.

[2] Park, R. & Barker, J. (2019). The Meat of the Problem. Behavioural Insights Team.

[3] Malochleb, M. (2018). How Environment alters sensory properties; Plant-based product claims flourish. Food Technology Magazine.

[4] Salameh, P., Jomaa, L., Issa, C., Farhat, G., Salamé, J., Zeidan, N., … & Lebanese National Conference for Health in University Research Group. (2014). Assessment of dietary intake patterns and their correlates among university students in Lebanon. Frontiers in public health, 2, 185.

[5] Sparkman, G., & Walton, G. M. (2017). Dynamic norms promote sustainable behavior, even if it is counter normative. Psychological science, 28(11), 1663-1674.