This paper reports a re-evaluation of seven sets of findings appearing in the research literature. The findings were derived from questionnaire studies of feelings of guilt and other feelings in diagnosed depressives and other psychiatric groups. The publications did not provide sufficient information to evaluate important inconsistencies in the findings. However, an examination of the structure and content of the questionnaires which were used made it possible to account for the discrepancies. There were three main conclusions. The first was that the only way to minimize the ambiguity of questionnaire responses was to avoid the use of multiple-item scales and only to use single-item questions which were constructed as simply as possible. The second was that depressives were not much different from other people in the degree of outwardly and inwardly directed hostility. The third was that, while guilt feelings, with certain provisos, tended to be associated with the diagnosis of depression, other feelings appeared to be both more prominent in and more strongly associated with depression. These feelings consisted of increments in feelings of loss, in the most general sense of the word, and decrements in general positive pleasant feelings.
IntroductionNegative feelings about the self are widely regarded as an important feature of depression. Data that might throw light on this notion can be found in published studies that have made use of questionnaires to elicit feelings. As a first step, it was decided to examine studies in which the feeling of guilt, and some of its synonyms, had been assessed. Other, non-guilt, feelings were also assessed in most of these studies. The opportunity was taken to compare the results for these other feelings with those for guilt and its synonyms. Seven relevant sets of data, presented in six different papers (Foulds et al., 1960;Friedman, 1964;Laxer, 1964;Foulds, 1965; Harrow et al., 1%6; Harrow & Amdur, 1971), were studied. This paper makes a largely independent re-evaluation of these data.