In social psychology, there is a small industry for articles reporting positive correlations between measures of self-control and various measures of socio-economic status and achievement. For example, Tangney, Baumeister and Boone (2008) found that self-control, measured on a self-report scale they devised, is correlated with better grades, somatic and mental health, and stable social relationships such as marriage. Moffitt et al. (2011) conducted a longitudinal study that followed children who had participated in the Mischel “marshmallow tests” to the age of 32 years old, and found childhood performance in that delayed gratification assignment to be correlated with measures of health and economic success, interpersonal adjustment, and  with criminal justice outcomes, even after controlling for childhood socio-economic factors.

Studies like these have been widely publicised, and the message in popular science media often leads with the idea that self-control is a stable trait that some have, some don’t. The ones who were dealt a losing hand in self-control got a losing hand overall, ending up with poor health, poverty, unstable relationships, and crime not out of ill will, but because they simply can’t hold it together. In short, the causal arrow goes from poor self-control to socioeconomic disadvantage.

This line of thinking has received plenty of criticism. Some have pointed out that the studies have been designed from a perspective assuming a middle-class lifestyle, and that self-control may not be as adaptive for people from all backgrounds.

For example, delayed gratification is less rational if you have grown up in an environment where you cannot trust that any promised future rewards will in fact materialise. And a study by Miller et al. (2015) found that while self-control did predict better grades in low-SES youth, it also lead to faster cellular aging, presumably due to the stress involved in  combining a high degree of self-control with economic uncertainty. (This impact on cellular aging was not observed in middle-class youth.) Perhaps a different set of capacities is adaptive for navigating such circumstances, and perhaps disadvantaged people excel in those capacities rather than delayed gratification?

It is tempting to reject the whole idea of self-control as oppressive, but I’m inclined to think that we can hold on to what’s valuable about this concept and introduce some nuance to how we think of the interaction of SES and self-control. Everyone sometimes finds themselves in a situation where their plans and wants pull them in two different directions. It is useful for us to have language and concepts for the capacity to navigate those situations. We can use that language to identify situations where we need to navigate our own motivational conflicts in order to act as we intend, and to identify and employ various self-control strategies and skills accordingly. We can get cunning about self-control: for example, when trying to cut down on booze, we can employ social support and avoid places where we are more likely to overindulge.

Yet I think we also should not explain away the robust correlations between low measures of delayed gratification and other measures of self-control, and measures of socio-economic status. We can’t simply tell each other to go and learn new self-control strategies and be done with it – though I definitely do endorse that we try to pick up some as we go along with our lives. One reason why learning self-control strategies is harder than just learning that the capital of Italy is Rome is that the steering of our own behaviour is deeply dependent on a group of cognitive mechanisms termed executive functioning. Executive functioning denotes a diverse set of cognitive functions, from working memory to inhibitory control. While situational strategies of self-control are less dependent on executive function (EF) than the brute inhibition of “just say no”, robust EF capacities remain helpful across the various things we call self-control. If one has difficulties in EF, difficulties in self-control follow.

It is probably unsurprising that EF, too, is correlated with measures of SES. Is it then, that low EF and not low self-control is the cause of the socioeconomic disadvantage that the self-control studies pick up?

Not so fast. Instead of thinking of a single causal arrow from EF to disadvantage, I am going to suggest a whole horde of causal arrows, a complex causal interaction where our neural and cognitive function is impacted by our lived circumstances and choices, and vice versa.

To do so, it’s helpful to look into established causes of interpersonal differences in executive functioning. Established causes and correlates of differences in executive function include stress, trauma, many forms of somatic illness including somatic pain, lack of sleep whether chronic or temporary, environmental toxins, malnutrition, many forms of psychopathology including Major Depressive Disorder, Bipolar Disorder, Schizophrenia, Post-Traumatic Stress Disorder and Attention Deficit/Hyperactive Disorder, and prenatal and postnatal care and parenting. (This list is by no means exhaustive.) The causal structures here are difficult to pin down. However, it is clear that EF is impacted by many different factors, most of which are more frequently encountered in conditions of hardship.

Explaining the fact that measures of self-control cluster together with disadvantaged life outcomes with a single causal arrow from self-control to SES is just plain false. But  given that executive functioning is a general-purpose set of capacities that we need to navigate complex life situations, it would also be dishonest to try to ignore the SES-EF or SES-self-control correlations or to reject any talk of these correlations as classist. The way these differences are, by and large, described in the literature is oppressive. But to do justice to the profound ways in which disadvantage creates differences in human agency, we need to understand how general the effects of disadvantage can be and why it is that socio-economic disadvantage can lead to difficulties in areas that, at first brush, seem not to be about money.

I am a Senior Researcher in ethics in the Nudging for Climate research consortium, at the University of Turku, Finland. My chief philosophical interests are agency and self-control, psychiatry, neuroethics, and behavioural policy.