assume that the Rorschach is a valid instrument for determining adjust ment in the way in which we have defined it? If so, then clearly there is no inherent connection between path ology and homosexuality. But caution is needed. As clinicians, we are well aware, in daily practice, of the limi tations of projective material anal yzed "blind". Nevertheless, the quantitative results are striking, and they are confirmed in part by observations of the judges, as well as and I say this with great caution by life history data.

But let us look at the results in the

second task given the judges, that of distinguishing between matched pairs of homosexual and heterosexual records. This is a much easier task than that which the clinician ordinarily faces, of identifying homosexuality in one record out of many; and yet it proved to be a very difficult one. As a judge compared the matched protocols, he would frequently com. ment, "There are no clues; or, "These are so similar that you are out to skin us alive;" or, "It is a forced choice," or, "I just have to guess." The difficulty of the task was reflected not only in the comments of the judges but also in the results. Judge "A" correctly identified 17 of the 30 pairs, and Judge "B" 18 of the 30. Thus neither judge was able to do better than chance. In seven pairs both judges were incorrect, that is, identifying the homosexual as the heterosexual, and vice versa; in twelve pairs, correct; and in the remaining eleven they disagreed.

Let us look at the problems the judges faced. In some pairs of records none of the clues usually considered to be signs of homosexuality occurred. In some pairs the "homosexual clues" appeared in both records. These "homosexual clues" were primarily anality, open or disguised; avoidance of areas usually designated as vaginal areas; articles of feminine clothing. especially under-clothing, and 'or art

objects elaborated with unusual detail; responses giving evidence of con. siderable sexual confusion, with cas tration anxiety, and or hostile or fearful attitudes toward women; evidence of feminine cultural identification, and or emotional involvement be. tween males. When these clues ap peared in neither or in both records, the judge was forced to look for other evidence, and most frequently depended upon peculiar verbalization, or responses with idiosyncratic meaning, or the "flavor" of the total record. When careful examination failed to reveal anything distinctive, the judge assumed that the more banal or typical record was that of the het erosexual, an assumption which was sometimes false.

After the judging was completed, and, indeed, even while it was in process, both judges commented on the fact that the records which they thought to be homosexual were un like the ones they were familiar with in the clinic. They were not the dis turbed records ordinarily seen. One judge, in the process of choosing, said, "It begins to look as if the homosexuals have all the good things: for example, M's and Fc." It may be pertinent to reiterate that I had made an effort to secure records of homosexuals who ordinarily would not be seen in a clinic. A discussion of the validity and reliability of homosexual signs is tangential to this symposium', but I would point out in passing that my data indicate the need for a thor ough-going reconsideration of this problem. At a minimum, healthy skepticism about many (but not all) so-called homosexual-content signs in the Rorschach is, I think, called for. The inability of the judges to dis tinguish the homosexual from the heterosexual records better

than

A paper on "Homosexuality in the Ror schach" is in process of preparation. It will contain a full discussion of homosexual signs, as well as other aspects of homo sexuality in the Rorschach

would be expected by chance fits, I think, the finding on adjustment of the two groups. Some of the records can be easily distinguished; the fact that the judges agreed in their identification of twelve pairs indicates this. These were records of individuals with strong emphasis on "femininity” and or anality. But apart from these, which constitute about a third of the group, the remaining two-thirds cannot be easily distinguished. If the homosexual records had been similar to those frequently seen in the clinic, that is, severely disturbed, there might have been greater probability that they could have been correctly identified, although this cannot be said with certainty. I have now seen about two hundred homosexual records and would be skeptical about my ability to identify correctly records. similar to many in this group.

Although it is not pertinent to this symposium3 to present in detail the findings of the statistical comparisons of the two groups of Rorschach protocols, it is relevant to point out in summary form that most of these comparisons have failed to produce differences of sufficient magnitude to satisfy tests of significance. Several examples will suffice to make the point. Although most studies of homo sexual protocols indicate greater pro ductivity on the Rorschach, the dif ference between the two groups in the present study does not reach sig. nificance, though there is a trend in this direction (t=1.389, df= 29, p= >.10). A detailed comparison of total M's and human figures was made. Of some 25 computations, of differences between means of M% in various categories (such as flexor or extensor), differences in form level, variation in form level, etc., the only ones which approached low significance were the sigma of form level (t=1.98, df=29, p=.05), and O-minus percent (t= 2.262, df=29, p=<.02).

Cronbach's warning about inflation

• See Footnote 4. •

of probabilities deters me from drawing too many conclusions from these two findings, although there is good theoretical rationale for them. The details of the analysis will be discussed more appropriately in a later paper. I cite these general findings at this time in order to show that despite considerable effort and the pursuing of many alluring possibilities, the ef

forts thus far to establish clear cut dif ferences between the two groups as a whole have been relatively fruitless. This, too, is consistent with the lack of significant differences between the adjustments of the two groups.

Ín addition to the overall adjustment ratings, each judge gave sum mary statements about each subject in a number of categories, including methods of handling aggression, af fectional and dependency needs, and form of impulse control. When these statements were tabulated and subjected to statistical analysis, again no clear-cut differences emerged. For example, the statements about affec tional and dependency needs have been tabulated in eleven categories, such as repressed or absent, ego-alien, integrates well, controlled by (that is, a dependent character). Four homosexuals were described as having af. fectional and dependency needs repressed or absent, while three heterosexuals were similarly described Six homosexuals and six heterosexuals were described as integrating well these needs. It was said of one homosexual and one heterosexual that af fectional and dependency needs were ego-alien. Chi square for differences

between the number of heterosexuals and homosexuals assigned to all cate gories is 5.736, df=10, insignificant. (Part Two of this article will appear in the January issue of Mattachine Review.)

The complete data will be reported in the future publication previously referred to.