- Top of page
Background. Recently, several studies have shown that strength of self-control is a crucial factor in determining positive outcomes in individuals’ lives. Most attention has been directed to the relationships that self-control has with learning and academic achievement.
Aims. This article analyses the effects of self-control not only on school grades but also on the experience of life balance and flow. It is theorized that students with a higher level of self-control are better able to distribute their time in a satisfying way over academic and leisure matters, and are better able to shield their studying against distractions.
Samples. A total of 697 eighth graders with a mean age of 13.4 years participated in the longitudinal study.
Method. Students completed a questionnaire containing measures of self-control, school grades, subjective life balance, and flow while studying at the beginning and at the end of the school year. Structural equation modelling was used to analyse the relationships between the constructs.
Results and conclusions. Results of cross-lagged analyses show that self-control predicted school grades, life balance, and flow. The findings suggest that self-control may assist adolescents to be better prepared, not only for school, but also for coordinating their investments in different areas of their lives.
- Top of page
Within educational psychology, the positive relationship between self-control as the ability to control or override impulsive responses and school performance has been demonstrated recently by cross-sectional studies (e.g., Bertrams & Dickhäuser, 2009a). Moreover, with reference to a broad range of desirable outcomes, self-control is described as a crucial ability and a key to life-success that goes above and beyond the positive relationship with academic achievement (Baumeister, Leith, Muraven, & Bratlavsky, 1998; Gailliot, Plant, Butz, & Baumeister, 2007; Tangney, Baumeister, & Boone, 2004). Following this train of thought, the aim of the present study is to investigate the predictive power of self-control with regard to school grades and to flow during studying, as a measure of academic motivation and enjoyment of the learning activity, and with regard to a measure related to well-being that goes beyond mere academic achievement: the subjective experience of life balance.
Developmental tasks during adolescence can be found in the area of academic achievement, but adolescents also strive for goals in other life fields, such as with family and peers. Adolescence is a crucial developmental life stage in which goals, interests, and hobbies are formed (Krapp, 2003). Based on this, we regard adolescents’ experience of a balanced life as a desirable outcome supplementary to good school grades. The term life balance denotes how satisfyingly individuals distribute their time to those areas they feel are important or that they highly value (Gröpel & Kuhl, 2006; Kuhnle, Hofer, & Kilian, 2010). Possible life areas include school, social relationships, and body/health (as an area referring to physical well-being). Flow during studying is included as a motivational outcome variable. Flow is characterized by the pleasure of engaging in activities for their own sake, with beneficial effects on learning and acquisition of skills (Csikszentmihalyi & LeFevre, 1989; Nakamura & Csikszentmihalyi, 2002).
In situations in which persons aim for several goals at once, self-control appears to be an important tool because it may help guiding behaviour towards a directed aim or goal, and to postpone impulses that demand immediate satisfaction which may distract from pursuing a long-term goal (e.g., going out with friends instead of studying for an exam) (Duckworth & Seligman, 2006; Kendall & Wilcox, 1979). Within a motivational framework, self-control can be considered as the capacity to select between rivalling motivations, and to decide in favour of the dominant motivation (predecisional phase in the terminology of the rubicon model, which includes deliberations about the incentives of the possible alternatives, Heckhausen, 1991). We argue that students high in self-control are more successful because it has been found that they are better at selecting appropriate goals and budgeting time and effort properly (Baumeister et al., 1998). Additionally, self-control abilities are seen as helpful during pursuit of the chosen activity (actional phase, when individuals focus on achieving the desired outcome after the decision for one alternative, Heckhausen, 1991) because they may help to control impulses, affect, and mental processes, and therefore maintain focus on the task at hand (Baumeister et al., 1998). This is particularly necessary because even after a decision in favour of a specific task is made, for example, doing homework, other intentions are still present (e.g., chatting with friends). The incentives of a nonchosen alternative may interfere with the performance of the chosen task (e.g., Hofer et al., 2007). In such situations, students with higher self-control may better be able to shield their current activity from distracting intentions and to reach a state of flow (Csikszentmihalyi & LeFevre, 1989).
Hence, the aim of this paper is to demonstrate the predictive role of self-control, not only with regard to school performance, but also with regards to the experience of subjective life balance and flow during studying.
Self-control and school grades
We follow a definition of self-control derived from Baumeister and colleagues (Baumeister, Gailliot, DeWall, & Oaten, 2006; Baumeister, Heatherton, & Tice, 1994), postulating that self-control is generally necessary to exert control over states and responses. A high level of self-control is necessary for the regulation of behaviour in line with long-term goals pursued by studying (e.g., learning for an exam) (Mischel & Ayduk, 2002). On the one hand, self-control seems important for classroom behaviour because students have to control their impulses in order to behave properly and to concentrate on the content of the lesson instead of talking with fellow students (McLaughlin, 1976). On the other hand, self-control also seems important during after school activities, when students have to organize and plan their free time of their own accord. In fact, the relation between self-control and school grades has been demonstrated in several studies (e.g., Duckworth & Seligman, 2006; Wolfe & Johnson, 1995). Since correlations stemming from cross-sectional studies do not allow for an interpretation of influence, longitudinal studies are required. Unfortunately, only a very few address the question at hand. A study showed that when measured at the age of 4, the ability to delay immediate gratification for the sake of a later activity that is connected with higher rewards predicted substantially higher SAT scvores when these same students entered college (Shoda, Mischel, & Peake, 1990). Particularly, the study presented by Duckworth and Seligman (2005) highlighted common variance between several measures of self-control and school grades, even if grades at the first measurement point were controlled for. In our study, we want to replicate these findings and expect to find self-control predicting grades over a period of a single school year.
Self-control and life balance
Beyond academic achievement, we are interested in a variable that taps life satisfaction as a dependent variable. Comparable to studies with adult samples (Allen, Herst, Bruck, & Sutton, 2000; Greenhaus, Collins, & Shaw, 2003), we conceptualized the subjective experience of life balance as an indicator that adolescents are leading a satisfactory life (Gröpel, 2005). In the present study, we expect this variable, as well, to be influenced by self-control.
We follow Gröpel and Kuhl (2006), who define life balance as subjective judgement of an appropriate proportion of time spent in major life domains, so that a person experiences life as satisfying and balanced. Conversely, people who report spending too much or too little time in different domains of their lives are considered to have poorly balanced lives. The feeling of being in balance is not necessarily related to spending equal time in these areas but to the perception of being in balance. A review of the consequences associated with conflicts in work-life balance revealed the widespread and serious consequences associated with this kind of imbalanced life (Allen et al., 2000). The authors concluded that this failure in balancing life has dysfunctional effects on individual work life, home life, and on life satisfaction.
Similar to adults, adolescents maintain various concurrent life scenarios (e.g., being a friend, a good student, a responsible son/daughter) in which they are simultaneously striving for several goals. Adolescent students have many preoccupations related to age-specific tasks; they strive for academic achievement but also pursue goals in other areas of their lives, such as those related to family and leisure activities (Hofer & Peetsma, 2005; Salmela-Aro, Aunola, & Nurmi, 2007). Through the analyses of time-use data, school and leisure are identified as two significant life contexts in adolescents from Western societies (Hofferth & Sandberg, 2001). As time spent in one context can compete with time spent in another (Motl, McAuley, Birnbaum, & Lytle, 2006), the successful combination of work and social life is seen as a critical task (Ratelle, Vallerand, Senécal, & Provencher, 2005). Especially, conflicts between school and leisure occur quite often with adolescent students (Fries, Schmid, Dietz, & Hofer, 2005). To lead a balanced life, adolescents have to judge how much of their time should be invested in different pursuits, and then allocate that time in a way they feel is appropriate to achieve optimal results in all of their pursuits.
We regard self-control as a likely determinant of life balance for students because it seems relevant to accurate and effective timing in the face of various needs, tasks, and role demands. Reaching a satisfactory experience of a balanced life may depend on students’ abilities to regulate their behaviour effectively. For instance, a student needs time not only to learn but also to engage in leisure activities, and to recover. Thus, time has to be distributed accordingly. Up to now, very few studies have been orientated towards self-control competencies in predicting and influencing life balance (Gröpel & Kuhl, 2006). Within these studies, the linkage is often investigated as the relationship between time management as a self-control competence, and adults’ work–family conflict (Jex & Elacqua, 1999), but not with regard to the issue of adolescents’ school–leisure conflicts. In sum, we propose that self-control predicts the subjective experience of a balanced life.
Self-control and flow
We choose the tendency to experience flow during studying as a third outcome variable because flow is a motivational construct that enhances learning. Csikszentmihalyi (1991, 1997) used the term flow to describe an optimal state wherein individuals are completely involved in a current task. In cases of perceived fit of skills and challenge, students may experience fluency of action, pleasure of studying, and absence of boredom or anxiety (Acee et al., 2010; Nakamura & Csikszentmihalyi, 2002). People are intrinsically motivated through the positive experience of the activity itself. As a result, they enjoy the activity and try to obtain this experience again, thus increasing their skills due to practice (Csikszentmihalyi, Rathunde, & Whalen, 1993; Engeser, Rheinberg, Vollmeyer, & Bischoff, 2005). Originally, Csikszentmihalyi (1997) stated the tendency to experience flow as a personality variable. More recently, domain specificity of flow has been discussed (e.g., Martin & Jackson, 2008; Jackson, Martin, & Eklund, 2008). Our measure is domain specific in that it grasps the state of flow during studying.
As the experience of flow is characterized by a state of concentration that is not hindered by any other intentions or negative emotions, the chance of entering the flow process should be more likely if students are able to successfully shield themselves from distractions such as telephone calls, conversations of other people, and their own thoughts (Csikszentmihalyi, 1975), as well as from negative emotions such as sadness or fear (Csikszentmihalyi, 1997). Disturbances of this kind, if not shielded successfully, may hinder attention paid to a specific task and therefore reduce the possibility of experiencing flow due to an impaired focus (Csikszentmihalyi, 1997). Many adolescents are prone to distraction during study by alternative actions related to nonacademic tasks. Due to time scarcity (Robinson & Godbey, 1997), conflicts between different motivations may arise (Dietz, Hofer, & Fries, 2007; Fries, Dietz, & Schmid, 2008; Kuhnle et al., 2010) that prevent the emergence of flow-like states during study. Research on self-regulated learning has recognized that students not only need to have learning strategies at their disposal, but also have to regulate their academic motivation (e.g., Wolters, 2003). Self-control is regarded as a personality variable that helps to control motivation effectively. Although self-control is not needed during the process of flow, self-control regulates students’ behaviour, affect, and attention, and is associated with the ability to persevere during work. Based on this, self-control might aid in the facilitation of reaching the state of flow. Hence, self-control should increase the tendency to be completely absorbed in an activity, as experienced in flow. In sum, we want to investigate the predictive power of self-control with regard to flow during studying.
Based on the foregoing thoughts, the following three hypotheses are tested within a full cross-lagged panel design with two measurement points, where the dependent variables at first measurement are held constant.
The first aim is to test the relation between self-control and school grades. The hypothesis tested in this study is that self-control at the beginning of the school year (T1) is positively related to the grades at the end of the school year (T2) (Hypothesis 1).
It is further expected that self-control (T1) is positively related to the experience of a balanced life (T2) (Hypothesis 2).
Finally, it is hypothesized that self-control (T1) is positively related to the experience of flow in learning situations (T2) (Hypothesis 3). For all hypotheses, possible reverse relationships to these theoretically postulated relations are considered, but no hypotheses regarding these relationships are postulated.
- Top of page
Since students in this sample were nested within classes, hierarchical data analysis was used. Table 1 shows the intraclass correlation coefficients as proportions of total variance due to variance between classes. In addition to the intra class correlations (ICCs), Table 1 displays the overall means and standard deviations of the variables.
For school grades, 23% of the total variance was between classes; therefore, when testing the SEM, the maximum likelihood parameter estimates with robust standard errors (MLR) method of Mplus (Muthén & Muthén, 2007) was used. Based on this estimation method, the standard errors and the chi-squared test statistic were robust with regard to nonindependence of observations due to cluster sampling and the chi-squared test statistic was robust with regard to violations of the normality assumption (Muthén & Satorra, 1995). Therefore, this method was used although we had no predictors on level two. In consideration of the estimation method used, the Satorra-Bentler chi-squared test statistic (S-Bχ2; Satorra & Bentler, 2001) was used for the comparison of fit rather than the standard chi-squared test statistic (Muthén & Muthén, 2009).
As can be seen in Table 2, all postulated relations were significant and pointed in the expected direction. The retest stability of the variables ranged from .50 to .76.
Table 2. Intercorrelations
| 1. Life balance||–|| || || || || || || || || |
| 2. Self-control||.20**||–|| || || || || || || || |
| 3. Flow||.14**||.42**||–|| || || || || || || |
| 4. School grades||.10*||.40**||.17**||–|| || || || || || |
| 5. Life balance||.50**||.19**||.12**||.08*||–|| || || || || |
| 6. Self-control||.17**||.68**||.36**||.34**||.22**||–|| || || || |
| 7. Flow||.08*||.37**||.56**||.14**||.16**||.47**||–|| || || |
| 8. School grades||.13**||.42**||.17**||.76**||.15**||.39**||.16**||–|| || |
| 9. Age||−.02||−.11**||.04 ||−.23**||−.06||−.05||−.01||−.19**||–|| |
| 10. Sex||.04||−.14**||.02 ||−.10**||−.06||−.11**||.01||−.10**||.08*||–|
In a first step, the measurement model with eight latent and 36 manifest variables was specified. Results of this confirmatory factor analysis indicated that the proposed measurement model fit the data well (χ2= 421.64, p < .001; χ2/df= 1.33; CFI = 0.986; RMSEA = 0.022). All factor loadings were statistically significant (p < .01). The measurement errors within the same subscales or items were correlated across time to control for unknown confounding variables (e.g., Aartsen, Smits, van Tilburg, Knipscheer, & Deeg, 2002).
In a first step, the reciprocal model (Model 1) was calculated and its results fit the data very well (χ2= 422.12, p < .001; χ2/df= 1.31; CFI = 0.987; RMSEA = 0.021).
Figure 1 shows the reciprocal model (Model 1) with standardized path coefficients. The expected relations between the constructs were found and all coefficients were significant (p≤ .05). Self-control (T1) was positively related to school grades (T2), life balance (T2), and the experience of flow during the learning activity (T2). The other directions were not significant. Thus, all three hypotheses were supported. Through the inclusion of the autoregressive paths, the model explains variance of the other constructs beyond the prediction of one construct from its previously measured value (Burkholder & Harlow, 2003).
Figure 1. Reciprocal model (Model 1) with standardized path coefficients. Note. T1 indicates Time 1 and T2 indicates Time 2 testing. Parameter estimates are standardized. Solid path coefficients are statistically significant p≤ .05. Dotted path coefficients are not statistically significant (N= 697). The coefficients between the T2 variables reflect the correlation among the residuals after controlling for the corresponding T1 variables and self-control T1. Grades are coded so that higher values indicate better results.
Download figure to PowerPoint
In a second step, a more restricted version of Model 1 is specified in which certain paths were constrained to be equal to zero. Considering the theoretical assumptions, in Model 2 (Figure 2), paths from self-control (T1) to school grades (T2), to life balance (T2), and to flow (T2) were specified. This model also fits the data well (χ2= 423.05, p < .001; χ2/df= 1.30; CFI = 0.987; RMSEA = 0.021). The reciprocal model (Model 1) did provide a marginally better fit than the model with only the theoretically postulated relations (Model 2). However, the postulated model did not indicate a significant worsening in fit compared to the reciprocal model. The obtained S-Bχ2 difference test (S-Bχ2= 1.217, p= .749) was not significant, which suggests that on the basis of parsimony, the theoretically postulated model (Model 2) provides a better fit to the data, and furthermore supports the postulated relationships.
Figure 2. Theoretically postulated relationship (Model 2) with standardized path coefficients. Note. T1 indicates Time 1 and T2 indicates Time 2 testing. Parameter estimates are standardized. Solid path coefficients are statistically significant p≤ .05. Dotted path coefficients are not statistically significant (N= 697). The coefficients between the T2 variables reflect the correlation among the residuals after controlling for the corresponding T1 variables and self-control T1. Grades are coded so that higher values indicate better results.
Download figure to PowerPoint
For testing the equivalence of the found solution across multiple groups (boys/girls and Realschule/Gymnasium), firstly, we conducted a test of invariance. In a second step, we included these variables as covariates in the models. A multigroup analysis was run to check whether the structure model of the SEM holds within the female and male students. The first model with no constraints and the second model with the same pattern of path coefficients for female and male students were compared. The results of the S-Bχ2 indicated that the unconstrained model did not fit the data significantly better than the constrained model (S-Bχ2= 10.431, p= .843). Furthermore, invariance over both school types was calculated and structural invariance was shown (S-Bχ2= 16.631, p= .410). Both multigroup analyses supported structure invariance; therefore, it was not necessary to conduct separate models. Even when including gender and school type as covariates in the calculation of the two models, results confirmed that the theoretically postulated paths from self-control to school grades, life balance, and flow remain significant (p < .05; one-sided testing).
- Top of page
The cross-lagged model used in this study gives insight into the relations between self-control on the one hand, and school grades, life balance, and flow on the other. Following the model comparisons and the significances of the paths, the three hypotheses were supported. We could replicate the relationship between self-control and school grades (e.g., Duckworth & Seligman, 2005), which goes beyond results stemming from cross-sectional studies alone (e.g., Tangney et al., 2004). The inclusion of an objective achievement test would add to these results. Nevertheless, the result is noteworthy since self-control is also correlated with the other two outcome variables, thus leaving less unexplained variance for predicting grades. Furthermore, in considering the autoregressive effects of self-control, school grades, flow, and life balance, the retest reliabilities were quite high, being .68 for self-control and .76 for grades. Thus, even modest effects may be seen as substantial.
The same holds for the prediction of the other two outcome variables by self-control. Apart from predicting school grades, self-control also predicts life balance as a broader construct. This extends findings from Bertrams and Dickhäuser (2009b), who found cross-sectional correlations between self-control and life satisfaction. Therefore, it seems promising for future research to study the relations of self-control with variables related to the satisfaction of life.
Finally, self-control seems to be relevant to experiencing flow in a learning activity. The specific contribution of self-control to this outcome variable was found to be the highest one. Given the high correlation between flow and self-control, this finding seems remarkable. It is quite important that adolescent students with high self-control abilities can improve the experience of academic activities because, for adolescents, doing homework and studying is usually associated with a low level of happiness and motivation (Csikszentmihalyi & Larson, 1984). Many students do not study because they enjoy the activity itself but do it simply because they want to achieve long-term goals, such as earning good school grades; this makes it hard for students to consistently study (Wong & Csikzentmihalyi, 1991). Flow during studying is relevant because it is a pleasurable and intrinsically rewarding experience. It is shown that flow has beneficial effects on learning (e.g., Csikszentmihalyi et al., 1993). Also in our study, flow is significantly related to school grades. Possibly, flow has to be considered as an intervening variable mediating the association between self-control and time spent on homework (Duckworth, Quinn, & Goldman, 2010). If there is some general cross-domain proportion of flow, the ability to reach flow during studying might enhance the ability to generally achieve flow-like states in other areas of students’ lives. Conversely, flow in other areas might support the ability to achieve flow during studying.
Taken together, the findings highlight the role of self-control as a personality variable that seems relevant for adolescents when coping with their developmental tasks (e.g., interaction with peers, learning at school) and striving for multiple goals with limited time resources (e.g., Dietz et al., 2007; Lerner, Freund, De Stefanis, & Habermas, 2001), as self-control generally is necessary when coping with actual developmental challenges because individuals have to compensate for a lack of social structure and normative orientation (Wrosch & Freund, 2001). We speculate that during the predecisional phase, and also during the actional phase, (Heckhausen, 1991) self-control abilities are helpful in selecting and pursuing a chosen activity in order to reach a goal successfully. Self-control may support students in selecting an activity relevant to a long-term goal (e.g., studying) instead of pursuing after immediate satisfactions, as well as in maintaining the selected behaviour.
One implication of these findings for educational psychology is the plea to include outcome variables other than academic achievement. Life balance is a candidate for this. Nevertheless, caution should be added because of limitations of this study that have to be discussed. First, the variables are based on self-reports, a method that has been criticized when used imprudently (e.g., Elliott, 2004). Although some researchers state that self-report data are needed and reflects the behaviour across all situations, even for children (Beitchman & Corradini, 1988), the possibility of socially desirable answers has to be taken into consideration. On the other hand, the variable life balance, per se, is subjective and cannot be measured in terms of behavioural observation. Second, especially with adolescents, it is quite frequent that at least within certain periods they may feel their life as quite balanced in terms of time distribution while their parents or teachers have the strong impression that they should allocate their time much differently. For instance, many parents think their child uses too much time playing computer games or chatting on the phone. In a given case, it is quite hard to decide whether the time distribution has developmental benefits. However, the examples point to the necessity of complementing the construct of life balance by other indicators, for instance, by teachers’, parents’, and peer ratings, to get a rounder picture of the balance of a student's life. Also, time-use analyses (e.g., Flammer, Alsaker, & Noack, 1999) could be helpful in gaining insights into the time adolescents with a certain degree of reported life balance actually invest in different areas of life. And third, the retest stability of self-control was higher than for life balance (.50) and flow (.56) because the higher variance in these two constructs could be explained by self-control, which may at least partially contribute to the explanation of why the theoretically postulated relations were higher than the reverse relationships.
Flow during study can be regarded as a motivational outcome variable. Flow is desirable from an educational perspective and is rather independent from life balance. We could show that self-control offers a possibility to reach the flow experience. To our knowledge, the scenario technique used in this study has never been used in measuring flow; therefore, cautious interpretations are appropriate. The influence of the scenario technique in comparison to the experience sampling method, a method mostly used to measure flow by asking for self-reports about actual behaviour and state at random occasions (Nakamura & Csikszentmihalyi, 2002), should be further explored. We relied on a well-established scenario mainly used in the field of motivational interference (e.g., Hofer et al., 2007) that, however, has not been used before to measure flow in a specific situation. Whether the findings can be generalized over other studying situations remains to be confirmed in further research.
Taken together, all three outcome variables tap relevant areas of students' life and all are affected by self-control. Due to the fact that self-control is correlated with a range of positive and desirable outcomes (Tangney et al., 2004), and as shown in the present study with school grades, life balance, and flow, it is encouraging that the enhancement of self-control abilities seems to be possible. In early studies, self-control training had been successfully used in order to reduce inattentiveness and disruptive behaviour in children (e.g., McLaughlin, 1976). Recently, Gailliot et al. (2007) found that 2 weeks of self-regulated exercise, such as reducing the consumption of tobacco, or trying to control one's eating habits or emotions, increased the capacity of self-control and reduced self-control depletion. Therefore, one practical implication of this study could be that the training of self-control, especially in adolescents, is not only a promising approach to improving students’ ability to concentrate on learning tasks (Clark, 2002), but also to enhance their experience of life balance. Another implication of this study can be said to be that the enhancement of self-control abilities as a further possible intervention method seems to be quite promising.
As the present study shows that grades, life balance, and flow as outcome variables in educational psychology are influenced by self-control, the findings are also relevant to the field of positive psychology (Seligman & Csikszentmihalyi, 2000). They centre on the question of how students’ strengths can be supported rather than how their deficits can be cured. Life balance as an educational outcome is a genuine positive variable. The results also teach a view of students that comprises their whole functioning as persons with a broad variety of goals, motives, and intentions. Furthermore, it was possible to show that the subjective experience of balance is relevant to more than just adults, and further studies exploring additional implications of disturbances of life balance for the daily experience of students seem worthwhile.