Abstract
The relative predictive strengths of eight weather factors were examined using as separate dependent variables monthly figures for community mental health intake, welfare caseload, calls to a telephone hotline, medical patient caseload, felony arrests, juvenile complaints, drunk-driving arrests, and mortality rates in two non-urban areas of Ohio. Z-score transformations of subjective discomfort of the weather factors as indicated by three independent samples were analyzed with a stepwise multiple regression. With the exception of hotline calls, each of the social indicators in the two localities was significantly predicted by a different weather factor, and the weather factors, taken together, accounted for about 10% of the variance of each social indicator. For each geographical area, combined weather factors accounted for no more than 30% of the variance of any local social indicator. Problems of overgeneralization and exaggeration of the effects of weather factors on social indicators in previous studies were discussed. A balanced approach to behavioral effects of geophysical variables must be achieved (Ammons, 1978).