Recent research on unemployment and health has depended heavily on macro-social and ecological data analyses. In order to redress this imbalance, the report summarizes the methods and results of a prospective study of a permanent factory shutdown, thereby illustrating and dissecting some of the strengths and weaknesses of the micro-epidemiological level of analysis. This leads to a broader analysis of research strategy questions which consider macro and micro approaches and variations on the two.

The interdependence of the three areas of patient improvement assessments used in a comparative study has been studied. The three dimensions are clinical psychopathology, social dysfunction and burden on family. A high correlation has been found between these dimensions. The differential effect of two modalities of treatment delivery on the interdependence of these three dimensions is discussed.

This paper introduces statistical methods suitable for the analysis of response, survival or failure times and, in particular, latencies measured in experiments on the speed of recall of memories. The discussion includes the use of simple descriptive statistics, as well as an explanation of the role of linear-logistic and log-linear models.

This paper discusses methods for assessing speech dominance in cases for unilateral ECT. The use of formal laterality tests has been strongly advocated in the literature, but is seldom adopted in clinical practice. Assessment of handwriting type (left/right, inverted/non-inverted) has been put forward as an alternative, but is not valid. Consideration of handedness alone is the most common practice, but is of little value. From current evidence on the relation of handedness to speech dominance, the paper evaluates the likely outcomes of always administering right-sided ECT, and of giving formal tests to left-handers only.