Studying the meaning of dreams is difficult and subject to many methodological pitfalls. One of these is vagueness and inconsistancy in defining dream content variables that are used as independent measures. This study evaluated the effect, upon accurate definition of the independent variable, of repeating the dream report and of using associative content as dream material. The importance of controlling these two sources of dream material was demonstrated: when the patient was allowed to repeat the dream and to report associative content, the number of initial references to death and separation variables nearly doubled in comparison to the standard single report of manifest content. Furthermore, there was minimal or no correlation of the number of these variables between manifest content and associative content, providing additional evidence for the unique contribution of associative content. This article also describes a new interview technique, the Staged Interview Technique, which provides a means to control, as well as measure, these potential distortions of the independent variable.