Don’t grade schools on character skills

In a recent New York Times op-ed, Angela Duckworth raises important concerns about the dangers associated with attaching consequences to school children’s development of character skills. Duckworth’s primary concerns regard the inherently imprecise ways in which character skills are currently measured and the potential for strategic manipulation—or even outright cheating.

These concerns are valid. However, Duckworth’s critique (and the broader literature on the importance of non-cognitive or character skills) overlooks an additional feature of such skills, which raises serious doubts about policies that would uniformly penalize or reward schools for the development of students’ non-cognitive skills. Unlike cognition or certain character skills like grit (which Duckworth has studied for many years), other character skills have mixed effects on student outcomes that may be helpful or productive in some contexts, but harmful or counterproductive in others.

Therefore, even if measurement is correct, and there is no gaming or cheating by teachers or schools, it is not obvious which character skills we want to reward or punish.

To illustrate this point, in recent research, my research team and I look at roughly 7,000 men and women who were born in the same week in Great Britain in 1959.  When these individuals were 11-years-old, their teachers were asked to complete inventories of student misbehavior in the classroom.  The goal was to identify different types of childhood misbehavior or maladjustment.

In general, misbehavior in the classroom was—and continues to be—viewed as reflecting a set of negative character skills, which lower schooling attainment and cause problems for individuals throughout their lives.

Our results indicate that the story is much more nuanced. We show evidence that some forms of childhood misbehavior, known by developmental psychologists as “externalizing behaviors,” and manifesting as aggression, hyperactivity or hostility, are indeed bad for schooling.  However, they are also valued in the labor market, predicting higher wages.  Surprisingly, this is true for both men and women.  In fact, high-externalizing women not only command higher wages, but also work longer hours.  In other words, the research shows robust evidence that some of the “character skills” underlying misbehavior, despite their negative effect on schooling, can be quite valuable in other domains.

The conclusion is not that we should promote externalizing behaviors, such as aggression or hostility. This would almost surely bring its own set of problems. Rather, this research suggests that interventions devised to reduce externalizing behaviors could be short-sighted. They might promote educational attainment in the short-run, but stifle character skills that are productive in the long-run. 

“[T]his research suggests that interventions devised to reduce externalizing behaviors [such as aggression or hostility] could be short-sighted. They might promote educational attainment in the short-run, but stifle character skills that are productive in the long-run.”

Looking more broadly, it is no longer controversial that character skills (sometimes called soft skills) are important factors that predict a wide range of behaviors and outcomes, including education, employment, health, marriage, and mortality.  Moreover, there is a fairly wide consensus that character skills are malleable at young ages, at least in comparison to cognition.  This has opened up the possibility to design policies that target children’s character skills, often with the noble intention of reducing inequality by helping disadvantaged youth. 

The mixed effects of misbehavior highlight the perils of policies that would penalize supposedly negative character skills. Well-intentioned policies affecting character skills could stifle valuable skills and therefore have unintended negative consequences—for the children targeted by these policies and for the economy as a whole. Instead of high-stakes evaluations of character skills, what is needed is more in-depth study of the long-run and varied effects of character skills on lifetime outcomes, including well-being and labor market performance.