1408 W. van Eerde / Personality and Individual Differences 35 (2003) 1401–1418 were significant (effect sizes reported k=25, n=247 vs. effect sizes reported as nonsignificant k=28, n =160; t(35.60)=2.62, P
4. Discussion W. van Eerde / Personality and Individual Differences 35 (2003) 1401–1418 1409 This <strong>meta</strong>-analysis provided a <strong>nomological</strong> <strong>network</strong> for the construct <strong>procrastination</strong>, by systematically relating age and gender, intellectual ability, personality variables, motives, affect, and performance to <strong>procrastination</strong>. The largest average correlations were found <strong>of</strong> <strong>procrastination</strong> with the personality factor conscientiousness (negative), self-efficacy (negative), and self-handicapping. How should these results be interpreted? The very large effect size indicating the negative relation between conscientiousness and <strong>procrastination</strong> leads to doubts whether <strong>procrastination</strong> can be distinguished from conscientiousness. Looking at the facets <strong>of</strong> conscientiousness, self-discipline has been shown to have the highest negative correlation with <strong>procrastination</strong> (e.g. Johnson & Bloom, 1995, p. 130, r= 0.75), and this correlation is just as high as was found for the facet self-discipline and the total scale <strong>of</strong> conscientiousness (e.g. Griffin & Hesketh, 2001, r=0.76). This would be an argument to consider <strong>procrastination</strong> a facet <strong>of</strong> conscientiousness, and to measure self-discipline, rather than <strong>procrastination</strong>, as an individual difference variable. Self-efficacy and self-handicapping were suggested to be indicative <strong>of</strong> the tendency to avoid negative performance feedback. However, this idea was not confirmed by the results for fear <strong>of</strong> failure and perfectionism, which would also relate to this tendency, but for which much lower effect sizes were found. Thus, <strong>procrastination</strong> appears be related to a higher extent to one’s selfimage than to the threat <strong>of</strong> receiving negative performance feedback. However, the larger effect size for self-handicapping may also be the result <strong>of</strong> the artefact that the constructs self-handicapping and <strong>procrastination</strong> overlap considerably. Conceptually, <strong>procrastination</strong> has been defined as one <strong>of</strong> the ways to self-handicap (Urdan & Midgley, 2001). And empirically, a recent study (Ross, Canada, & Rauss, 2002) also demonstrated the similarity between <strong>procrastination</strong> and self-handicapping in relation to the Big Five Model <strong>of</strong> Personality. Another finding in this <strong>meta</strong>-analysis is that it is more likely that younger rather than older people procrastinate, even within this restricted sample <strong>of</strong> a majority <strong>of</strong> college students. It may indicate that one learns to procrastinate less, or that students who did not improve may have dropped out <strong>of</strong> college. Also, contrary to the individual studies, these results indicate that it is slightly more likely that men procrastinate more than women, even within this sample with a majority <strong>of</strong> women. Still, this is a small effect size, and a heterogeneous one, but it indicates a significant difference between the sexes. Considering that the majority <strong>of</strong> categories <strong>of</strong> the effect sizes in this <strong>meta</strong>-analysis is heterogeneous, the search for moderators is important. Some <strong>of</strong> the possible moderators that need to be investigated are task characteristics, especially the difference between tasks that are too challenging or not challenging enough. Theoretically, the relation between <strong>procrastination</strong> and performance should be moderated by ability, and task characteristics such as autonomy, difficulty, and the degree to which one should be flexible and adaptable (see also Griffin & Hesketh, 2001, for these arguments relating to conscientiousness). Delay should not have the same effect on all performance measures, and that is indeed what the results <strong>of</strong> this analysis demonstrate. That is, missing a deadline and using less time to prepare are directly related to <strong>procrastination</strong>, but completing a task has no relation with <strong>procrastination</strong>, and the other performance categories were heterogeneous, indicating that other moderating factors may play a role. Different courses with different requirements may affect the relation. The extent to which performance outcomes are affected should theoretically be (partly) mediated by the degree <strong>of</strong> dysfunctional psychological outcomes. That is, if individuals are concerned