[Confounding effects in social science studies and their control. The examples of personality, social support,...

Social scientific investigations on coping with illness, on the relevance of life events or social support for psychic disorders are concerned with a variety of constructs with interdependent structures. Hypotheses in interrelationships are set up and tested on the construct level. A problem quite often overlooked here is that almost all constructs overlap to a considerable extent both with regard to their content as well as operationally, i.e. they are confounded. Interrelations of constructs which are strongly content-related or operationally confounded are thus in danger of demonstrating artificial results. Data from an extensive epidemiological research project on the incidence and prevalence of psychic disorders in the general population are used applying structural equation models for latent variables to demonstrate confoundation effects for the constructs personality, social support and state of health. Appropriate control methods are described. The results show that partner support maintains a strong influence on the psychic and physical state of health even confoundation of content has been excluded. Application of this method of content confoundation control is discussed with regard to research on coping life event research etc.