Is Self-Perceived Health Defined Differently Depending on Education? An Application of Machine Learning Techniques

Jordi Guma-Lao , Centre for Demographic Studies

Self-perceived health is a subjective health outcome that summarizes all the health conditions and is widely used in population health studies. However, it is still unclear as to which health conditions are actually taken into account when making an individual assessment of one’s own health. In addition, this health outcome has been used to explore health inequalities by education. However, education has also been shown to be associated with the knowledge that individuals have about their own health status and its evaluation. The aim is to assess the influence of objective health conditions in predicting responses to questions about self-perceived health among European women and men between the ages of 50 and 64 according to education using data from the 6th wave of SHARE survey. First, random forests analysis will be used to identify the health conditions with a higher predictive capacity for self-perceived health, second, classification trees (J48 algorithm) will be applied to predict self-perceived health according to education and sex. Preliminary results confirm the high predictive capacity of self-perceived health based on only four of the most widely used health indicators in the literature, namely chronic diseases, IADLs, ADLs, and depression. Future research will confirm that with these four health outcomes is possible to obtain predictions of self-perceived health with a high level of accuracy. Thus, after confirming the evidence of the sex differences in the path to predicting self-perceived health, the next step will be to stratify the analysis according to education for women and men.

See extended abstract

 Presented in Session 52. Education and Health