Self-Control in Daily Life

Abstract

In this daily diary study, we investigated the flexibility-enhancing effects of positive affect on the self-reported success of self-control strategies followed in daily life; 297 participants completed a 13-day daily diary that included measures of positive affect, desire, and habit strength as well as three self-control strategies (i.e., monitoring, distraction, and stimulus control). We found specific effects of positive affect on self-control strategies: Individuals with higher positive affect were most successful when following a strategy of distraction (e.g., thinking about something else), particularly when faced with strong tempting desires. These results reinforce the idea that positive affect is associated with both cognitive flexibility and distractibility, which may help people distract them from tempting desires.

Keywords

self-control temptations desires affect daily diary ambulatory assessment

Introduction

People are faced with numerous challenges to self-control in everyday life. They may be trying to eat healthier, exercise more, drink less alcohol, work harder, or curb what they say (or do not say)—all in an effort to modify unwanted behavior. To date, most studies have examined self-control success and failure in controlled laboratory settings (e.g., Hagger, Wood, Stiff, & Chatzisarantis, 2010). Recent studies, however, expand on these approaches emphasizing the importance of everyday behavior in self-control (e.g., Berkman, Falk, & Lieberman, 2011; Hofmann, Baumeister, Förster, & Vohs, 2012). Studies about the effectiveness of everyday self-control strategies are still scarce (Quinn, Pascoe, Wood, & Neal, 2010). Specifically, more research is needed that examines the effectiveness-specific strategies people employ in daily life to regulate temptations and habits and to identify possible moderating factors associated with self-control success. One promising candidate is positive affect: Laboratory research showed that positive affect can facilitate or hinder self-control depending on self-control demands (Aspinwall, 1998; Wenzel, Conner, & Kubiak, 2013). However, to the best of our knowledge, it is still unclear whether and how this evidence translates to self-control as it occurs in the everyday life of individuals. Thus, the goal of the present daily diary study was to investigate how positive affect influences self-control success by facilitating (or thwarting) different self-control strategies.

Self-control, often defined as the altering of one’s own impulses to attain goals (Muraven, Tice, & Baumeister, 1998), requires both stable maintenance of current goals and flexible goal switching (Goschke, 2003). On the one hand, currently active goals need to be maintained and protected from distraction in order to achieve goal attainment. On the other hand, flexibility is needed when switching between goals or when disengaging from an unachievable goal. Thus, successful self-control requires a context-sensitive balance between stability (maintaining goal intentions) and flexibility (switching goal intentions). If this balance is off, individuals could suffer from perseveration or they could suffer from impulsivity and distractibility (Dreisbach & Goschke, 2004).

Research has shown that positive affect influences this stability–flexibility balance. For example, positive affect is associated with increased cognitive flexibility in adopting abstract future goals, particularly self-control goals, whereas negative affect is linked to focused attention on immediate goals like mood management (Fishbach & Labroo, 2007; Labroo & Patrick, 2009). Positive affect is also known to broaden one’s attentional focus (Fredrickson, 2004) and lead to a broader repertoire of action plans (Fredrickson & Branigan, 2005). Dreisbach and Goschke (2004) showed that positive affect enhances flexibility and distractibility at the cost of reduced stability and focus on a cognitive task. When new information was linked to a target response, participants in a positive affective state performed better than participants in a neutral affective state since directing attention to the novel (helpful) information promoted the processing of the target. Conversely, when novel information was linked to the distractor, participants in a positive state performed worse because the increased focus on the new (distracting) information interfered with the target response. In a similar vein, Wenzel, Kubiak, and Conner (2014) demonstrated previously that the flexibility-enhancing effect of positive affect also holds for self-control and that greater cognitive flexibility explained the influence of positive affect on consecutive self-control performance: If novel information in the second task was linked to its target, positive affect improved performance compared to neutral affect. In turn, positive affect led to worse performance if the novel information was linked to the distractor in the second task. These results demonstrate that the effect of positive affect on cognitive and self-control depends on whether the novel information individuals switch to is beneficial or not for the given task goal.

However, so far, the evidence for the flexibility-enhancing effect of positive affect is based on rather short laboratory tasks. It is unclear whether such an effect occurs in daily life as well. Ambulatory assessment represents a necessary complementary strategy to investigate the ecological validity of results found in laboratory research (Kubiak & Stone, 2012). In order to assess demands for stability and flexibility in self-control situations in daily life, we adopted the self-control strategies from Quinn, Pascoe, Wood, and Neal (2010) since they affect the stability–flexibility balance differently, as explained later. Quinn et al. identified three main strategies of self-control based on the delayed gratification research: vigilant monitoring of one’s own behavior, distraction (e.g., thinking about something else), and stimulus control that is reducing or avoiding the salience of a cue (e.g., leaving the situation). Whereas in monitoring, one focuses attention on the behavior or the cognitions that should be controlled, distraction concentrates attention on something else. Quinn et al. (2010) further demonstrated that the effectiveness of these self-control strategies depends on whether the unwanted behaviors are temptations or habits. Temptations arise when an impulse like eating a tempting donut interferes with a long-term goal like reducing weight (Nordgren & Chou, 2011). Habits are automatically activated responses in a consistent context (Wood & Neal, 2007) like reading the newspaper every morning or taking a walk every afternoon. Quinn et al. (2010) found that monitoring was less successful than stimulus control in resisting temptations, whereas monitoring was more successful than stimulus control for regulating habits, suggesting that focusing on a tempting impulse is detrimental (Hofmann, Deutsch, Lancaster, & Banaji, 2009), but focusing on a habit is not. However, contrary to Quinn et al.’s expectations, distraction did not differ from stimulus control or monitoring.

In the present research, we propose that positive affect interacts with self-control strategies in explaining differences in self-control success in everyday life. Specifically, we predict that positive affect will enhance the success of distraction in daily life when confronted with strong temptations, whereas it impairs the effectiveness of monitoring. Distraction favors flexibility over stability, in that to distract oneself from a temptation, individuals have to modify the response disposition to the tempting situation by stepping back and disengaging oneself from the impulse in order to think about something else. If positive affect helps people to improve cognitive flexibility by facilitating processing of new information (Dreisbach & Goschke, 2004; Wenzel, Kubiak, & Conner, 2014), it should increase self-control success when people are using distraction as a self-control strategy in daily life. If the effectiveness of distraction depends on positive affect, it could explain the null findings for distraction by Quinn et al. (2010). In contrast, monitoring favors stability over flexibility, in that stable attention is required in order to ensure that a response is not performed. Thus, positive affect should negatively interact with monitoring. This two-way interaction (positive affect by self-control strategy) should also be influenced by temptation and habit strength, in that distraction should be particularly effective in regulating strong temptations compared to the other self-control strategies when people feel positive.

Method

Participants

A total of 327 undergraduate students at the University of Otago, New Zealand, participated for an opportunity to earn research credits or payment (maximum of NZ$55). Those 327 participants reported 3,670 observations. Data from nine of these participants (n = 37 observations) were excluded because they completed fewer than 50% of the daily surveys. On 922 days (25.4% of actual answers), participants did not report an unwanted behavior, which reduced the total observations for our analyses to a total of 2,711 and the total number of participants to 310. Since missings in other measures of interests were not imputed, another 13 participants (n = 214 observations) were excluded due to missings, leaving a total of 297 participants (65% female) and 2,497 observations for analysis. These participants were aged 17 to 30 years (M = 19.93 years, SD = 2.32). In terms of ethnicity, 73.6% self-identified as European, 10.4% as Asian, 4.0% as Maori/Pacific Islander, and 12.0% from other ethnic backgrounds.

Procedure

This study was part of a combined laboratory and 13-day daily diary study called the Daily Life Study. At the start of the study, participants were invited to the laboratory, where they provided informed consent and completed self-report measures of demographics over the computer. Starting the next day, for 13 days total, participants logged onto a secure website between 3 p.m. and 8 p.m. to complete a brief 5-min questionnaire that included the following measures embedded within other measures of daily experience. Thus, the daily diary comprised one query per day that retrospectively assessed the domain and nature of the unwanted behavior and how it was regulated. After 2 weeks, participants returned to the laboratory and were debriefed and reimbursed for their participation.

Measures

Domain

To control for the influence of the type of event that the individuals encountering at a given moment, participants were given a list of examples of unwanted behaviors derived from the list of frequently reported unwanted activities reported by Quinn et al. (2010), for example, procrastinating or sleeping. In each daily diary, participants picked one that they tried to inhibit or change that respective day. If they experienced more than one unwanted behavior on that day, they were instructed to indicate the strongest one, which was subject to the further measures.

Affect

Participants rated how they felt “before the situation with the unwanted behavior” using 2 items of the valence subscale of the short form of the German Multidimensional Mood Questionnaire (MDMQ; Steyer, Schwenkmezger, Notz, & Eid, 1997). The MDMQ assesses affect on three dimensions: valence (items: good and bad), energetic arousal, and tense arousal on a scale ranging from 1 (not at all) to 5 (extremely). Mean scores were computed for the valence dimension (with one of the items reversed scored) with higher scores representing higher valence (i.e., more positive affect). Summary statistics for affect and the other measures are presented in Table 1.

Table 1.

Range, Means, Standard Deviations, Reliability, and Intercorrelations on Self-Control Success, Desire Strength, Habit Strength, Valence (Both Before and After the Act of Self-Control).

Measure	Range	M	SD	R _KRN	R _CN	1	2	3	4
1. Self-control success	1–5	2.64	1.11	.82	–
2. Desire strength	1–5	2.22	1.23	.85	–	.01
3. Habit strength	1–4	2.55	0.80	.88	–	−.10	.05
4. Valence (before act of self-control)	1–5	3.55	0.80	.81	.56	.12	.04	−.04
5. Valence (after act of self-control)	1–5	3.52	0.88	.77	.66	.42	.08	−.10	.39

Note. The presented variables are not person-mean centered. R_KRN = between-person reliability; R_CN = within-person reliability (day-to-day changes, see Shrout & Lane, 2012).

Self-control strategies

Participants then checked the strategy they used to inhibit or change the unwanted behavior, with (1) monitoring (36.2% of the total number of diary entries across individuals), (2) distraction (26.8%), (3) stimulus control (16.0%), (4) nothing (17.2%), or other (3.8%). Definitions and examples of each strategy were listed in the diary instructions. The examples were monitoring (using a watch to be on time; monitoring alcohol use), distraction (thinking about something else; doing another activity), stimulus control (removing myself from the situation or removing the opportunity to do it such as avoiding certain triggers; leaving early from a pub), and nothing (I didn’t try any strategies to inhibit or change unwanted behavior). We excluded the “other” category since we were interested in the interactions of positive affect and specific self-control strategies.

Desire strength

Since temptations are based on desires, the enactment of these should lead to positive feelings in that moment. Therefore, temptations were measured by asking participants “how much would performing the unwanted behavior make you feel good” in that situation from 1 (not at all) to 5 (extremely), with higher scores indicating greater desire strength.

Habit strength

Habit strength (M = 2.60, SD = 0.78) was assessed with the single item “How often have you performed the unwanted behavior in the past?” on a scale ranging from 1 (monthly or less often) to 4 (several times per day).

Self-control success

The dependent variable of self-control success was measured by “how successful was the attempt to change or inhibit the unwanted behavior?” from 1 (not at all) to 5 (completely).

Control variables

Control variables in all analyses were age, gender, domain, day of study, and day of the week. Since daily diaries involve retrospective assessment of recent thoughts and behavior, individuals may not be able to distinguish their affect before the unwanted behavior from their affect afterward. For instance, the affect measures may be biased by individuals’ feeling about how they managed these events, in that if they managed them well, positive affect after the situation increased, which may lead to higher positive affect measure before the situation. Therefore, participants also rated how they felt “after the situation with the unwanted behavior” using the same MDMQ valence scale. There was no evidence for multicollinearity, with both measures correlating only moderately, r = .39. Thus, controlling for affect after the self-control, an independent affect coefficient for positive affect before the situation could be obtained.

Analytic Approach

We computed multilevel models with random intercepts and coefficients in Stata 13 (Stata Corporation, College Station, TX) with daily diary observations (Level 1) nested within participants (Level 2). All continuous Level-1 independent variables were person-mean centered, and all continuous Level-2 independent variables were grand-mean centered (Enders & Tofighi, 2007). Since we were interested only in the within-subject processes, the averaged continuous Level-1 independent variables were not entered into the model (Bolger & Laurenceau, 2013). Categorical variables were dummy-coded (self-control strategies reference: “doing nothing”; gender reference: “male”; domain: other) since we were primarily interested in Continuous × Categorical interactions (Cohen, Cohen, West, & Aiken, 2003).

Results

Adherence with the daily diary protocol was good, with participants completing more than 11 of the 13 possible daily diaries on average (M = 11.11, SD = 2.14; 86% adherence). To test whether participants completed fewer diaries across the course of the study, we computed a χ² goodness-of-fit test that did not differ from an equal distribution, χ²(12) = 17.19, p = ns. Consequently, compliance did not deteriorate over the 2 weeks of the study.

We first fitted the unconditional model without predictors on self-control success, which revealed that 74.7% of the total variance in self-control success was attributable to within-person variability. We, then, included all Level 1 and Level 2 variables to compute the main effects and added the two- and three-way interactions in the next step. We tested whether treating Level 1 predictors as random improved model fit, which was the case, χ²(4) = 103.60, p < .001. Since including random covariances did not further improve the model fit significantly, χ²(6) = 4.33, p = .614, we only included the random effects without their covariances. The final model (Table 2) accounted for 37.3% of the variance in self-control success. To facilitate the presentation of the results in Table 2, we do not present the coefficient for each strategy-by-moderator interaction but an omnibus test, which indicates a significant interaction. This omnibus can be interpreted like the F value in an analysis of variance, and F values can be obtained by dividing the χ² values by its degrees of freedom. To interpret significant differences, we then show the simple slopes (Cohen et al., 2003) for each self-control strategy (Table 3).

Table 2.

Fixed Effects Estimates (Top) and Variance Estimates (Bottom) of Self-Control Success as Function of Self-Control Strategies (SCS), Desire and Habit Strength, and Valence.

Within-Subjects Fixed Effects	Estimate	SE	z	p ^a	95% CI [Lower, Upper]
Step 1
SCS	χ²(3, N = 2,497) = 136.58			<.001
Desire strength	0.01	0.02	0.22	.828	[−0.04, 0.05]
Habit strength	−0.05	0.04	−1.44	.149	[−0.12, 0.02]
Valence (before act of self-control)	0.04	0.03	1.23	.218	[−0.02, 0.10]
Valence (after act of self-control)	0.43	0.03	17.29	<.001	[0.38, 0.48]
Step 2
Desire Strength × Valence	0.04	0.03	1.41	.158	[−0.02, 0.10]
Habit Strength × Valence	0.05	0.05	1.08	.280	[−0.04, 0.15]
SCS × Desire Strength	χ²(3, N = 2,497) = 6.85			.077
SCS × Habit Strength	χ²(3, N = 2,497) = 6.27			.099
SCS × Valence	χ²(3, N = 2,497) = 8.98			.030
Step 3
SCS × Desire Strength × Valence	χ²(3, N = 2,497) = 15.06			.002
SCS × Habit Strength × Valence	χ²(3, N = 2,497) = 3.13			.366
Intercept	2.30	0.39	5.84	<.000	1.53	3.07
Random parameters
SCS	0.18	0.03	5.56	<.001	0.14	0.26
Desire strength	0.04	0.01	3.39	.001	0.03	0.08
Habit strength	0.07	0.03	2.56	.010	0.03	0.15
Valence	0.02	0.02	1.26	.207	0.01	0.11
Intercept	0.21	0.04	5.96	<.001	0.14	0.28

Note. CI = confidence interval. N = 297 persons, 13 days, 2,497 observations. Self-control success was measured on a scale from 1 to 5; χ² tests reflect omnibus tests for interactions of the self-control strategies dummy variables.

^aAll p values are two-tailed except in the case of the random parameters, where one-tailed p values are used (because variances are constrained to be nonnegative).

Table 3.

Simple Slopes for Each Self-Control Strategy (SCS) and Valence, Desire Strength, Habit Strength, Valence × Desire Strength.

Measure	Estimate	SE	z	p ^a	95% CI [Lower, Upper]
Valence
Monitoring	−0.05	0.05	−0.93	.353	[−0.14, 0.05]
Distraction	0.17	0.06	3.03	.002	[0.06, 0.28]
Stimulus control	0.04	0.08	0.42	.675	[−0.13, 0.20]
Doing nothing	0.11	0.08	1.38	.169	[−0.05, 0.26]
Desire strength
Monitoring	−0.02	0.04	−0.49	.623	[−0.09, 0.05]
Distraction	0.10	0.04	2.48	.013	[0.02, 0.18]
Stimulus control	−0.03	0.05	−0.51	.611	[−0.12, 0.07]
Doing nothing	−0.01	0.05	−0.18	.859	[−0.11, 0.09]
Habit strength
Monitoring	0.01	0.05	0.26	.791	[−0.09, 0.12]
Distraction	−0.01	0.06	−0.07	.948	[−0.13, 0.12]
Stimulus control	−0.20	0.08	−2.42	.016	[−0.36, −0.04]
Doing nothing	−0.11	0.08	−1.47	.143	[−0.27, 0.04]
Valence × Weak Desires
Monitoring	−0.11	0.07	−1.70	.089	[−0.25, 0.02]
Distraction	−0.01	0.07	−0.22	.829	[−0.15, 0.12]
Stimulus control	0.07	0.11	0.66	.506	[−0.14, 0.28]
Doing nothing	0.25	0.10	2.52	.012	[0.06, 0.44]
Valence × Strong Desires
Monitoring	0.02	0.07	0.33	.739	[−0.11, 0.16]
Distraction	0.35	0.09	3.97	<.001	[0.18, 0.52]
Stimulus control	−0.01	0.11	−0.01	.992	[−0.23, 0.22]
Doing nothing	−0.04	0.11	−0.37	.714	[−0.25, 0.17]

Note. CI = confidence interval.

Effectiveness of Self-Control Strategies

In the first step, we entered the main effects into the model (Table 2), which indicated significant differences between the three self-control strategies. Simple main effects analyses with pairwise comparisons of the marginal means revealed that individuals reported significantly more self-control success if they had used any of the three self-control strategies monitoring (M = 2.79, SE_M = 0.05, p < .001), distraction (M = 2.80, SE_M = 0.05, p < .001), and stimulus control (M = 2.98, SE_M = 0.07, p < .001) compared to doing nothing (M = 2.05, SE_M = 0.07). Moreover, among the self-control strategies, stimulus control was more effective than monitoring, p = .007, and distraction, p = .013, with no difference between the latter two strategies, p = .861. Neither desire or habit strength nor valence exhibited significant main effects on self-control success.

In the second step of the analysis, we included the two-way interactions. Self-control strategies significantly interacted with valence, p = .030, and marginally significantly interacted with desire strength, p = .077, and habit strength, p = .099. As illustrated in Table 3, simple slope analysis revealed that positive affect only influenced distraction significantly, in that distraction was more effective with increasing positive affect. In turn, the success of monitoring and stimulus control as well as doing nothing was not significantly influenced by positive affect. The same pattern was revealed for the simple slope analysis with desire strength, in that only distraction was more effective with increasing desire strength. For habit strength, stimulus control was less effective when dealing with stronger habits, whereas monitoring was not impacted by habit strength.

Both two-way interactions between self-control strategies and valence of desire strength, respectively, were qualified by a three-way interaction between self-control strategies, desire strength, and valence in the third step of the hierarchical analysis, p = .002. Simple slope analyses revealed that the positive effect of positive affect on the success of distraction increased with increasing desire strength, with a large effect for strong desires (+1 SD below the mean; see Figure 1, right panel) and no significant relationship for weak desires (−1 SD below the mean). Moreover, the simple slope of monitoring for weak desires was marginally significant, that is, positive affect yielded a negative effect when participants reported using monitoring for weak desires (see Figure 1, left panel). These results demonstrate the flexibility-enhancing effect of positive affect, in that the success of distraction in regulating strong tempting desires highly depends on valence.

Figure 1.

Marginal means of self-reported self-control success as a function of valence for weak (−1 SD of mean of desire strength, left panel) and strong temptations (+1 SD of mean of desire strength, right panel). The bars at each marker reflect 95% CI.

Choice of Self-Control Strategy

Next, we investigated how the primary constructs influenced the choice of self-control strategies. To this end, we computed a multilevel multinomial logistic regression with self-control strategies as the outcome. We examined self-control success, desire strength, habit strength, valence, and its interactions as the predictors, using the gsem command in Stata 13. As illustrated in Table 4, participants facing strong habits chose more likely monitoring and less likely distraction or stimulus control. Although desire strength did not directly influence the likelihood for any self-control strategy, it interacted with self-control success: Less effective regulators (1 SD < M of self-control success) chose monitoring marginally significantly more often, b = .03, SE = 0.02, p = .076, 95% CI [−0.00, 0.06], and distraction significantly less often, b = −.04, SE = 0.02, p = .020, 95% CI [−0.07, −0.01], when confronted with strong compared to weak desires, whereas effective regulators (1 SD > M) did not, b = −.02, SE = 0.02, p = .348, 95% CI [−0.05, 0.02] and b = .01, SE = 0.02, p = .399, 95% CI [−0.02, 0.04], respectively. However, as noted before, the interaction on the likelihood of monitoring failed to reach significance.

Table 4.

Simple Effects of the Multilevel Multinomial Logistic Regression With Self-Control Strategies as the Outcome and Self-control Success, Desire Strength, Habit Strength, Valence, and Their Two-Way Interactions as the Predictors.

Measure	Estimate for Monitoring [95% CI]	Estimate for Distraction [95% CI]	Estimate for Stimulus Control [95% CI]	Estimate for Doing Nothing [95% CI]
Self-control success	0.01 [−0.02, 0.03]	0.01 [−0.02, 0.03]	0.04*** [0.02, 0.06]	−0.05*** [−0.08, −0.03]
Desire strength	0.01 [−0.01, 0.03]	−0.01 [−0.04, 0.01]	−0.01 [−0.02, 0.01]	0.01+ [−0.00, 0.02]
Habit strength	0.08*** [0.04, 0.12]	−0.04* [−0.07, −.00]	−0.04* [−0.07, −0.01]	−0.01 [−0.03, 0.00]
Valence	0.01 [−0.03, 0.05]	−0.05** [−0.09, −0.02]	0.03* [0.01, 0.06]	0.01 [−0.00, 0.03]
Self-Control Success × Desire Strength	−0.02 [−0.05, 0.00]	0.02* [0.00, 0.05]	−0.00 [−0.02, 0.02]	0.00 [−0.01, 0.1]
Self-Control Success × Habit Strength	0.01 [−0.03, 0.05]	0.01 [−0.03, 0.04]	−0.01 [−0.04, 0.03]	−0.01 [−0.3, 0.01]
Self-Control Success × Valence	−0.02 [−0.06, 0.02]	0.05* [0.01, 0.08]	−0.02 [−0.06, 0.02]	0.01 [−0.01, 0.02]
Desire Strength × Valence	−0.00 [−0.04, 0.03]	−0.00 [−0.03, 0.03]	0.01 [−0.02, 0.03]	−0.00 [−0.02, 0.01]
Habit Strength × Valence	−0.01 [−0.07, 0.05]	−0.01 [−0.06, 0.05]	−0.00 [−0.05, 0.05]	0.02 [−0.01, 0.04]

Note. In order to improve clarity of the presentation of our findings, we only present the averaged marginal effects. CI = confidence interval.

*p < .05. **p < .01. ***p < .001.

With increasing valence, participants were less likely to choose distraction and more likely to choose stimulus control, even if they reported to be more successful when using distraction under positive affect. Moreover, valence interacted with self-control success if distraction was chosen: Less effective but not more effective regulators reported to choose distraction less often when regulating strong desires, b = −.09, SE = 0.02, p < .000, 95% CI [−0.14, −0.05] and b = −.02, SE = 0.02, p = .318, 95% CI [−0.07, 0.02], respectively. None of the interactions for stimulus control and doing nothing reached significance.

Discussion

We investigated the role of affect on the success of different self-control strategies in daily life. In general, stimulus control was the most effective self-control strategy. In turn, distraction was less effective overall, which was offset under certain circumstances: When participants were experiencing greater positive affect, distraction was reported to be a successful self-control strategy, especially when confronted with strong desires. The results for distraction are in line with the flexibility-enhancing effect of positive affect (Dreisbach & Goschke, 2004) as well as the broaden and build model of positive emotions (Fredrickson, 2004) and provide ecologically valid evidence for the flexibility-enhancing effect. On the grounds of laboratory research, positive affect is linked with higher flexibility and higher distractibility. Therefore, positive affect should help distract oneself from tempting desires, which was found in our study. Thus, the lower self-control success of distraction in general may be due to the increased demand of additional cognitive flexibility, which is facilitated by positive affect. Our findings illustrate why the effects of affect on self-control may be complex at times: When faced with tempting desires, it is better to distract oneself when feeling positive than negative. In short, it is not sufficient to learn which strategy works for which situation—it is also important to consider one’s affective state.

Participants with higher self-control success surprisingly did not report weaker habits or stronger desires in general. Basically, stronger temptations or habits should be mirrored in lower self-control success on average: Quinn et al. (2010) found a significant main effect for habit strength, in that regulating strong habits was less successful than regulating weak habits, whereas the main effect of habit strength in our study failed to reach significance. It has to be noted, though, that we used person-mean-centering of habit strength for our analyses, which differs from the analytical approach in the study by Quinn et al. (2010) where grand-mean-centering was used. We deliberately opted for person-mean-centering, as grand-mean-centering does not allow for separating between within- and between-person processes (Hamaker, 2012), which was crucial for our research questions. Interestingly, when using the grand-mean-centered approach by Quinn et al. (2010) on our data set, we find significant negative effect of habit strength on self-control success, b = −.09, SE = 0.04, p = .012, as well as a marginally significant positive effect of desire strength, b = .04, SE = 0.02, p = .101, which is in line with the Quinn et al. (2010) findings. Since the person-mean-centered measure of habit strength was not significant in our study, the effects found when using the grand-mean-centered measures have to be due to between-person differences. Thus, our results demonstrate that participants who face stronger habits than the ones they usually have to deal with are not less successful in regulating them (within-person process). Instead, participants who generally face stronger habits than other participants reported less success at doing so (between-person process). However, the random effect of the within-subjects habit strength measure was significant in the final model (Table 2), in that approximately 95% of the participants were within −0.56 and 0.46 (−0.05 ± 0.26 × 1.96) of the typical value for habit strength. Some individuals reported less success and some more success when confronted with stronger habits. Future research should examine individual differences such as trait self-control in order to explain the variability between the individuals in more detail.

In the case of desire strength, neither the within-subject nor the grand-mean-centered main effect was significant, although the latter approached significance (p = .101), in that stronger desires were more difficult to control on average. This may seem odd since stronger temptations should pose a greater self-control dilemma and be more likely to result in indulgence. Counteractive self-control theory (Fishbach, Zhang, & Trope, 2010) may explain these null findings, as it suggests that temptations could, in fact, also strengthen a long-term goal by indicating a need for self-control. Thus, some individuals indulge when strongly tempted, whereas others could increase their self-control efforts in order to avoid putting their long-term goal at risk. This individual difference in the influence of desire strength on self-control is supported by the significant random effect of desire strength (Table 2). Future research could tap into this variability by directly assessing short- versus long-term goals (e.g., Hofmann, Baumeister, et al., 2012), which enables investigating whether trait self-control is associated with strengthening long-term goals when faced with tempting desires.

Our results also shed some light on some of the null findings by Quinn et al. (2010): Contrary to their expectations, distraction did not differ from stimulus control or monitoring in the success of regulating strong desires. We found that the success of distracting from strong desires depended on positive affect, in that distraction was only more effective than stimulus control or monitoring when individuals reported positive affect. In line with the results by Quinn et al. (2010), we found self-control control to be less successful in regulating strong habits. We could not replicate, however, that monitoring was the least successful strategy for desires and the most successful one for habits. One possible explanation might be that Quinn et al.’s (2010) measure of habit strength not only assessed how often the behavior was performed but additionally included the location where the unwanted behavior has been performed in the past. Thus, when an unwanted behavior is often performed but in different locations, it is a weak habit in Quinn et al.’s (2010) study, but a strong one in ours, which may explain the differences.

We also examined the choice of specific self-control strategies when facing habits and tempting desires in order to clarify whether participants choose the more effective strategy. Two of our findings of these analyses are noteworthy: First, participants chose stimulus control more likely with increasing habit strength, which was in line with the decreased effectiveness of stimulus control for strong habits. Moreover, although the success of monitoring and distraction was not influenced by habit strength, participants chose them more (monitoring) or less likely (distraction), which is in line with the theoretically predicted effectiveness. Thus, since continually monitoring of persistent habits is wearisome (Baumeister, Vohs, & Tice, 2007) and may increase the likelihood for indulgence (Hofmann, Vohs, & Baumeister, 2012), individuals might switch to a less effortful strategy at some point in order to avoid the strenuous monitoring.

Second, desire strength did not influence the choice of a particular self-control strategy alone. Only the interaction with self-control success was significant: More effective regulators (high self-control success) were not more likely to pick the (theoretically) more effective strategy. Less effective regulators chose monitoring and distraction more likely when a different strategy would be more effective. Third, although distraction was more successful under positive affect, participants chose it less likely with increasing valence. To sum up, based on our results, it seems that individuals may not necessarily recognize the optimal strategy since even individuals who reported higher self-control success did not systematically choose the self-control strategies that were most effective given our data regarding affect, desire, and habit strength. Thus, future research clearly should include an in-depth examination of predictors and concomitants of the choice of particular self-control strategies in daily life.

There are a number of limitations of the present research. First, people were asked to remember how they felt before enacting the control strategy. It is possible that people might misremember how they felt to be in line with their current feelings (Ross & Conway, 1986). However, the daily diary was completed relatively early after the events’ occurrence which should reduce a potential retrospective memory bias. We also found interaction effects and differences in marginal means for negative and positive affect and self-control strategies, which cannot be simply explained by a retrospective memory bias. Finally, we also assessed affect after enacting the control strategy and controlled for its influences on daily self-control success.

Second, we assessed requirements of stability and flexibility only indirectly by assessing self-control strategies that differently affect the stability–flexibility balance. Future research could study the influence on this balance more directly, for example, by using an ambulatory assessment design with several prompts per day that assess current demands for stability or flexibility.

Third, we adopted the measures of habit and temptation strength from Quinn et al. (2010). There, temptation strength is a measure of the strength of desires (in conflict), in that it assesses the positive affective consequences only of enacting a strong desire (feeling better). However, the measure fails to capture negative affective consequences such as guilt (Hofmann & Fisher, 2012) that may arise since giving in to desires may also jeopardize a long-term goal. Thus, future research should either improve measurement of temptation strength by separate desire strength from conflict strength (Hofmann, Baumeister, et al., 2012) or assess temptation strength more directly.

Fourth, we did not distinguish between different facets of positive affect, such as motivational intensity, that might differently influence cognitive flexibility and, thus, self-control. According to Harmon-Jones, Gable, and Price (2012), high motivationally intensive positive affect before goal attainment (e.g., enthusiasm) is associated with narrowed attention in order to foster the completion of that goal. Conversely, low motivationally intensive positive affect after goal attainment (e.g., satisfaction) is associated with broadened attention in order to facilitate one’s search for new goal opportunities. However, temptations are high in approach motivation, and our results suggest that positive affect rather helps than hinders the success of distraction in regulating strong temptations. Nevertheless, this distinction offers a promising future research avenue to investigate the influence of positive affect on self-control in more detail.

Fifth, the correlational design cannot provide causal evidence for the effect of positive affect and, thus, cannot rule out alternative explanations like affect covarying with other factors possibly affecting self-control. While our data cannot replace experimental studies about the flexibility-enhancing effect of positive affect (e.g., Dreisbach & Goschke, 2004; Wenzel et al., 2013), it demonstrates its potential in explaining the role of affect on self-control efforts in naturalistic situations and corroborates the experimental evidence mentioned above with ecologically valid data.

To conclude, the impact of affective experience on self-control depends on the chosen self-control strategy. We found that self-control success is higher under positive affect when flexibility is required. Feeling good helps people to distract themselves from strong tempting desires and focus on something new by fostering the required flexibility. Therefore, when trying to distract yourself from that tempting donut, it may be better to do it when feeling good. Otherwise, it may be better to avoid the presence of this tempting treat.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Aspinwall

L. G.

(1998). Rethinking the role of positive affect in self-regulation. Motivation and Emotion, 22, 1–32.

Baumeister

R. F.

Vohs

K. D.

Tice

D. M.

(2007). The strength model of self-control. Current Directions in Psychological Science, 16, 351–355.

Berkman

E. T.

Falk

E. B.

Lieberman

M. D.

(2011). In the trenches of real-world self-control: Neural correlates of breaking the link between craving and smoking. Psychological Science, 22, 498–506.

Bolger

Laurenceau

J.-P.

(2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. New York, NY: Guilford.

Cohen

West

S. G.

Aiken

L. S.

(2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Hillsdale, NJ: Erlbaum.

Dreisbach

Goschke

(2004). How positive affect modulates cognitive control: Reduced perseveration at the cost of increased distractibility. Journal of Experimental Psychology: Learning, Memory and Cognition, 30, 343–353.

Enders

C. K.

Tofighi

(2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12, 121–138.

Fishbach

Labroo

A. A.

(2007). Be better or be merry: How mood affects self-control. Journal of Personality and Social Psychology, 93, 158–173.

Fishbach

Zhang

Trope

(2010). Counteractive evaluation: Asymmetric shifts in the implicit value of conflicting motivations. Journal of Experimental Social Psychology, 46, 29–38.

10.

Fredrickson

B. L.

(2004). The broaden-and-build theory of positive emotions. Philosophical Transactions-Royal Society of London Series B Biological Sciences, 359, 1367–1378.

11.

Fredrickson

B. L.

Branigan

(2005). Positive emotions broaden the scope of attention and thought-action repertoires. Cognition & Emotion, 19, 313–332.

12.

Goschke

(2003). Voluntary action and cognitive control from a cognitive neuroscience perspective. In Maasen

Prinz

Roth

(Hrsg.), Voluntary action: Brains, minds, and sociality (pp. 49–85). Oxford, England: Oxford University Press.

13.

Hagger

M. S.

Wood

Stiff

Chatzisarantis

N. L. D.

(2010). Ego depletion and the strength model of self-control: A meta-analysis. Psychological Bulletin, 136, 495–525.

14.

Hamaker

E. L.

(2012). Why researchers should think “within-person”: A paradigmatic rationale. In Mehl

M. R.

Conner

T. S.

(Eds.), Handbook of research methods for studying daily life (pp. 43–61). New York, NY: Guilford Press.

15.

Harmon-Jones

Gable

P. A.

Price

T. F.

(2012). The influence of affective states varying in motivational intensity on cognitive scope. Frontiers in Integrative Neuroscience, 6, 1–5.

16.

Hofmann

Baumeister

R. F.

Foerster

Vohs

K. D.

(2012). Everyday temptations: An experience sampling study of desire, conflict, and self-control. Journal of Personality and Social Psychology, 102, 1318–1335.

17.

Hofmann

Deutsch

Lancaster

Banaji

M. R.

(2009). Cooling the heat of temptation: Mental self-control and the automatic evaluation of tempting stimuli. European Journal of Social Psychology, 40, 17–25.

18.

Hofmann

Fisher

R. R.

(2012). How guilt and pride shape subsequent self-control. Social Psychological and Personality Science, 3, 682–690.

19.

Hofmann

Vohs

K. D.

Baumeister

R. F.

(2012). What people desire, feel conflicted about, and try to resist in everyday life. Psychological Science, 23, 582–588.

20.

Kubiak

Stone

A. A.

(2012). Ambulatory monitoring of biobehavioral processes in health and disease [Editorial to the Special Issue on Ambulatory Monitoring]. Psychosomatic Medicine, 74, 325–326.

21.

Labroo

A. A.

Patrick

V. M.

(2009). Psychological distancing: Why happiness helps you see the big picture. Journal of Consumer Research, 35, 800–809.

22.

Muraven

Tice

D. M.

Baumeister

R. F.

(1998). Self-control as limited resource: Regulatory depletion patterns. Journal of Personality and Social Psychology, 74, 774–789.

23.

Nordgren

L. F.

Chou

E. Y.

(2011). The push and pull of temptation: The bidirectional influence of temptation on self-control. Psychological Science, 22, 1386–1390.

24.

Quinn

J. M.

Pascoe

Wood

Neal

D. T.

(2010). Can’t control yourself? Monitor those bad habits. Personality and Social Psychology Bulletin, 36, 499–511.

25.

Ross

Conway

(1986). Remembering one’s own past: The construction of personal histories. In Sorrenlino

R. M.

Higgins

E. G.

(Eds.), Handbook of motivation and cognition: Foundations of social behavior (pp. 122–144). New York, NY: Guilford Press.

26.

Shrout

P. E.

Lane

S. P.

(2012). Psychometrics. In Mehl

M. R.

Conner

T. S.

(Eds.), Handbook of research methods for studying daily life (pp. 302–321). New York, NY: Guilford Press.

27.

Steyer

Schwenkmezger

Notz

Eid

(1997). Der Mehrdimensionale Befindlichkeitsfragebogen (MDBF) [The multidimensional mood questionnaire – Manual]. Göttingen, Germany: Hogrefe.

28.

Wenzel

Conner

T. S.

Kubiak

(2013). Understanding the limits of self-control: Positive affect moderates the impact of task switching on consecutive self-control performance. European Journal of Social Psychology, 43, 175–184.

29.

Wenzel

Kubiak

Conner

T. S.

(2014). Positive affect and self-control: Attention to self-control demands mediates the influence of positive affect on consecutive self-control. Cognition & Emotion, 28, 747–755.

30.

Wood

Neal

D. T.

(2007). A new look at habits and the habit-goal interface. Psychological Review, 114, 843–863.