Understanding Differences in Health Behaviors by Education

Why do better-educated individuals consistently exhibit healthier behaviors—such as lower rates of smoking, obesity, and heavy drinking—compared to those with less education? This paper investigates the mechanisms behind this "education gradient" by testing whether it stems from differences in material resources, specific health knowledge, cognitive ability, or underlying psychological tastes like risk aversion and discounting.

Using multiple large-scale datasets from the United States (including the NHIS, NLSY, and HRS) and the United Kingdom (NCDS), the authors systematically add potential explanatory variables to regression models to see how much they reduce the observed effect of education on health behaviors. They categorize these potential mechanisms into four groups: material resources (income, insurance), information (specific knowledge), cognitive ability (general processing skills), and utility function characteristics (discount rates, risk aversion, personality).

• Material Resources: Income and economic status account for roughly 20 percent of the education gradient.
• Cognitive Ability: This is a major driver. Education appears to enhance general cognitive processing skills, which in turn improves health behaviors. This factor accounts for approximately 30 percent of the gradient.
• Knowledge and Social Factors: Specific factual knowledge about health risks and the quality of one's social network each account for about 10-12 percent of the gradient.
• Utility and Personality: Contrary to common economic theories, measures of risk aversion, time discounting, and personality traits (like self-control or depression) explain virtually none of the education gradient.

This research suggests that the health benefits of education are not merely about having more money or knowing more facts. Instead, education appears to fundamentally improve an individual's ability to process information and navigate complex health decisions. These findings imply that public health interventions might be more effective if they are designed to accommodate different levels of cognitive processing rather than simply providing more information.