This study aimed to determine the dimensionality of concepts of schizo
phrenia using 11 different diagnostic systems, and then to identify th
e nature of these dimensions by their relationship to a range of signs
and symptoms. The sample consisted of 479 patients admitted with a fi
rst episode of functional psychosis. The underlying structure of the 1
1 diagnostic systems was best represented by an oblique 3-factor solut
ion. Whereas the second and third factors could be meaningfully interp
reted by their correlations with signs and symptoms, the first factor,
anchored by 'modern' nosologies such as DSM-III-R, was more clearly s
pecified by what it is not (the absence of affective symptoms) rather
than by what it is (the presence of characteristic psychotic symptoms)
. A logistic regression of DSM-III-R diagnosis on to separate diagnost
ic components supports the contention that duration of illness and aff
ective exclusion criteria discriminate the presence of DSM-III-R schiz
ophrenia much better than the three characteristic psychotic symptom g
roupings.