James S. Fleming

Measures of Transgender Behavior

Using factor analysis, we sought to identify the components of transgenderism. Subjects were 455 transvestites and 61 male-to-female transsexuals, all biological males. A 70-item questionnaire was used, along with other structured questions concerning preferred and usual sex partners. Five factors were identified and interpreted: Transgender Identity, Role, Sexual Arousal, Androallure, and Pleasure. These factors represent the most salient dimensions of transgenderism. All five-factor Means for transvestites and transsexuals differ. An examination of overlap of group distributions for each factor showed such overlap to range from only 6% for Identity to 46% for both Androallure and Pleasure. Factor intercorrelations for the obliquely rotated factors ranged from −.37 to .27. While transvestites and transsexuals have different lifestyles, their transgender cognition and behavior seem constructed upon different combinations of the same variables. A second-order analysis of these five factors yielded two factors: Sexual Arousal loaded highest on the first factor (.91), and for the second the highest loading variable was Androallure (.57). Each of these highlights the primary importance of sexual arousal in transgender cognition and behavior. Studying possible age effects, we found that the younger versus older transvestite groups had significantly different scale Means for Androallure and Pleasure; there were no age differences between older and younger transsexuals on any of the five scales. Six percent of transvestites reported a male as their usual sex partner; 25% of the transsexuals reported a female as their usual sex partner. For each group, one-third indicated their usual sex practice was without any partner, while only 5% said they preferred this practice. We propose that the five variables identified offer a comprehensive approach to the description of individual differences in transgender experience and expression.

A Test of Shavelson's Hierarchical Multifaceted Self-Concept Model in a Filipino College Sample.

Shavelsons hierarchical, multifaceted self-concept model (HMFM) has proven useful to educational and psychological researchers. The reliability and factorial validity of a well-known HMFM instrument - the Fleming-Courtney Self-Rating Scales (SRS) - were examined for a sample of 198 Filipino college students. Confirmatory factor analysis compared six structural models. A six-factor, oblique model that included two social factors (Social Acceptance and Social Anxiety} fit the data best, and a second-order factor analysis supported the hierarchical structure of the SRS. The men were significantly higher than the women on three self-concept scales, but the only large effect was for Physical Abilities, consistent with findings in a U.S. college sample. Results provided some support for the utility of both the SRS and HMFM for Filipinos.

The personal and academic self-concept inventory : factor structure and gender differences in high school and college samples

The Personal and Academic Self-Concept Inventory (PASCI) measures global, social, physical, and academic components of self-concept, as well as social anxiety. Several competing structural models were tested by confirmatory factor analysis for a sample of high school (N = 222) and college (N = 338) students. Seven oblique factors which included two social factors (Social Acceptance and Social Anxiety) fit the data best, and a second-order model supported the hierarchical structure of the data. Internal-consistency and test-retest measures indicated that the scales were reliable. Girls and women scored lower than the boys and men on Physical Ability and on Math Ability, consistent with other findings. Self-Concept scales were stable across high school grades, but global self-concept was higher in the college group, also consistent with prior research.

An Index of Fit for Factor Scales

An index is defined for each column of a factor matrix to measure the goodness of fit for scale defining items, where the latter are determined in advance by the investigator. Let MS(S), MS(NS), and MS(T) be the mean-square loadings for scale items, nonscale items, and total items, respectively, for a given factor. A measure denoted IFFS (index of fit for factor scales) is defined as the signal-to-signal plus noise ratio: IFFS = 1 - MS(NS)IMS(T). The highest possible value of IFFS is 1.00, with .50 indicating that scale items are no better than nonscale items in defining the construct. An overall measure is also defined. The index is complemented by Kaisers index of factor simplicity (IFS) as computed for each scale item. Evaluative criteria and examples are presented.

FACTREL : a general program for relating factors between studies via the Kaiser-Hunka-Bianchini method

FACTREL relates factors from two studies with the same variables using Kaiser, Hunka, and Bianchinis (1971) method. Both test and factor vectors for the two sets are placed in a common space, then one is rotated to maximize the sum of the row cross-products of loadings between the sets, with the resulting cosines between factor axes serving as the primary measures. Cosines between corresponding tests are also computed, as is their average. This method of relating factors is potentially useful in meta-analytic studies when only the factor loadings (and in oblique cases, factor correlations) are available in the published literature. Written in FORTRAN-77, FACTREL is a general, interactive program that handles both orthogonal and oblique cases.

Spreading the word: funding faculty to train faculty in the California State University

The mission of the California State University is to furnish a quality learning environment for its teaching, research, and public service functions. The primary teaching function is transference of knowledge from faculty to students. The structure and size of the CSU, and the availability of funds, are not typical of the average institution. Nonetheless, we believe that the model presented below is potentially adaptable by others. Many colleges and universities either belong to instructional or computing consortia, or at least have the option of joining together in a formal or informal way to address common needs. On a smaller scale, the model can be implemented even at the level of the single campus, or perhaps targeted for a discipline within the university. When funds are in short supply, but space and equipment are available, faculty release time may be possible (with the good will of the administration), and your local Computer Center may also be able to contribute with resources and people to assist.