Participation in Learning and Depressive Symptoms

Abstract

This paper reports the findings of research on relationships between depression and participation in learning using data from a large sample of older adults. The objective was to establish whether learning can reduce the risk of depression. Data were obtained from the English Longitudinal Study of Ageing, a nationally-representative sample of adults aged 50 and above. The eight-item Center for Epidemiologic Studies Depression Scale (CES-D) was used to measure depressive symptoms. Participation in learning covered the acquisition of qualifications; taking formal training courses, learning at arts, music or evening classes; and participation in gym/exercise classes. Multiple regression analyses were conducted with the change in the measure of depressive symptoms between two waves of the longitudinal survey as the response variable. There was limited evidence that participation in learning reduced the risk of depression in later life. Only learning leading to qualifications was significantly associated with the outcome after controlling other factors. But learning leading to qualifications was undertaken only by a small minority of the sample, mainly the youngest and most highly educated. Other types of learning were not significantly associated with depression. It appears that learning has, at best, a small role to play in addressing the risk of depression in later life.

Notes

Note. Learning activity at Wave 2.

Notes. Absolute values of t statistics in parentheses.

*significant at 10%; **significant at 5%; ***significant at 1%.

Both models contain a full set of controls, including depression score at Wave 1, confounding factors at Wave 1, and measures of change in circumstances between Wave 1 and Wave 2. All models weighted for nonresponse.

Notes. t statistics in parentheses.

*significant at 10%; **significant at 5%; ***significant at 1%.

All models contain a full set of controls, including depression score at Wave 1, confounding factors at Wave 1, and measures of change in circumstances between Wave 1 and Wave 2. All models weighted for nonresponse.