T hompson's article on Openmindedness and Indiscriminate Antireligious Orientation (1974) presented some excellent evidence on the relationships among Dogmatism and intrinsic and extrinsic religious orientations and raised some important conceptual questions. The size and selection of the sample are particularly to be commended. Thompson is quite right to insist that the Allport-Ross (Allport, 1968) intrinsic and extrinsic religion subscales constitute two separate continua instead of a single bipolar continuum. He is also correct in his suggestion that Strickland and Shaffer (1971) may have found no significant relationship between extrinsic religion and authoritarianism because they had confounded the intrinsic and extrinsic orientations in their measurement. When I measured intrinsic and extrinsic religious orientations separately, the extrinsic dimension was consistently related to authoritarianism, and the intrinsic dimension was also related positively but less strongly and less consistently (Kahoe, 1974). However, I should make three kinds of comments: (a) to question Thompson's employment of among the religious orientations, (b) to point out an unrecognized or unacknowledged response style problem that calls into question his major conclusion, and (c) to present some data supporting the validity of the intrinsic religion scale among Catholics. Thompson deviated from normally preferred practice in psychology when he converted the two independent religious dimensions into types-intrinsic, extrinsic, indiscriminately proreligious, and indiscriminately antireligious. While it may be useful at times to classify people into types on the basis of independent continua, it is conceptually sounder and more common to conceive and study separate traits that vary independently of one another. Thompson attributed his scoring procedure to Hood (1971), but whereas Hood did present persuasive evidence for the psychometric independence of the intrinsic and extrinsic dimensions, he did not render them into the typologies Thompson used. As I read Hood, anyway, he advocates the kind of independent ordering of the variables that I am arguing for here. Thompson's own data actually make more sense to me when viewed in those terms, and I think some of his conceptual problems arise out of his forcing a typology, Allport's precedent to the contrary notwithstanding. Toward the end of his article Thompson did report the correlations between I)ogmatism and the intrinsic and extrinsic scales separately. These statistics are strikingly similar to some of my own data that Thompson had no access to. With a sample of 333 college freshmen, about 85% Baptist, extrinsic religion correlated .30 with I)ogmatism, compared to .28, .28, and .38 for Thompsons's subsamples;