IMPROVING EFFICIENCY OF CALORIE LABELS: USING NUDGES TO OVERCOME STRATEGIC IGNORANCE

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: TERMINATED
• Funding Source: AFRI COMPETITIVE GRANT
• Reporting Frequency: Annual
• Accession No.: 1012176
• Grant No.: 2017-67023-26232
• Project No.: WYO-00632-17
• Proposal No.: 2016-09907
• Program Code: A1641
• Project Start Date: Jul 1, 2017
• Project End Date: Jun 30, 2019
• Grant Year: 2017

Project Director
Thunstrom, L.

Recipient Organization
UNIVERSITY OF WYOMING
1000 E UNIVERSITY AVE DEPARTMENT 3434
LARAMIE,WY 82071-2000

Performing Department
Economics and Finance

Non Technical Summary
The goal of this project is to increase effectiveness of calorie labels at changing consumption, and improve the welfare of such labels. Salient labels that convey nutritional information are used to encourage healthier food choices. For instance, the FDA has finalized a law on mandatory menu labelling for chain restaurants. For rational consumers, such costless information should lead to healthier food choices. However, field evidence shows calorie menu labels have little or no effect on food choices. This is likely caused by consumers having private incentives to avoid calorie information since it comes at a psychological cost (emotional discomfort). To increase the effectiveness of salience calorie labels, while reducing the "emotional tax" they may cause, it is important to reduce psychological costs to calorie information. In this seed grant project, we examine the impact of four nudges that, based on theoretical and empirical findings, may be successful in enhancing effectiveness of calorie labels. We will use a controlled laboratory environment to implement nudges that (i) increase self-control, (ii) increase self-efficacy, (iii) ask consumers to consider that they underestimate calories, (iv) inform consumers of strategic ignorance of calorie labels. The knowledge gained from the laboratory in this seed grant project will serve as a basis for an application in which we propose to take the most efficient nudges to test in the field.

Animal Health Component

0%

Research Effort Categories

Basic

40%

Applied

30%

Developmental

30%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
607	7310	3010	50%
607	5010	3010	50%

Knowledge Area
607 - Consumer Economics;

Subject Of Investigation
7310 - Experimental design and statistical methods; 5010 - Food;

Field Of Science
3010 - Economics;

Keywords

calorie consumption

calorie labeling

salience nudges

strategic ignorance

Goals / Objectives
The goal of this project is to increase the effectiveness of calorie salience nudges at changing consumption.To reach this goal, we identify 'supporting' nudges that reduce incentives to ignore calorie information, in ways that improve calorie labels' effect on social welfare.

Project Methods
We will examine the impact of the proposed 'supporting' nudges in an economic laboratory experiment, building on the design of our experiments in previous related studies. These previous experiments, which tested our theory of strategic self-ignorance empirically, used two restaurant-style ready meals: a beef and a chicken meal. The meals differed substantially in calorie content: one contained around 500 calories, the other around 900 (about the same as the average fast-food meal; see Dumanovsky et al., 2009). Nevertheless, when subjects in a focus group were told these numbers--as were the subjects in the actual experiments--they were unable to tell which meal was low and which was high calorie. These features ensured that the calorie information we offered in the experiment, namely which meal was which, was both important and instrumental. That is, subjects knew that the calorie difference was large enough to matter, yet were unable to infer the calories of the specific meal they consumed without our information.All subjects in the control group were told which meal was which, while subjects in the treatment group could choose to either take or ignore this information. Subjects were asked to finish their meals on site, and their calorie consumption was measured by weighing the leftovers. We found that a substantial number of subjects engaged in strategic self-ignorance--they consumed significantly more calories under self-selected ignorance than did corresponding subjects in the exogenously informed control group.Whereas our previous experiments used Scandinavian subjects, our proposed experiment will use U.S. subjects recruited in Laramie, Wyoming. We will again use a design with two meals (one high calorie, here: 1000 calorie, and one low calorie, here: 500 calorie), prepared by the University of Wyoming Catering department in collaboration with their in-house nutritionists (contact person: David Asmuth, executive chef; see attached letter of intent). Subjects will again be told up front how many calories the two meals contain, but not which meal is which.In this experiment, we also propose an additional method to examine if treatment-group subjects who choose to remain ignorant do so strategically, i.e., as an excuse to consume more than they would if they were informed. In Thunström et al. (2016a), we compared the density distributions of calorie consumption levels across the treatment and control groups. In Nordström et al. (2016), we also tried, as an alternative method, asking subjects in the exogenously informed (ignorant) control groups if they would have ignored (taken) information had they been given that option. However, we found evidence of substantial hypothetical bias in the answers, forcing us to again rely on the method used in Thunström et al. (2016a). In this project, we will add a treatment in which we give information to subjects who state that they would want none (treatment group 2 below). Doing so will allow us to compare these subjects' consumption to that of subjects in the first treatment group who prefer ignorance and are not given information.Specifically, we will have the following treatments, involving 630 subjects overall:Control group: exogenously informed (60 subjects)Treatment group 1: offered a choice of ignoring or taking information, and then informed endogenously, in accordance with their choice (95 subjects)Treatment group 2: same as Treatment group 1, but informed exogenously (95 subjects)Treatment groups 3-6: same as Treatment group 1, but after being given Nudges 1-4 (95 subjects in each group = 380 subjects)To measure the welfare effects of calorie information, we will follow Allcott and Kessler (2015), who asked for subjects' hypothetical willingness-to-pay (WTP) for information and allowed for negative values. We will measure self-control using the scale developed by Tangney et al. (2004) and self-efficacy using the Weight Efficacy Life-Style Questionnaire (WEL) developed by Clark et al. (1991). The WEL focuses on eating-related self-efficacy, and has been widely used as a measure of such self-efficacy. We will survey subjects for risk preferences, interest in healthy eating, diet-related health, and general background characteristics.Results will be analyzed using both regression techniques and non-parametric methods (Kolmogorov-Smirnov tests for equality of distributions, kernel density estimates, lowess smooths, etc.)

Progress 07/01/17 to 06/30/19

Outputs
Target Audience:This final year of the project has been busy in disseminating results from the project. We have presented (or will present) results from the project at the following academic conferences and seminars: "Reference-Dependent Preferences for Information" AEA Annual Meeting 2020 American Economic Association San Diego, CA. January 2020. "Reference-Dependent Preferences for Information" North American Economic Science Association Conference 2019 Economic Science Association Los Angeles, CA. October 2019. "Reference-Dependent Preferences for Information" Department of Agricultural and Resource Economics, Colorado State University Fort Collins, CO. September 2019. "Reference-Dependent Preferences for Information" Department of Economics, University of Chicago. Chicago, IL. May 2019. "Nudging Reloaded: Increasing the value of information" North American Economic Science Association Conference 2018 Economic Science Association Antigua, Guatemala. October 2018. "Nudging Reloaded: Supplementing a Nudge with Other Nudges" Graduate Student Group Seminar Department of Economics, University of Wyoming Laramie, WY.,USA. October 2018. The first academic article from the project is published, open access, thereby being available both to the public and academic community: Thunström, L. (2019). Welfare effects of nudges: The emotional tax of calorie menu labeling. Judgment and Decision Making, 14(1), 11. https://www.sas.upenn.edu/~baron/journal/18/18829/jdm18829.pdf Further, the PI (Dr Linda Thunstrom) was asked to write a blog post (Feb 15, 2019) for behavioraleconomics.com about distributional effects of nudges. Behavioraleconomics.com has 38,000 members and (from their website) "Behavioraleconomics.com is home tothe largest online behavioral science network (Behavioral Economics Group), the popular Behavioral Economics Guide,and a behavioral science blog that publishes the latest ideasand practicesfrom academia, business, and public policy." The blog post was more braodly about the effects of nudges, but draws on our findings in this project. Hence, the results from this project have been commmunicated to the academic community, and partly to a wider community of researchers, civil servants, and others interested in the effects of nudges. Once the two remaining WPs are finished, we will be able to communicate the results to an even wider audience, directly involved in the policy making process. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project has been instrumental to the Ph.D. dissertation of our graduate student Tabare Capitan, who will include two scientific articles from the project in his dissertation. He has learned how to design economic experiments to answer important policy questions, how to practically conduct the experiments and how to analyze experimental data. The entire research group has learned about new statistical methods for data analysis and gained novel insights into how calorie salience nudges are valued by consumers, impact consumer food choices, as well as can be made more efficient. How have the results been disseminated to communities of interest?Results from the project have been communicated, or will be communicated (have been accepted for presentation) at the following conferences: "Reference-Dependent Preferences for Information" AEA Annual Meeting 2020 American Economic Association San Diego, CA. January 2020. "Reference-Dependent Preferences for Information" North American Economic Science Association Conference 2019 Economic Science Association Los Angeles, CA. October 2019. "Reference-Dependent Preferences for Information" Department of Agricultural and Resource Economics, Colorado State University Fort Collins, CO. September 2019. "Reference-Dependent Preferences for Information" Department of Economics, University of Chicago. Chicago, IL. May 2019. "Nudging Reloaded: Increasing the value of information" North American Economic Science Association Conference 2018 Economic Science Association Antigua, Guatemala. October 2018. "Nudging Reloaded: Supplementing a Nudge with Other Nudges" Graduate Student Group Seminar Department of Economics, University of Wyoming Laramie, WY.,USA. October 2018. The PI (Dr. Linda Thunstrom) wrote a blog post (Feb 15, 2019) for behavioraleconomics.com about distributional effects of nudges. Behavioraleconomics.com has 38,000 members and (from their website) "Behavioraleconomics.com is home tothe largest online behavioral science network (Behavioral Economics Group), the popular Behavioral Economics Guide,and a behavioral science blog that publishes the latest ideasand practicesfrom academia, business, and public policy." What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? We have achieved the goal of the project: we have empirically identified supporting nudges that help increase the value of calorie salience information/nudges. We did so via generating experimental laboratory data and surveys specifically designed to identify what nudges might be effective, and the magnitude of those effects.

Publications

• Type: Journal Articles Status: Published Year Published: 2019 Citation: Thunstr�m, L. (2019). Welfare effects of nudges: The emotional tax of calorie menu labeling. Judgment and Decision Making, 14(1), 11. https://www.sas.upenn.edu/~baron/journal/18/18829/jdm18829.pdf

Progress 07/01/17 to 06/30/18

Outputs
Target Audience:During this period, we havemainly focused on experimental design, data collection and preliminaryanalysis. Hence, we have not yet started communicating our results (we are scheduled to do so during the coming report period). The audience reached during this period has therefore mainly consisted of fellow academics with whom we have discussed our results. Changes/Problems:We had intended to analyse four supporting nudges in our laboratory experiment, but did only have enough statistical power to analyse two of them (with the budget we had). A compromise was therefore to analyse all four nudges in an online experiment (which is cheaper), which complements the laboratory experiment. In the laboratory, we served meals that participants ate on site. We had planned to have participants chose between2 meals. However, the meal provider ended up being able to compose a more restaurant-like menu (with 4 meals to chose from), at a lower price than they had initially given us. This significantly improved our design, as the experiment became more similar to a real restaurant situation. The exprimental sessions were so complicated to run that it was impossible for our graduate student to run them without extra help. We therefore hired another graduate and an undergraduate student to assist in the laboratory. What opportunities for training and professional development has the project provided?The project has provided invaluable training to our graduate student (Tabare Capitan), in both the policy issues relating to calorie menu labeling and the research methods (laboratory experiment) used to examine the impact of nudges. The latteris true also for the students providing technical assistance in the project. Further, the Co-PIs/PI of this project have also developed valuable new knowledge and skills, needed to design the experiment in this project. Finally, the results provide new knowledge and insightsthat benefit all involved. How have the results been disseminated to communities of interest?As stated before, this report period has been focused on data collection, so we have nothing to report here for this period. What do you plan to do during the next reporting period to accomplish the goals?Analyse the data we have, write up two working papers based on that data, and communicate results on conferences -- to policy makers and academia.

Impacts
What was accomplished under these goals? We have finished a working paper showing that menu labeling has emotional impacts, which might in turn impact consumer responses to such labels. We have collected data to examine how 'supporting' nudges impact incentives to ignore calorie menu labeling. We are therefore close to accomplishing the overall goal of this project, i.e., to be able to answer to what extent supporting nudges reduces ignorance of calorie information on menus.

Publications

• Type: Journal Articles Status: Under Review Year Published: 2018 Citation: Thunstrom, L.(2018) Incidence of an Emotional Tax: The Case of Calorie Menu Labeling, Working paper