Research on self-control has heavily analyzed how individuals implement self-discipline and has touched on this subject to a more minimal extent. The time gap following a self-discipline violation and the subsequent event is actually more critical than the act of self-discipline being broken. A field of research covering a range of activities, such as eating, moral conduct, and academic performance, has found a surprising, yet consistent, result. The response society believes is disciplined, and those who advise against this unstructured leniency believe to be indulgent, actually predicts recurrence.
The researchers intentionally allocate varying amounts of attention to different research areas. Self-control research focused on starting, maintaining, and ending amounts of self-control paints a picture of depletion of a resource which leads to breaches of self-control. This breach of self-control has shown to occur in every area of self-control; prevention has failed. Not addressing this half of the equation is not a case of poor research. Failure to solve prevention is more appealing to the general research audience and to readers.
This contradicts the general assumption of discipline writing, where self-criticism is perceived to be the stinging mechanism that brings an individual back to an established standard. The standard is not sustained by self-criticism. If anything, self-criticism causes a breakdown of the psychological barrier that enables a standard to be temporarily maintained.
I. The Sequence Most Advice Gets Backward
In the world of discipline, most people think of a lapse and a response to a lapse as one long failure. The breakdown of the rule and the person’s subsequent remorse are seen as one failure. Behavioral scientists see it differently. They see a lapse as a specific behavior that occurs at a particular moment in time. The lapse also occurs at a particular moment in time. The response is a decision about the lapse that follows and whether to take any additional action. Baumeister and colleagues see self-regulation as a series of decisions rather than a long-lasting decision and see the response that follows the lapse as one decision of many in the series.
Analyzing the two events allows us to understand what an intervention is likely to achieve. An intervention designed to prevent a lapse would involve setting up new schedules, re-engineering contexts, and using better commitment devices than the current ones. An intervention designed to address the response to a lapse would use different strategies. Treating the two as one failure type would likely lead to response-focused solutions that attempt to prevent failures, but the research cited in this paper reveals that such failures remain one of the greatest failures that a company can have.
How the response is framed changes what should be relevant. Advice that helps to avoid the first slip because of a lack of willpower, poor planning, or little motivation focuses on the first link in the chain. It cannot consider the internal response as prevention has failed. With more research showing that the second link (the internal response) determines the outcome much more than the first link (the external stimuli), advice focusing on the first link is taking a poor, outdated approach.
II. The Diet Study That Isolated the Mechanism
Adams and Leary set out to differentiate the second link of this connection in their 2007 paper. Restrictive eating tendency participants (screened with a researcher-developed measure) numbering 84, female, undergraduates, were randomly assigned to one of three groups: consumed a preload and self-compassion induction; consumed a preload and no intervention; or a no preload control group. Participants in the two preload groups consumed a Dunkin’ Donuts cake doughnut. Participants were shown a nature documentary designed not to contain content addressing eating, body image, or emotionally charged content. Because the preload was the same for both groups, the researchers could differentiate the effect of the self-compassion manipulation from the effect of the food.
The self-compassion exercise involved the researcher acknowledging participants' guilt over eating study doughnuts, normalizing the behavior, and linking reduced self-judgment to Neff's self-kindness, common humanity, and mindfulness. The no-intervention group received no statement of this nature. All participants engaged in the measure of interest, a bogus taste test, by consuming three bowls of candy.
The result was not the simple story often repeated in secondary summaries of this research. The authors explicitly report that they did not replicate the classic disinhibition effect, the pattern in which any preload predicts overeating relative to no preload. The effect of the self-compassion manipulation on subsequent eating was found to be influenced by trait restrictive eating, which suggests the manipulation was more likely to be clear to participants high on the trait of conscious, effortful avoidance of forbidden food and was not clear to participants at the other end of the spectrum.
Among the most rigid dieters, self-compassion after the doughnut produced eating behavior indistinguishable from having never eaten the doughnut at all.
Some results showed divergence when participants were restricted from eating. For example, participants in the self-compassion condition consumed less candy than the no-preload control group, t (48) = 1.99, p = .05. These participants acted like most people who engage in compensation, eating less after they consume a preload. However, highly restrictive participants in the no self-compassion group did not exhibit compensation and ate similar amounts to the no preload group. The direct comparison of the self-compassion and no self-compassion preload groups among highly restrictive participants was on the edge of statistical significance, t(50) = 1.81, p = .08. A similar pattern was seen for negative affect. Highly restrictive participants in the no self-compassion group reported more negative affect than the no preload group, t(46) = 4.02, p < .001. In contrast, highly restrictive participants in the self-compassion group showed no differences when compared to the no preload group, p= .20.
Here, the direct comparison of self-compassion versus no intervention following identical lapses in self-control was on the edge of significant for the eating measure and was significant for the confirmatory measure of negative affect. The authors recognized an additional limitation in their study. The no self-compassion group did not receive any statements about food from the researcher, so some effect in the self-compassion group may be attributed to the researcher’s involvement and not the specific self-compassion message. The authors suggest that a well-designed study would include an additional preload group that received a food-related, but not self-compassionate statement.
The data only support certain conclusions. Popular framings overstate the findings. For people with a strong preference for rigid dietary restrictions, Pastun and colleagues’ brief self-compassion intervention after a diet violation (as compared with no intervention) stopped participants from compensatory eating. It also decreased feelings of negative affect. The negative effect was greater without the intervention. The data do not support the argument that self-compassion reduces overeating in the general population. Also, the data do not support the argument that the guilt-and-overeating story holds in the general population.
This study relates to Heatherton et al.’s earlier study on ‘counter-regulatory eating,’ which is also known as the ‘what the hell effect.’ In this earlier study, dieters who experienced a threat to their self-esteem (a task that failed but was purported to be easy) consumed significantly more in the subsequent task, while non-dieters in the same study did not increase their food consumption. Adams and Leary’s study builds on the earlier study; they do not confirm the effect as the threat and disinhibition, and the protective effect of self-compassion, from overeating are restricted to dieters with a rigid relationship with food, compared to all dieters.
III. Defining Self-Compassion Against the Obvious Objection
The term self-compassion invites a specific and reasonable objection before its research base can be taken seriously: that it functions as a permission structure, a psychologically comfortable justification for the same behavior a person is ostensibly trying to change. Self-indulgence and self-compassion are not synonymous. Kristin Neff clearly understood the distinction early in her research publications. Neff chose to be deliberate, specifically distinguishing self-compassion from self-indulgence, as opposed to assuming it was understood.
Each of Neff’s three components requires something from the individual. The first, self-kindness, is responding to your own failures of understanding rather than with harsh self-criticism. The second is common humanity; placing your failure within the context of all human experience as opposed to seeing it as an example of how inadequate you are as an individual. The third is mindfulness, which is holding the failure in your mind so as not to blow it all out of proportion to the point of seeing it as a catastrophe, but not totally ignoring it. None of the three components entails decreasing the standard to which you hold yourself. Each deal with the aftermath of a failed self-evaluation.
The mindfulness component distinguishes the construct from careless self-forgiveness. Neff’s construct is not about looking the other way or downplaying the failure. It demands that you look failure right in the eye and is in fact more demanding than harsh self-criticism.
IV. Self-Compassion and Subsequent Motivation
Self-compassion functions as a permission structure, but what does that mean for self-compassionate behavior? If it functions similarly to a permission structure, self-compassionate behavior should mean taking less effort. Psychologists Juliana Breines and Serena Chen tested this theory in four different studies in 2012, and each study did not corroborate their finding with a permission structure.
The value of doing four separate studies instead of one is the variety of outcome measures. A study that only deals with self-reported motivation would let permission structure skeptics argue that participants just reported what they thought they would feel rather than what they would actually do. Resolving the objection involves changes in belief, moral motivation, time-related behavior, and information seeking, all requiring similar outcomes but differing significantly.
Participants in the first experiment were more able to describe their personal weaknesses as changeable rather than as fixed if they were instructed to consider the weaknesses in a self-compassionate way.
In the second study , participants who applied self-compassion to a recalled moral transgression reported greater motivation to make amends and to avoid repeating the transgression than participants who did not. This third experiment focused on observable behavior over self-reports. Participants who showed self-compassion after receiving a failing test were found to spend more time studying for the next test compared to participants who did not. This was an experiment that was much more difficult to explain using general mood effects compared to explaining with self-reported motivation. The self-compassionate group in experiment four actively sought social comparison with higher performers, unlike the control group that avoided it.
Self-compassion did not lower effort in a single one of the four experiments. It raised it.
The consistency across four independent experimental paradigms, incremental belief change, moral motivation, behavioral study time, and social comparison seeking, strengthens the claim considerably beyond what any single study could establish. Self-criticism appears to motivate the avoidance of failure itself, since sitting with a harshly judged mistake carries an emotional cost most people are not inclined to absorb voluntarily. Self-compassion appears to reduce that cost enough to make direct engagement with the failure tolerable, which is the precondition for doing anything productive about it.
V. A Structural Response to the Post-Lapse Interval
These can be synthesized into a direct, three-part answer to the interval after a lapse, instead of an overall disposition to be easy on oneself.
The first component is the classification of the lapse. However, this component's evidence suggests a tendency for some participants to rigidly adhere to dietary restrictions, rather than a general effect. This shows the component is most relevant to individuals with an all-or-nothing approach to their standards. Don't assume a onetime skip means a consistent pattern for this group. Neff’s common humanity component provides the specific reframe: that this event should be viewed as an ordinary occurrence and something that is shared by many, as opposed to an instance of unique failure.
The what the hell effect rests on a false binary assumption that a day is completely whole or completely ruined. There is no in-between. Heatherton, Herman, and Polivy’s original counter-regulatory eating research examined the cost of the binary through ego threat, rather than a dietary slip. Adams and Leary’s research proposes that the same binary is most dominant in people with high rigid, rule-based standards. The corrective is more procedural and less emotional. Treat a lapse as data and continue the plan at the next convenient opportunity, as opposed to waiting for a ‘Monday’ or ‘first of the month,’ which rarely aligns with the lapse.
The fourth component is the direction of effort after the lapse occurs. The third study by Breines and Chen provides the most interesting data on this, as they measured actual study time after the lapse, as opposed to other studies that only measured the intended effort. Self-criticism also directs available energy to the failure, where the result is no behavior, while self-compassion makes the energy ready for the next goal or effort. This energy of self-criticism is also used to ruminate.
This does not change the basic standard. What is different is the management of the narrow interval following a skipped standard. This is the interval that almost all conventional advice on discipline considers a ‘secondary’ element of the failure and not, as evidence-based research suggests, the decision point.
One last point is to differentiate this argument from a much broader argument with which it may be confused. The research discussed in this paper does not advocate for the indiscriminate use of self-criticism to lower your standards. Both Adams and Leary’s and Breines and Chen’s participants demonstrated an interest in their respective failures. The criticism is more specific. It suggests that the method most people reach for almost automatically in the 60 seconds following failure and almost always without thinking is the most ineffective method. It cannot fulfill the only standard self-criticism is supposed to deal with, which is to motivate individuals to get back to their original standard.
The One Thing to Do Today
In the next instance of a lapse, try to subvert the self-critical response with the statement, “This happens all the time, and we proceed from this point.” To proceed to the next step in the plan, take the next smallest action without resuming from a specified starting point.
Have advice, a letter, or a suggestion for us? Send it to kanpoteau@gmail.com.
This is part of a series of research-based articles looking at discipline within the scope of that research and the findings that surprise the public and help shape everything published under the Deliberate Project.
See you next time,
The Deliberate Project Team
References
1. Adams, Claire E., and Mark R. Leary. “Promoting Self-Compassionate Attitudes Toward Eating Among Restrictive and Guilty Eaters.” Journal of Social and Clinical Psychology, 2007.
2. Neff, Kristin D. “Self-Compassion: Theory, Method, Research, and Intervention.” Annual Review of Psychology, 2023.
3. Breines, Juliana G., and Serena Chen. “Self Compassion Increases Self-Improvement Motivation.” Personality and Social Psychology Bulletin, 2012.
4. Heatherton, Todd F., C. Peter Herman, and Janet Polivy. “Effects of Physical Threat and Ego Threat on Eating Behavior.” Journal of Personality and Social Psychology, 1991.