LETTERS TO THE EDITORS
To the Editors:
I enjoyed the article “Lifelong Impact of Early Self-Control” (September–October) by Terrie Moffitt, Richie Poulton, and Avshalom Caspi, reporting on an interesting study that has significant social policy implications.
They presented strong evidence that people who show good self-control capabilities at three years of age tend to be more successful in various ways later in life, and then discussed at length how self-control might be enhanced in order to improve people’s lives. On an intuitive level, I find it credible that a person’s self-control capability would be instrumental—causally related—to success in life. However, some of the analysis presented seems to suggest that self-control, while indicative of success, may not be instrumental.
The analysis shows that higher levels of self-control at age three are strongly correlated with success in life. But self-control levels at older ages are only weakly correlated with self-control at age three. This hints that self-control levels when the subjects are older than three do not correlate as strongly with success. If that is indeed the case, then it would suggest that some other underlying factor affects both self-control at age three and success later in life, and that self-control per se is not the cause of success.
My question for the authors is, How does the correlation between self-control and success vary according to the subject’s age when self-control was assessed? If the correlation between adult success and self-control increases with age, then the case that self-control is causally related to success and that enhancing self-control should be beneficial would be buttressed. If, on the other hand, self-control assessed later is less correlated with adult success than self-control assessed at an early age, then it would suggest that some other underlying factor is influencing both self-control and success.
John Freidenfelds, Ph.D.
Dr. Moffitt and collaborator Dr. Salomon Israel respond:
To answer the question, “Which predicts life outcomes more strongly, childhood or adult self-control,” we turned to data from the Dunedin Study, which was described in our recent article. It is important to appreciate that the same method cannot be used to measure self-control in children and adults, because the expression of self-control often differs at successive stages of human development. For example, children with low self-control have difficulty taking turns on playground equipment; adults with low self-control are disorganized at work and make unwise snap decisions.
Each Study member’s self-control in childhood was measured through ratings of age-relevant behaviors, as described in our article, and were compiled to make one composite scale. In contrast, self-control in adulthood was measured through ratings of an age-relevant personality trait called conscientiousness. These ratings were made by informants who knew them well during our latest assessment carried out at age 38. These ratings also yielded a scale.
Here, we report two examples of life outcomes from our article: adult credit ratings and health problems (metrics detailed in the article). First, the Pearson correlation coefficient (r)—a measure of the strength of a linear relationship between two variables, where higher magnitudes represent a stronger relationship—measuring the association between childhood self-control and adult credit ratings was r = .26, and between adult conscientiousness and adult credit ratings was r = .27. Second, the association between childhood self-control and adult health problems was r = –.19, and between adult conscientiousness and adult health problems was r = –.18. Therefore, despite the differing content of the child and adult self-control measures, their statistical associations with life outcomes generated coefficients of similar magnitude.
Our article reported that self-control measurements were less stable from childhood to adulthood than measures of the IQ. In part, this lesser stability could arise because measures of self-control given to children versus adults differ in content and format (unlike the Wechsler IQ tests, which were designed to test intellectual skills in a standard way from age 7 to 70). But the long-term instability of self-control measures probably also means that many people change their rank on self-control over time, relative to their age-peers. Simply put, all of us get better at self-control as we grow up. Few small children can wait to eat a marshmallow, but most adults find it easy. However, some of us improve more than others. In the feature article, we reported one factor that improved some Study members’ self-control rank: working as a supervisor responsible for subordinate employees. Regardless of what caused improvement, Study members who improved most had better adult outcomes than expected. Whether self-control is ascertained in children or adults, more of it is associated with better health and wealth, and lower proclivity to crime.
We caution against overemphasizing the question of stability in self-control. People improve skills, even if changing a psychological trait seems more challenging. For instance, the track record is weak for efforts to enhance intelligence, but this does not stop us from teaching skills that use intelligence, like reading and math. Likewise, it might not be necessary to alter the core personality trait of conscientiousness in order to teach people the skills necessary to look after their health and plan for retirement.