ABSTRACT
The population groups selected for longitudinal studies may be of two kinds: either cohorts or geographically determined population groups. Pure cohorts should preferably be birth cohorts, since all other kinds of cohorts constitute a selection of survivors. Birth cohorts investigated for psychiatric purposes were those of J. Klemperer and K. H. Fremming, differing with regard to yield, Klemperer being able to identify only about 50 per cent of his probands, as compared to Fremming’s 92 per cent. General Paresis of the Insane was the most frequent of those psychiatric disorders which, during the second half of the nineteenth century, was regarded as having an organic origin. The cohort of schizophrenics which served as a basis for the cancer-schizophrenia study was, as mentioned, derived from a nationwide census in 1957 of hospitalized psychiatric patients in Denmark.
