| The goal of this investigation was to evaluate the construct validity of psychometric prediction of liability for schizophrenia. Information was integrated from three sources: clinical diagnosis and phenomenology, psychometric assessment, and degree of genetic liability for the disorder. A sample of 244 twin pairs from the Washington University (St. Louis) Twin Study previously had been given consensus DSM-III diagnoses based on structured interviews. These data were studied for heritability of various definitions of schizophrenia and related disorders, as well as for a variety of psychotic symptoms. The most heritable definition of schizophrenia included DSM-III schizophrenia, schizoaffective disorder, and schizophreniform disorder. Symptoms of hallucinations, delusions, and emotional withdrawal were found to have a substantial genetic basis, with no apparent influence of shared family environment. About 60% (151) of the pairs had complete MMPI data, which were scored for a variety of special schizotypy and affective scales. Sixteen of 25 measures studied significantly distinguished schizophrenic twins from individuals with other diagnoses. Of greater import, two measures significantly distinguished the nonschizophrenic cotwins of schizophrenics from the cotwins of probands in other groups. One measure, the 8-6K scale, was found to predict a substantial portion of the variance in diagnosis. Structural modeling analyses of the biometric basis of this prediction suggested that all of this predictable variation was due to genetic sources. Although these results are preliminary, this study suggests a need for further investigation of these psychometric indicators as markers of genetic liability for schizophrenia. |
