Response to Hunsberger

misinterpreted some of our assertions. This brief note is a response to Hunsberger's note, with responses to each of his points in the order in which he presented them. First, we did not suggest that our factor analysis may have differed from that reported by Fullerton and Hunsberger due to the use of BMDP rather than SPSS. We described our use of BMDP in detail only in response to comments of a reviewer who felt that the paper would be strengthened by more adequately indicating our procedures and methods. Second, Hunsberger states that much of our rests on the assumption that rotation of the factor matrix is necessary to adequately assess the data. He apparently misinterpreted our argument since we subscribe completely to Nunnally's guidelines for rotation. As stated in our paper, Rotation simplifies the structure of a factor matrix, thus enabling the researcher to choose a solution which clearly identifies the direct cluster of variables which form the factor or (331). Moreover, the thrust of our paper was not the need for rotation, but rather the difference between Fullerton and Hunsberger's finding of one factor vs. our two factor result. Hunsberger correctly points out that the determination of the number of factors to be interpreted can be made by one of several methods, and lists the following options: 1) using only factors with an eigenvalue greater than 1.0; or 2) performing a Scree test to evaluate whether there is a large enough gap between eigenvalues to justify discarding the smaller factors. He also points out that there are no firm guidelines in deciding how many factors one should retain and rotate after the initial solution. We agree with all of these points. We also note that in some cases the researcher may wish to use only the strongest factors which would be most easily interpreted. In other cases, the last and weakest factors, although they may have small eigenvalues, may represent the most interesting and unexpected findings in the research, and therefore may be retained. The latter was the case in this research. In our paper, we outlined a two dimensional analysis which we feel sheds light on the concept under study. Both factors have eigenvalues greater than 1.0. The first factor appeared to represent a traditional aspect of Christian Orthodoxy. The second factor, although explaining much less of the variance than the first, appeared to have substantive meaning (discovered through rotation of the factor matrix) as a moder, rational form of Christian Orthodoxy. These two identifiable dimensions provide a multidimensional

Philosophy · Scientific Research and Discoveries · Statistical and numerical algorithms