This paper examines recurrent continuous morale in old age within a statistical modelling paradigm. The proposed statistical modelling relates recurrent morale to a set of explanatory variables which includes subjective as well as objective measures. In order to assess the degree to which explanatory variables influence morale, an adequate statistical model must handle the possibility that substantial variation between respondents will be due to unmeasured and potentially unmeasurable variables (residual heterogeneity), multicollinearity and past behaviour effect. A regression model with a \"variance component\" specification handles residual heterogeneity by incorporating individual-specific normal error within the modelling framework separately from the structural error.

However, despite adopting a mixture model specification, assumption violation is still possible due to the inclusion of the subjective variables. One solution to such problems is the application of instrumental variables. The instrumental variables method is seldom used with social survey data. The main criticism is the arbitrary selection of variables as instruments. Longitudinal data, because of their temporal structure, provide natural instruments. This property of panel data is incorporated within longitudinal modelling to correct for model mis-specification. This paper further discusses the theoretical implications of the adopted pragmatic approach to improve control to account for time varying omitted variables (such as the unobserved effects of underdiagnosed and under treated depression among the elderly living in community).

These applications are illustrated using morale in old age from a longitudinal survey of the elderly. The results suggest a strong presence of heterogeneity effect, i.e. current levels of morale appear to be individual-specific and independent of its previous levels.