Revisiting Internet Use and Depression Among the Elderly

Abstract

Four years ago in a paper entitled Internet Use and Depression Among the Elderly, my colleague and I presented evidence indicating that Internet use reduces symptoms of depression among older Americans (55 or older). This showing was based on the sample of over 7,000 retired persons taken from the Health and Retirement Study (“HRS”) conducted at the University of Michigan. Applying a variety of advanced empirical techniques, we found that Internet use is associated with a 20% reduction in depression classification based on a commonly used measure of depressive symptoms (a eight-item CESD score). In this PERSPECTIVE I present the results of an update to the earlier analysis that uses additional data and more sophisticated econometric techniques. Specifically, while the data we used in our first paper was large, the HRS is nonetheless a longitudinal survey. Since depression, at least in some forms, is often a recurrent condition that may persist over a number of years, we decided to exploit the time series component of the HRS to study the role of past depression and Internet use on current depression in older adults. To do so, we use a dynamic probit model, which in some cases is enhanced by matching algorithms. This updated analysis of the data again reveals that Internet use reduces depression categorization and does so by 34%, a reduction similar to, but slightly larger than, that found in our earlier research (20-28%). Thus, consistent with our earlier findings, as depression is estimated to cost the United States about $100 billion annually, the data indicate that policies which expand Internet use among the elderly may have significant economic payoffs.