It’s time again for New Year’s resolutions, which provide one of the greatest natural experiments in the world for those of us who study health behavior change. As I have probably noted here before, the average duration of a new behavior is between 1 and 6 months, so you can look for the early-January crowds in your gym to diminish right around Groundhog Day. This finding fits well with Prochaska’s stages of change model, which suggests that 3-6 months is the amount of time it takes to move from the action to the maintenance stage — i.e., for a new behavior to become a habit. After that transition, the new behavior is much more likely to “stick” for the long term. One method that many people use to get through this transitional time is to harness the power of streaks.
A streak can be defined as any behavior that is repeated (often daily) over a period of time, without interruption. The lack of interruption is crucial to this definition: The reason it’s useful to track a streak at all is because it layers on a new source of motivation, which is simply to keep the streak going for its own sake. There are then two separate sources of motivation which can be mutually reinforcing: (a) the person’s original motivation to engage in the behavior, and (b) their motivation to maintain the streak. Furthermore, these two sources of motivation have a push-pull dynamic that works on the person from opposite directions. While the original motivation is often gain-focused (e.g., “I want to exercise to I prove my health”), the streak-maintenance motivation is inherently loss-focused (i.e., “I’d hate to lose my streak”). Kahneman and Tversky’s Prospect Theory suggests that the desire to avoid losing something that you have can be a particularly powerful one, often referred to as the “endowment effect.” A streak can be easily seen as a personal possession, and the potential loss of a streak is therefore experienced as a threat that motivates action. This is useful because potential gains such as future good health are less inherently motivating on a day-to-day basis, and the threat to health from skipping one day of exercise usually feels pretty small and remote. (The exception is under conditions of uncertainty, like a health crisis, when someone is likely more strongly motivated to seek gains because their future health status is unknown). But people who feel OK with their current status are more strongly motivated by loss aversion. Some writers have suggested that the power of streaks comes from the dopamine release that occurs in the brain when a streak counter ticks upward for the day, and that's definitely a salient reward. But the endowment effect is probably a better explanation: The threat of losing a streak is immediate and visceral, and can therefore have a much greater motivating effect.
Exercise is not the only potential application for streaks, although due to the widespread use of fitness apps it is probably the most common. Meditation or mindfulness apps commonly track streaks, and they can be used to track other regular practices like yoga. Streak-tracking is part of the reason so many of us are hooked on the daily “Wordle” puzzle. Writers sometimes log their writing time to track streaks, and that's a strategy I recommended in my post about building writing habits (the every-2-weeks schedule for this blog makes it pretty easy to keep up, but I haven’t missed a posting since starting in 2019). And programs to quit smoking or abstain from alcohol and drugs have long used daily streak tracking as a way to help their users stay clean and sober, especially during that crucial 3 months after initiating a change. This blog notes that some other businesses with less healthy connotations, like Starbucks and Netflix, do not use streaks to encourage our continued use of their services, even though they are surely tracking our behavior. (Actually, storefront businesses like Starbucks often do use loyalty cards to trigger rewards, which are a form of streak tracking).
Technology is not a necessary ingredient for streaks, but it has helped to make this method more widely known and used. The simplest method for tracking streaks might be simply marking days off on a calendar. But many smartphone-based habit-tracking apps are also now available, including one that's simply called "Streaks." The language-learning platform Duolingo has a particularly effective streak tracking feature, which lets you know the number of days on which you completed at least one lesson. And I have written previously about the digital badges awarded by Apple Health, which use exercise streaks as one basis for recognition. Gamification is the idea that people are more likely to do things when they are fun, and is another of the behavior-economics tools highlighted in the latest edition of the book Nudge. Some additional gamification strategies that can be readily combined with streaks are visual displays of one's own data over time; offering badges or medallions when a user meets certain milestones; displaying leaderboards of top performers; allowing for personal competition with a group of friends on the same site; gradually modifying the target behavior using stretch goals; or even being able to trade in earned points for some type of reward.
Are there risks associated with this behavior-change technology? Streaks seem to be very powerful in encouraging a particular behavior, perhaps so much that a few writers have worried they actually constitute a form of addictive behavior. Runners, for instance, might continue exercising even through an injury in order to maintain a streak. But a more common concern is that when a streak is finally broken, it can lead to what the addictions field calls an “abstinence-violation effect.” This means that once someone has already had a single lapse, subsequent lapses become more likely. (In substance use, the lapse is usually a failure of "abstinence" from alcohol or drugs, which is the source of the effect's name; in exercise or other positive behaviors, it happens instead when someone fails to engage in the behavior after a streak). You might think that a lapse would prompt someone to redouble their efforts, but in fact the effect tends to go the other way. If streaks were just about receiving a tiny dopamine reward, that reward should actually be more salient after it was missed for a day -- in the language of economics, it ought to have greater marginal utility. But that's not what usually happens. An explanation of streaks based on the endowment effect results in the opposite prediction: People no longer "have" the streak, so they also no longer have any motivation to maintain it.
You might think of abstinence violation as a "what the hell effect": Once you have already lost your streak, there's no particular incentive to re-start it. If you've missed a day of running, why not take the second day to rest and recover as well? If you have already eaten one slice of cheesecake your diet is blown anyway, so why not enjoy a second one? This pattern can be true even for very important health behaviors with potentially serious consequences. A recent daily-diary study found that when young men with HIV missed a single day of medication, they were then more likely to miss the subsequent day's dose as well. Too many missed doses of HIV medication can result in a treatment-resistant virus, so the positive incentives for the behavior are still in effect, and perhaps even become stronger as each day of medication is missed. But the abstinence-violation effect at the end of a streak is stronger yet, perhaps manifesting itself as a sense of relief that one doesn't have to worry about the streak anymore.
An interesting study last year by J. Silverman and A. Barasach looked at the boundary conditions for streaks, and identified some rules of thumb for those who want to use this behavior-change method more effectively:
- Provide a way to log the streak. Streaks can be motivating whether or not they are logged, as seen in the study of people's HIV medication use (participants weren't formally tracking it) and in a study of exercise where people who met a step goal on one day were more likely to also meet it the next day even if there was no tracking involved. However, the effect of a logged streak on behavior is even greater. Silverman and Barasach proved this with a clever design in which everyone had actually completed a 4-day streak, but some participants were told that the software had malfunctioned and couldn't show their 4th day as completed. Those participants were less likely to exercise the next day even though they knew their streak was actually intact. This demonstrates that the effect of seeing a streak is independent of the actual streak behavior.
- Emphasize the streak itself as a goal. Reminder messages like "don't lose your streak!" are probably quite effective, because participants in Silverman and Barasach's research reported that maintaining a streak was an independent goal for them, even when the target behavior was something intrinsically enjoyable like watching amusing videos. They also reported a higher sense of goal attainment when the streak was intact. These subjective findings provide additional evidence that keeping a streak going can be its own reward.
- Allow various behaviors to 'count' for the streak. In another study, Silverman and Barasach found that when people were allowed to categorize a variety of behaviors as being "part of the streak," such as different types of word games, they felt the same sense of accomplishment and level of motivation to continue as if they had maintained the identical behavior throughout. This finding suggests that it might be beneficial to define behavioral goals broadly, such as "exercise" including both strength training and a range of cardiovascular activities, rather than more narrowly as in a "run streak." The Duolingo app again provides a good learning streak example, where users can complete a lesson, practice vocabulary, or read a story in order to reach the same goal of extending their streak by a day. People find it easier and more satisfying to maintain their streak when some variety in how they achieve the daily goal is allowed, and having this type of flexibility doesn't seem to decrease the motivating power of the streak.
- Avoid highlighting broken streaks. Although visually displaying an intact streak can enhance its motivational effect, a visual display showing a broken streak also makes the abstinence violation effect stronger. Silverman and Barasach found that when people couldn't see their broken streak, it weakened their motivation but did so less than when the broken streak was visually displayed. Therefore, in cases where a streak has in fact been broken, it might be better to hide that information from the user. For instance, one could just encourage the person to resume the target behavior without emphasizing the broken streak.
- Make streaks repairable. When users in another of Silverman and Barasach's studies were given the ability to "repair" their streak they didn't feel quite the same level of accomplishment as those with a true intact streak (at some level, they knew their streak was broken). But on the other hand, they didn't show the strong negative impact of the abstinence-violation effect, and their motivation stayed higher than that of people whose streak was broken and not repaired. A simple "repair" mechanism might help to overcome the negative side effects of a broken streak, and if such opportunities are explicitly limited (e.g., "you used your streak freeze! Be sure to practice today or your streak will really be gone!") they might actually prompt someone to resume the target behavior.
- Blame someone else for the broken streak. A final interesting finding from Silverman and Barasach comes from a study in which people were given the opportunity to attribute a broken streak to factors outside their own control (in this case, a "user quote exceeded" message on their computer that meant they were unable to continue their streak). In this case, the broken streak had a less negative effect on their behavior than a break that was self-attributed (here, the self-attribution was due to a very tough word puzzle that didn't count toward the person's streak of successes -- a failure actually arranged by the researchers, but that the participant was likely to blame on him- or herself). A last-ditch method to prevent broken streaks from spiraling into abstinence-violation relapses might therefore be to give the person an excuse for why their streak wasn't maintained: e.g., a message that says "I'm sorry, your data from yesterday couldn't be uploaded" even when they were in fact uploaded and the result would have been a broken streak. At the individual level, it's probably beneficial to blame the weather or lack of time when you lose your streak -- idiosyncratic factors that were beyond your control. Whether that's actually true probably doesn't matter; you're better off with a little bit of reality distortion in this case!
Comments
Post a Comment