Intermittent Incentives to Encourage Exercising in the Long Run

Last registered on May 20, 2020

Pre-Trial

Trial Information

General Information

Title
Intermittent Incentives to Encourage Exercising in the Long Run
RCT ID
AEARCTR-0005873
Initial registration date
May 20, 2020

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
May 20, 2020, 10:47 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Region

Primary Investigator

Affiliation

Other Primary Investigator(s)

PI Affiliation
Rady school of management, UCSD
PI Affiliation
Coller school of management, Tel-Aviv University

Additional Trial Information

Status
Completed
Start date
2017-11-23
End date
2019-06-30
Secondary IDs
Abstract
We report the results of a field experiment in which students are encouraged to exercise. We compare a no-incentive control with a treatment in which participants are paid per visit, and to two intermittent incentive schemes: monetary rewards at increasing intervals and monetary rewards with unpredictable timing. The advantage of irregular incentives, as already recognized in the psychology literature, is that they lead individuals to hold two contradictory expectations simultaneously: that of being rewarded and that of not being rewarded. Individuals continue to pursue the reinforcement because the anticipation of a reward is greater than the anticipation of no reward. With repeated experience, they learn to live with the paradox and to cope with the frustration of sometimes not being rewarded for a proper response. This mechanism may facilitate the maintenance of the habit over the long run, after the incentives have been removed. In line with this theory, we find the two intermittent incentive schemes increased participants’ frequency of exercising after the incentives were stopped and over an extended period of time more than the per-visit scheme.

Registration Citation

Citation
Arad, Ayala, Uri Gneezy and Eli Mograbi. 2020. "Intermittent Incentives to Encourage Exercising in the Long Run." AEA RCT Registry. May 20. https://doi.org/10.1257/rct.5873-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The participants consisted of 213 students attending Tel Aviv University who were selected after a pre-screening process. They were offered the chance to participate in an experiment aimed at increasing their physical activity. Participants received a six-month membership (January-June 2018) to Tel Aviv University sports center. Financial incentives for exercising were provided only for the first two months (January-February 2018), allowing us to observe whether the effects of the incentives persisted during the subsequent four months. We also observed whether participants continued to exercise at two six-month intervals, after the free membership had ended (12 and 18 months after the start of the experiment).
To facilitate comparison, we set the mean reward per visit to the gym at 20 NIS for the three incentivized treatments.
We implemented the following incentive schemes:
1. Control: No monetary incentives were given to visit the gym.
2. Per-visit: Participants received 20 NIS for each visit to the gym.
3. Increasing: The number of visits required to receive the next reward and the size of the next reward increased after the payment of each reward. A reward of 20 NIS was paid after the first visit, a reward of 40 NIS after two additional visits (i.e., after the third visit), a reward of 60 NIS after three additional visits (i.e., after the sixth visit), and so on. Participants were aware of this setup, and information on the timing of the next reward and its size was available to them from the mobile app. Under this incentive structure, participants can more easily receive a reward early on, when the need to compensate for the high initial costs of exercising is bigger. Over time, the reward is paid less frequently, and participants gradually get used to visiting the gym without a reward. Thus, the incentives are frequent enough to induce the initiation of the exercise habit, but the habit will not become strongly associated with receiving a reward, due to its decreasing frequency.
4. Unexpected: Participants received 50 NIS after X visits to the gym, where X is a number between 1 and 4 and determined randomly by lottery. After a participant in this group received a reward, a new random number was drawn, and a new count of visits began. Participants knew only that a number between 1 and 4 had been drawn, but did not know which and therefore could not know when they would receive the reward. Because X is bounded, participants knew they would always get a reward by the fourth visit. On their first visit, these participants received 20 NIS in order to overcome any feelings of uncertainty that they would indeed receive a reward, and to help establish the researchers’ credibility among the students. After that visit, X was chosen by lottery.
Intervention Start Date
2018-01-01
Intervention End Date
2018-06-30

Primary Outcomes

Primary Outcomes (end points)
For the membership period: amount of gym visits at incentivized period (months 1-2), non-incentivized period (months 3-6), and the last months of the interventions (months 5-6).
For the follow-up: the amount of time a participant exercised 12 months, and 18 months after the beginning of the experiment
Primary Outcomes (explanation)
Membership period: counting entrances by gym chips. No more than one entrance to the gym per day was counted.
Follow-up period: self report on amount of exercise

Secondary Outcomes

Secondary Outcomes (end points)
We calculated the changes in the physiological and psychological indicators between the introductory session and the concluding session (body fat, weight, pulse, consideration of future consequences, propensity to plan, risk preferences, and happiness), examined how the amount of gym visits affected these changes, and compared them across treatments.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The participants consisted of 213 students attending Tel Aviv University who were selected after a pre-screening process. They were offered the chance to participate in an experiment aimed at increasing their physical activity. Participants received a six-month membership (January-June 2018) to Tel Aviv University sports center. Financial incentives for exercising were provided only for the first two months (January-February 2018), allowing us to observe whether the effects of the incentives persisted during the subsequent four months. We also observed whether participants continued to exercise at two six-month intervals, after the free membership had ended (12 and 18 months after the start of the experiment).
Past research (e.g., Charness and Gneezy, 2009) usually found the most significant effects among participants who did not previously exercise, which guided us in the recruitment. To this end, all university students were invited to answer a short 3-minute lifestyle questionnaire that was distributed online (see the Appendix). To incentivize the students to answer the questionnaire, they were told that 5% of them would be randomly selected to receive 50 NIS (at the time of the experiment, 3.6NIS=$1). The questionnaire included several questions about lifestyle, including filter questions used to determine who was eligible to participate in the experiment.
The students filling out the questionnaire did not know the details of the experiment and its incentives and thus did not have an incentive to lie. We designed the selection criteria to identify individuals who (1) did not exercise at all or exercised only once a week but not in a gym and not swimming, (2) had a commute time of up to 120 minutes from their residence to the university (participants who live closer to the gym can more easily exercise there), and (3) wanted to exercise more. Initially, our criteria were supposed to be stricter, with criterion (1) being not exercising at all, and criterion (2) being a commute time of up to 30 minutes. However, we did not have enough students who answered our stricter criteria. We then sent the selected students invitations to participate. We told them that based on their answers to the lifestyle questionnaire, they were eligible to participate in a research project aimed at encouraging them to exercise in the university gym and that they would be given a free six-month membership (a value of 1,500 NIS).
Participants attended a 90-minute introductory session. The 29 sessions were held between December 17 and December 28, 2017, in the Interactive Decision-Making Lab at Tel Aviv University. Each session was attended by participants from the same treatment in order to minimize the chance of exposure to participants from other treatments. We balanced the treatment assignment according to age, gender, and commute time.
Before being informed about the incentives, participants were asked to sign a consent form to participate in the experiment; 223 agreed to sign the consent form (six participants dropped out at this stage). By signing the form, participants gave consent not only to participate in the study, but also for us to access their future grade transcripts, their Israeli SAT scores, and their matriculation exam scores. After signing the consent form, participants received the experiment’s instructions for the treatment to which they were assigned. The instructions were read aloud to them and fully explained. The instructions included the incentive they would be offered to exercise, how to use the mobile app for the experiment, how rewards would be distributed during the experiment, and so on. Participants also filled out a medical questionnaire, as required by Israeli law, in order to ensure exercising was safe for them.
Participants then answered the following psychological questionnaires: (1) the Propensity to Plan scale (Lynch, Netemeyer, Spiller, and Zammit, 2010), which measures an individual’s tendency to plan, a trait that is typically necessary in order to exercise on a regular basis; (2) CFC—Consideration of Future Consequences (Strathman, Gleicher, Boninger, and Edwards, 1994), which estimates the extent to which participants take into account future consequences, and in this case, those of the decision to exercise; (3) DOSPERT—Domain-Specific Risk-Taking (Blais, and Weber 2006; Weber, Blais, and Betz, 2002), which measures risk tolerance and may be relevant in this context because one of the intermittent incentive schemes in the study involved a certain degree of risk; and (4) a happiness questionnaire, based on a subset of questions from the Oxford Happiness Questionnaire (Hills and Argyle 2002), that measures differences in levels of happiness between people who exercise and people who do not. Participants also answered additional questions about general lifestyle. While participants were answering the questionnaire, they were taken out one by one to a separate room where a nursing student measured their medical indicators (pulse, weight, and body fat percentage) using non-invasive devices—an Omron Body Composition Monitor BF511 for weight and fat percentage, and an Omron M3 device for pulse and blood pressure.
To facilitate comparison, we set the mean reward per visit to the gym at 20 NIS for the three incentivized treatments. Details regarding the treatments can be seen in the treatment section.
As part of the membership, each participant received one personal training session, during which they received an exercise program suited to their needs. The goal of providing an individualized program was to enhance the exercise’s effectiveness (Jeffery, 2012).
The experiment had an associated mobile app (a customized website for smartphones). A direct link to the app was uploaded to all of the participants’ smartphones during the introductory session. The app contained information regarding the participant’s exercise program and membership, including the number of visits, number of visits in the previous week, the rewards that had already been received, the size and timing of their next reward, and so on. Each participant also received a text message on the last day of every week telling them how many times they had exercised in the previous week and whether they had reached their weekly goal (which was set at the recommended three visits). For example, a text might read as follows: “You exercised at the gym twice this week. Good job! You can do even better next week. The recommendation is to exercise at least 3 times a week.”
Visits to the gym were recorded by chip swipes at the entrance to the sports center and to the gym. An employee of the sports center at the entrance verified that the chip used belonged to the individual that swiped it (a picture of the member pops up when the chip is swiped). This system applies to all members of the sports center.
When participants were eligible for a reward for a particular visit, they received a text message approximately 15 minutes after they swiped the chip. We set this delay so participants would still be at the gym if they were actually exercising, in order to prevent participants from swiping the chip to enter, receiving the reward, and leaving immediately. An example of such a message follows: “You are entitled to a reward of 20 shekels for exercising at the gym today. You can pick up the money at the gym reception desk. In order to receive the payment, please enter the app or the link in order to confirm receiving the money.”
After receiving this message, participants had until the end of the day to pick up their reward from the gym reception desk. If a participant forgot to collect the payment on the same day, he or she could pick it up from the experiment’s administrator.
During the four months following the incentivized period (March–June 2018), participants could still access the gym and use the app, and they continued to receive weekly text messages, but they did not receive any rewards. Of the 213 participants who started the experiment, eight canceled their participation at some point during the six months: three from Control, one from Per-visit, two from Increasing, and two from Unexpected. Thus, 205 participants remained in the study until its completion.
During the first two weeks of June 2018, all the participants were invited to a concluding lab session similar to the introductory one held in December, and received 100 NIS for attending. During the session, participants answered the same questions as in the original questionnaire, along with some additional ones (see Appendix). One hundred seventy-one participants attended the concluding session (roughly 85% of the remaining participants).
Participants were sent two short follow-up questionnaires in which they were asked whether they had continued to exercise and where. The first online questionnaire was sent in January 2019, 12 months after the beginning of the incentivized period, and the second was sent in June 2019, 18 months after the beginning of the incentivized period and prior to the university’s exam period.
Experimental Design Details
Randomization Method
We held 29 experimental sessions. Each session was attended by participants from the same treatment in order to minimize the chance of exposure to participants from other treatments. We balanced the treatment assignment according to age, gender, and commute time.
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
213 students
Sample size: planned number of observations
213 students
Sample size (or number of clusters) by treatment arms
213 students
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Tel Aviv University
IRB Approval Date
2017-10-18
IRB Approval Number
NA

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
June 30, 2018, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
June 30, 2019, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Abstract
We report the results of a field experiment in which students are encouraged to exercise. We compare a no-incentive control with a treatment in which participants are paid per visit, and to two intermittent incentive schemes: monetary rewards at increasing intervals and monetary rewards with unpredictable timing. The advantage of irregular incentives, as already recognized in the psychology literature, leads individuals to hold two contradictory expectations simultaneously: that of being rewarded and that of not being rewarded. Individuals continue to pursue the reinforcement because the anticipation of a reward is greater than the anticipation of no reward. With repeated experience, they learn to live with the paradox and to cope with the frustration of sometimes not being rewarded for a proper response. This mechanism may facilitate the maintenance of the habit over the long run, after the incentives have been removed. In line with this theory, we find the two intermittent incentive schemes increased participants’ frequency of exercising after the incentives were stopped and over an extended period of time more than the per-visit scheme.
Citation

Reports & Other Materials