ABSTRACT
W e next turn to exam ine the relationship o f the count o f depressive sym ptom s to a risk index defined by a sum m ary scale o f 26 so cial, fam ily, and life event risk factors for p sy chopathology. H ere we observe a quite d iffer ent pattern. In this case, we see, in F igure 6, w hat looks like a genuinely dim ensional rela tionship. N o cut-point on the predictor risk factor scale or the depression scale produced a larger test statistic for the association betw een these two variables than the linear association o f the two (m eans ratio from Poisson regres sion = 1.14; 95% C I = 1 .10-1 .19 ; p < .0001). W hat w e see is that the “sam e psychopathol ogy” m ay apparently behave both categori
cally and dim ensionally in relation to d ifferent risk factors.