One way to study dream function is to use dream content as an independent variable. Erroneous or inconstant definition of the dream content, however, would undermine its use as a predictor. This study describes a dream interview technique that allows assessment and control of the distorting effect from using, in addition to a single report of manifest content, associative dream content and repetition of the dream report. Dream reports obtained by this technique were rated for the number of references to the following seven variables: death, mutilation, separation, illness, guilt, shame, and falling. The data showed considerable variability in designating the presence or absence of all seven dream content variables, depending on how the dream report itself was defined. It is concluded that to use dream content as an independent variable in the study of dream function, associative content and repetition of the dream report must be systematically controlled.