Would you like some coffee with your sugar? A natural field experiment on the efficiency and acceptability of setting zero sugars as a default in coffee-vending machines

ABSTRACT

This paper aims to replicate the effect of a nudge on behavior (efficiency) and acceptability in a natural field experiment. The nudge in our study consists in setting zero sugars as the default level of sugar in hot drinks–vending machines in a French university. We compared Campus A (default option set to 0 sugars) to Campus B (default option set to 3 sugars). We measured the efficiency of this default option by observing the level of sugar actually chosen by the participants, and we measured acceptability through a questionnaire. We hypothesized a high level of efficiency for the nudge and a higher acceptability in Campus A (default option set to 0 sugars) compared to Campus B (default option set to 3 sugars). Our results show that participants with the default option set to zero sugars (Campus A) consumed less sugar than those with the default option set to 3 sugars (Campus B). We also found a high level of acceptability on both campuses, though with no difference between Campus A (where the nudge was implemented) and Campus B (where a future nudge would be implemented). The discussion addresses the applied perspectives and ethical implications of these results.

KEYWORDS:

• Behavioral Policy
• Nudge
• efficiency
• default option
• acceptability

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/23743603.2023.2214964.

Notes

1. Significant sources of fluctuation include the source of the nudge (e.g. government or employer, Arad & Rubinstein, 2018), the intention behind the nudge (e.g. reducing cost, health protection, limiting overconsumption, Bang et al., 2018), the target of the nudge (e.g. an individual or a firm, Branson et al., 2012), and the beneficiary of the nudge (e.g. the targeted individual or the society, Hagman et al., 2015), if the nudge relies mostly on conscious or unconscious decision processes (e.g. giving advice or changing the default option; Felsen et al., 2013). Moreover, Sunstein, Reisch, and Kaiser (2018) and Sunstein, Reisch, and Rauber (2018) have shown that individual characteristics (e.g. age, sex, political orientation) are also predictors of an acceptability judgment. It appears, however, that these measures of acceptability are not immune to manipulation. For example, Davidai and Shafir (2018) have shown that nudges are judged less acceptable when they are jointly evaluated with traditional public policy compared to separate evaluation.

2. Data collection was initially planned in 2020, from 1 March to 30 April. We had to significantly postpone data collection because of the COVID-19 crisis, as almost all courses in French universities were taught online during this period.

3. First, three translators translated items from English to French and three other translators back-translated them into English. Second, French items were subjected to context adaptation. Third, 30 French native speakers rated content clarity for the six questions.

4. Participants who did not belong to the university where the study took place could have different sugar consumption habits, a different attitude toward sugar consumption reduction, or a different cultural or academic background. For these reasons, they were considered as outliers as we announced in the preregistration.

5. Analyses including these outliers are available in supplemental materials.

6. Data and outliers’ analyses can be found here: https://osf.io/wt8bm/?view_only=43de441510c046f4ba28c02c35e09135.

7. As recommended by Green and Salkind (2008), we divided the chi-square value by N-1 to obtain an index equivalent to ƞ².

8. There were no non-binary participants in the final Campus A sample. Consequently, we were not able to use this variable with three modalities (Female vs. Male vs. Non-binary). We decided to compare Female to others (Male + Non-binary). We ran the analyses without Non-binary participants and the results were equivalent.

9. As a note of terminology, the expression “dark nudge” is occasionally used to designate nudges aiming to steer people’s behavior in a direction that is not in their best interest. The notion of “sludge” (Thaler, 2018) is also sometimes used to refer to interventions aiming to refrain certain behaviors, in general, against the person’s best interests.

10. Steffel et al. (2016) argue that making a nudge by default explicit does not reduce its efficiency.