Our research has several limitations. First, our
sample was not a representative sample of the entire
population of people with schizophrenia and
diabetes. In addition, our sample tended to include
those with relatively stable, less severe symptoms
because we only included subjects who were
allowed to participate in this study by their
psychiatrists. Second, the number of subjects in
this study was small (N = 38), which may not be
enough to identify some real associations.