Ideological concerns in the operationalization of homophobia, Part II: the need for interpretive sensitivity with conservatively religious persons.

Rosik, Christopher H.

Building on the first article in this series (Rosik, 2007), the present study provided empirical analyses to determine the degree to which the relationship between conservative religion and homophobia as defined by Herek's (1998) Attitudes Toward Lesbians and Gay Men Scale (ATLG-R) was dependent upon items experienced as antireligious by Christian students. Three multiple regression analyses revealed that the associations between homonegative attitudes and respondents' intrinsic religiousness, religious practice, and beliefs about the authority of the Bible were predicted only by the "Condemnation-Tolerance" component after accounting for gender, age and the remaining components of the ATLG-R. These findings suggest the possibility of an ideologically based circularity in the relationship between conservative religion and the construct of homophobia as measured by the ATLG-R. Thus, for these respondents the ATLG-R may function as an empirically packaged method of disparaging their religiously-based values concerning homosexuality. It is requisite that mental health professionals cultivate greater sensitivity to such concerns.

**********

In the study of homophobia, one of the most robust findings has been the association between theologically conservative religion and homonegativity (Finlay & Walther, 2003; Herek, 1994; Johnson, Brems & Alford-Keating, 1997). People with intrinsic religious orientations, where faith is the central organizing principle of personal identity, have been found to be highly homophobic, although they do not appear to be racially prejudiced (Herek, 1987, 1994; Wilkinson, 2004). This was in contrast to persons with an extrinsic religious orientation, whose religious identification serves other more primary social or personal goals. Extrinsic individuals have been found to hold prejudice toward both homosexuals and ethnic minorities (Herek, 1994; Wilkinson, 2004).

As Rosik (2007) noted, ignorance of ideological differences between scales measuring homophobia and conservative religious traditions can result in findings that obfuscate rather than clarify the likely complex relationship between these variables. What limited research that has been done to investigate these possibilities appears to suggest complexity and the importance of interpretive sensitivity in considering homophobia among religious persons. Several studies have indicated that some religiously devout individuals distinguish between the value of gay men or lesbian persons and the value of their sexual behavior in ways that have meaningful implications, a distinction that virtually all scales of homonegativity obscure but which has clear legitimacy within a conservatively religious ideological surround (Bassett, et al., 2000; Bassett, Angelov, Mack, Monfort, Monroe, & Rosik, 2003; Bassett, Kirnan, Hill, & Schultz, 2004; Bassett, van Nikkelen-Kuyper, Johnson, Miller, Carter, & Grimm, 2005; Fulton, Gorsuch & Maynard, 1999; Rosik, Griffith & Cruz, 2007). Clinically important information may be lost when homophobia is used in a blanket fashion to describe clients who come from traditional religious communities.

Fulton, Gorsuch & Maynard (1999) discovered that the homonegativity of the most religiously devout students, those with higher intrinsic religious orientations, were more nuanced than that of the more extrinsically religious participants. Intrinsically oriented individuals tended to limit their homonegativity to the moral dimension of their evaluation and did not necessarily restrict their social contact with gay men and lesbians to any greater degree than they would with other disapproved groups (e.g., liars, overeaters, alcohol abusers). These persons apparently were comfortable being "in the world but not of it," while the extrinsically religious participants were less concerned about ideological consistency than about maintaining effective social boundaries. Their homonegativity was more broadly based and they tended to seek greater social distance from gay men and lesbians than from other disapproved groups.

In a similar vein, Wilkinson and Roys (2005) recently studied 180 college students to determine if they could distinguish between homosexual behavior, fantasies and emotions. Results indicated that gay men and lesbians were rated more negatively when they were described as engaging in sexual behavior than when the reference was only to having sexual fantasies or homoerotic feelings. Participants' religiosity appeared to be responsible for these differences in evaluations for gay men, but not for lesbians.

In the previous article in this series, Rosik (2007) identified four factors in Herek's (1998) Attitudes Toward Lesbians and Gay Men Scale Revised (ATLG-R) for a sample of conservative Christian undergraduate students. He then submitted the 20-item ATLG-R to an ideological surround analysis (Watson, Morris & Hood, 1992) and found four items registered as ideological neutral, four were viewed as proreligious and 12 were experienced as antireligious. Herek's primary "Condemnation-Tolerance" factor was found to consist predominately of items evaluating the morality and naturalness of homosexuality. Furthermore, all of the items forming this component were scored in a manner experienced as inconsistent with respondents' religious tradition (i.e., antireligious).

The present study extends Rosik's (2007) prior analysis by empirically examining the relationship of the ATLG-R's four factors to religious behavior, intrinsic religiousness and beliefs about the Bible to highlight potential ideological influences in conclusions about homonegativity. Additional implications for interpreting the link between homophobia and conservatively religious persons will conclude the study.

Study Predictions

Based on the previously identified ideological nature of the ATLG-R items and the valuing person versus behavior distinction of traditional Christian sexual morality, the following two predictions were explored:

1. Religious commitment, intrinsic religiousness and belief in the Bible as being authoritative would be significantly predicted by ATLG-R component 1, as it is comprised by items consistent with conservative Christian normative values about the morality and naturalness of homosexual practice that the scale identifies as homophobic ("antireligious items").

2. Religious commitment, intrinsic religiousness and beliefs about the Bible as being authoritative would not be significantly predicted by ATLG-R components 2, 3 or 4 as these factors contain items that generally reflect how to relate toward homosexual persons in society and contain a mix of ideologically neutral, proreligious and antireligious items.

METHOD

Participants and Materials

A description of the study methodology can be found in Rosik (2007). The following additional measures were also employed in the present analysis:

Religious Commitment Inventory (RCI-10). The RCI-10 is a 10 item brief measure of religious commitment developed by Worthington and colleagues (Worthington, et al., 2003). The inventory is based on Worthington's (1988) theory of religious values, where religious commitment is defined as the degree to which a person adheres to his or her religious values, beliefs, and practices and uses them in daily life. For example, "I enjoy working in the activities of my religious organization." Participants responded on a 5-point Likert-type scale (1 = not at all true of me to 5 = totally true of me). Total scores range from 10 to 50. Worthington et al., reported a full scale alpha of .95 for a religiously diverse sample of 468 undergraduate students and .88 for a sample of 150 Christian college students. The RCI obtained a Cronbach's alpha of .91 in the present analysis.

Single-Item Measures. In addition to assessing Christian identity, two other single measure self-report items were utilized as independent variables in the present study. One item provided assessment of intrinsic religiousness (as recommended by Gorsuch and McPherson, 1989) using a 6-point Likert scale (1 = strongly disagree, 6 = strongly agree). Using a 9-point Likert scale, participants were also asked to comment on their view of the Bible (1 = the ultimate source of truth, 9 = one source of truth among many).

Procedure

In order to empirically clarify the degree to which antirelgious items in the ATLG-R were responsible for any relationship between homophobia and conservative religion, hierarchical multiple regression analyses were conducted utilizing RCI, intrinsic religiousness, and participant's view of Biblical authority as dependent variables. Age, gender and the components from the ATLG-R factor solution generated in Rosik's (2007) earlier analysis served as predictor variables. By identifying the ideological salience of the items in each factor (antireligious, neutral or proreligious) and the significance of each factor for explaining the dependent variables, some light can be shed on the degree to which the relationship between homophobia and conservative religion might be an artifact of antireligious bias in the measurement of homonegativity. All analyses were again performed using SPSS 13.0 for Windows.

RESULTS

Multiple Regression Analysis

Building upon the factor analytic and ideological surround examinations of the ATLG-R, the hypotheses were tested by performing three hierarchical multiple regression analyses. A three stage model was utilized for each of these regressions. First, age and gender were entered together as control variables. Then ATLG-R components 2, 3 and 4 were entered as a group. Due to the antireligious ideology of all the items comprising component 1, this component was then entered last in order to determine the degree to which it would attenuate any associations of the other components with the dependent variables.

The means, standard deviations and zero-order correlations of all variables are presented in Table 3. Non-normality in the form of positive skewness was a problem for Age and Biblical authority. The ATLG-R component 1 was negatively skewed and therefore scores were transposed before applying transformations. Four outliers, which were over three standard deviations above the mean, were identified and eliminated for age. Square root transformations reduced skew to acceptable levels for component 1 of the ATLG-R. Inverse transformation reduced skew to acceptable levels for Biblical authority. Age retained only slight skewness after being inversely transformed. Examination of residuals indicated that assumptions for linearity, equal variances and independence of the error terms were sufficiently satisfied to proceed with the multiple regression analyses. Alpha levels were set at p < .01 to control for Type 1 error for the analysis.

As seen in Table 4, neither age nor gender were associated with RCI scores at any step of model development, as was also the case when intrinsic religiousness and beliefs about Biblical authority were being predicted. A significant [DELTA][R.sup.2] was present in the second step of the regression, with the majority of that variance apparently connected to component 2, which was significantly associated with the RCI scores. However, this correlation was completely eliminated when component 1 was added to the equation in step 3. Overall, the full model was significant (F(6,144) = 12.83, p < .001) and explained 35% of the variance in RCI scores. It thus appears that the items of components 2, 3 and 4 of the ATLG-R provided negligible illumination of respondents' religious behavior once the items of component 1 were included.

A similar pattern was found for the models predicting intrinsic religiousness and beliefs about Biblical authority, as displayed in Tables 5 and 6. With significant [DELTA][R.sup.2] for steps 2 and 3, an intrinsic orientation was significantly related to component 2 in the second step, but reduced to nonsignificance with the addition of component 1 in the third step. The full model was again significant (F(6, 144) = 12.70, p < .001) and accounted for 35% of the variance in ATLG-R scores. Of some additional interest is that the negative association of component 3 (privacy rights) with RCI neared significance (p < .02, two-tailed) in the final model.

The model predicting respondents' view of Biblical authority also revealed a strong negative correlation between the dependent variable and component 2 for the second step, which not surprisingly registered a significant [DELTA][R.sup.2]. Yet once again this relationship was removed (i.e., achieved a nonsignificant p < .05, two tailed) once component 1 was entered into the equation in step three. Adding this component resulted in a significant [DELTA][R.sup.2], as those having a higher view of the Bible's authority had significantly higher scores on component 1 in the full model. Again, the full model was significant (F(6, 144) = 13.01, p < .001) and accounted for 35% of the variance in the students' beliefs about the authority of the Bible.

DISCUSSION

The present study raises a number of intriguing and potentially important considerations that have heretofore been given minimal attention in the literature on the relationship between homophobia and conservative religion. After summarizing the findings, some further implications will be highlighted for understanding how homophobia/homonegativity may function as a construct when applied to religious conservatives.

Hypothesis 1 was generally supported. Religious commitment, intrinsic religiosity and belief in the authority of the Bible were all significantly predicted by the "Condemnation-Tolerance" component of the ATLG-R, even after accounting for variance attributable to participants' gender, age, and responses to the other three ATLG-R components. The "Condemnation-Tolerance" component, it should be remembered, was comprised only of items that conservatively religious students experienced as anti-religious and for the most part assessed respondents' views on the morality or naturalness of same sex behavior.

Similarly, hypothesis 2 was also supported. None of the religious dependent variables were significantly predicted by the items comprising the other components of the ATLG-R once the "Condemnation-Tolerance" component was entered into the model. Components 2-4 were comprised of a mix of ideologically neutral, proreligious and antireligious items that dealt with affective responses, social concerns and social policy toward lesbians and gay men. These items might be broadly considered by religious conservatives as regulating their relationships with homosexual persons, both on a personal and societal basis.

Implications for Interpreting the Homophobia-Religion Connection

Given these findings, mental health professionals studying the literature on homophobia and religion need to exercise great caution in interpreting the conclusions made by researchers in this area. Had the present examination of attitudes toward gay men and lesbians among conservatively religious persons taken the approach common to most studies, only the overall ATLG-R scale mean of 119, perhaps separated for men (M = 128.41) and women (M = 115.52), and the highly significant correlations between the RCI, intrinsic religiosity and view of Biblical authority (r = .54, r = .52, r = -.54, respectively, all p < .001) and ATLG-R scale would have been presented. Given the means Herek (1994) found for his student sample at a secular university (M = 97.79 for men, M = 95.31 for women), such large differences might lead to a global conclusion that devout Christian students displayed a significant degree of homophobia and anti-gay prejudice. Such an inference could in actuality be erroneously one-dimensional, given respondents' modest affirmation of basic civil rights and their ideological proreligious perception of the items rejecting employment discrimination and negative affective responses to lesbians and gay men. Failure to highlight these kinds of considerations might also contribute to the perception of religiously conservative clients in an unnecessarily derogatory or stereotyped manner.

The present analysis certainly suggests that homophobia is an ideologically embedded construct whose normative values about same sex behavior can be at variance with those who adhere to a conservatively religious ideology. There is a great deal of evidence for this from this series of studies, so much so that it is probably appropriate to consider homophobia and homonegativity as defined in the ATLG-R as terms that fundamentally denote a moral outlook as opposed to a phobic reaction when studying religious conservatives. In addition to the previous factor analytic and ideological surround analyses (Rosik, 2007), the present regression analyses indicated that Herek's "Condemnation-Tolerance" component was the only significant component predictor of RCI, intrinsic religiosity and beliefs concerning Biblical authority. Given the apparent potential for ideological circularity in the ATLG-R, the positive relationships between homophobia and measures of religious behavior/belief found in the present study may tell us little about "homophobia" per se and merely imply that students who hold religiously based views about same sex behavior are students who hold religiously based views about same sex behavior (Watson, 1993).

In summary, then, the conceptualization of homophobia and homonegativity as operationalized by the ATLG-R scale may need to be refined when applied to religious conservatives in order to account for underlying ideological presumptions, as indicated by the perceived antireligious nature of the majority of items. At least in the present sample of devout Christian students, the relationship between the religious variables and the ATLG-R components was significant and only significant for the antireligious items that comprised the "Condemnation-Tolerance" component, even after accounting for age, gender and the other scale components. Thus, homophobia as defined by the ATLG-R appeared largely explicable in terms of traditional religious moral sentiment rather than a pathological adjustment that inevitably leads to negative affects toward or the suppression of certain basic civil rights for gay men and lesbians. To the extent that other scales of homophobia and internalized homophobia have similar factor structures and ideological surrounds fundamentally experienced by religious conservatives as antireligious, they may share in this need for interpretive sensitivity.

In addition to the limitations earlier outlined by Rosik (2007), a reminder needs to be made regarding the present analyses that correlational statistics do not allow for definite statements on the causal directionality of the relationship between homonegative attitudes and conservative religious belief and behavior. While it is theoretically conceivable that individuals with preexisting homonegative sentiment are attracted to conservative moral religious teachings, it seems more plausible that religious teaching cultivates disapproval of homosexual behavior, especially among individuals who are raised within these faith communities.

Summary Conclusions

This study and its predecessor (Rosik, 2007) have illustrated the potential influence of divergent ideological surrounds when examining the relationship of homophobia and conservative religion. Item-level, factor analytic, ideological and multiple regression analyses conducted on devout Christian students suggests that the ATLG-R may evidence ideological partiality on a number of levels. It appears that this instrument 1) is constructed with a majority of items that are perceived to be antireligious, i.e., that negatively reference the normative value assumptions of conservative religion as regards same sex behavior, 2) relies heavily on these items to define the relationship between religious behaviors/beliefs and homophobia, as evident in the antireligious nature of all items contained in the "Condemnation-Tolerance" component, and 3) in so doing conceals the likelihood that some respondents exhibit a degree of tolerance along with strong moral disapproval regarding homosexual behavior (Rosik, Griffith, & Cruz, 2007). In addition, the findings raise some question as to whether previous differences found in heterosexuals' attitudes toward lesbians and gay men reflect greater homonegativity directed at gay men or simply differences in the way religious and cultural conservatives ideologically experience the ATL and ATG items.

There are certainly elements within religious conservatism that have treated the difficult circumstances faced by lesbians and gay men without proper compassion and in doing so brought disgrace to their faith. However, it is not at all clear that these excesses should best be remedied by engraving in conceptual stone an ideological position that deprecates by definition traditional religious beliefs about sexual morality and the natural order. Such issues deserve to be discussed openly and on their own terms within the field of psychology, thereby promoting rather than inhibiting the expression of diverse perspectives. In the search for workable solutions to the clinical and social dilemmas over homosexuality that respect the beliefs of both religious conservatives and gay men and lesbians, the discipline of psychology will need to exercise heightened circumspection and restraint in the use of the language of homophobia as presently defined.

REFERENCES

Bassett, R. L., Hodak, E., Allen, J., Bartos, D., Grastorf, J., Sittig, L., & Strong, J. (2000). Homonegative Christians: Loving the sinner but hating the sin. Journal of Psychology and Theology, 19, 258-269.

Bassett, R. L., Angelov, A. B., Mack, W. J. A., Monfort, K., Monroe, J., & Rosik, C. H. (2003). Spontaneous and deliberative attitudes toward gay and lesbian persons among Christian college students. Paper submitted for publication.

Bassett, R. L., Kirnan, R., Hill, M., & Schultz, A. (2004). SOAP: Validating the Sexual Orientation and Practices Scale. Paper submitted for publication.

Bassett, R. L., van Nikkelen-Kuyper, M., Johnson, D., Miller, A., Carter, A., & Grimm, J. P. (2005). Being a good neighbor: Can students come to value homosexual persons? Journal of Psychology and Theology, 33, 17-26.

Finlay, B., & Walther, C. S. (2003). The relation of religious affiliation, service attendance, and other factors to homophobic attitudes among university students. Review of Religious Research, 44, 370-393.

Fulton, A. S., Gorsuch, R. L., & Maynard, E. A. (1999). Religious orientation, antihomosexual sentiment, and fundamentalism among Christians. Journal for the Scientific Study of Religion, 38, 14-22.

Gorsuch, R. L., & McPherson, S. E. (1989). Intrinsic/extrinsic measurement: I/E-revised and single-item scales. Journal for the Scientific Study of Religion, 28, 348-354.

Herek, G. M. (1987). Religious orientation and prejudice: A comparison of racial and sexual attitudes. Personality and Social Psychology Bulletin, 13, 34-44.

Herek, G. M. (1994). Assessing heterosexuals' attitudes toward lesbians and gay men. In Greene, B., & Herek, G. M. (Eds.). Lesbian and gay psychology: Theory, research and clinical applications (pp. 206-228). Thousand Oaks, CA: Sage Publications.

Herek, G. M. (1998). Attitudes Toward Lesbians and Gay Men Scale. In Davis, CM. (Ed.), Handbook of Sexuality-Related Measures (pp. 392-394). Thousand Oaks, CA: Sage Publications.

Johnson, M. E., Brems, C., & Alford-Keating, P. (1997). Personality correlates of homophobia. Journal of homosexuality, 34,57-69.

Rosik, C. H. (2007). Ideological concern in the operationalization of homophobia, Part I: An analysis of Herek's ATLG-R scale. Journal of Psychology and Theology, 35, 132-144.

Rosik, C. H., Griffith, L. K., & Cruz, Z. (2007). Homophobia and conservative religion: Toward a more nuanced understanding. American Journal of Orthopsychiatry, 77,10-19.

Watson, P. J. (1993). Apologetics and ethnocentrism: Psychology and religion within an ideological surround. International Journal for the Psychology of Religion, 3, 1-20.

Watson, P. J., Morris, R. J., & Hood, R. W. (1992). Quest and identity within a religious ideological surround. Journal of Psychology and Theology, 20,376-388.

Wilkinson, W. W. (2004). Religiosity, authoritarianism, and homophobia: A multidimensional approach. International Journal for the Psychology of Religion, 14, 55-67.

Wilkinson, W. W., & Roys, A. C. (2005). The components of sexual orientation, religiosity, and heterosexuals' impressions of gay men and lesbians. Journal of Social Psychology, 145, 65-83.

Worthington, E. L., Jr., Wade, N. G., Hight, T., L., Ripley, J. S., McCullough, M. E., Berry, J. W., et al. (2003). The Religious Commitment Inventory-10: Development, refinement, and validation of a brief scale for research and counseling. Journal of Counseling Psychology, 50, 84-96.

Worthington, E. L., Jr. (1988). Understanding the values of religious clients: A model and its application to counseling. Journal of Counseling Psychology, 35,166-174.

AUTHOR

ROSIK, CHRISTOPHER H. Address: Link Care Center, 1734 W. Shaw Ave., Fresno, CA 93711. Title: Psychologist. Degree: PhD. Specializations: Missions, dissociative disorders, human sexuality.

CHRISTOPHER H. ROSIK

Link Care Center

Fresno Pacific University

The author would like to thank several anonymous reviewers who provided invaluable guidance for improving earlier versions of these articles. Correspondence concerning this article may be addressed to: Christopher H. Rosik, Ph.D., Link Care Center, Email: christopherrosik@linkcare.org

Table 1 Descriptive Statistics and Bivariate Correlation Matrix for
ATLG-R Components and Continuous Predictor Variables

 Predictor Variables
 ATLG-R ATLG-R
 Mean SD Age Component 1 Component 2

Predictor Variables
 Age 20.68 4.59 -- .05 -.20*
 Component 1 75.84 22.49 -- .72**
 Component 2 28.51 13.43 --
 Component 3 5.03 2.79
 Component 4 4.32 2.58
Dependent Variables
 RCI 34.13 9.23
 Intrinsic 4.35 1.28
 Biblical Authority 2.33 2.13

 ATLG-R ATLG-R Biblical
 Component 3 Component 4 RCI I Authority

Predictor Variables
 Age -.11 .03 -.02 .08 -.11
 Component 1 -.08 .38** .57** .56** -.50**
 Component 2 .02 .36** .43** .47** -.41**
 Component 3 -- .02 -.08 .10 .21
 Component 4 -- .24** .20* -.07
Dependent Variables
 RCI -- .74** -.43**
 Intrinsic -- -.43**
 Biblical Authority --

Note: RCI = Religious Commitment Inventory; I = Intrinsic Religious
Orientation; Component 1 = ATLG-R primary component (Herek's
"Condemnation-Tolerance" component); Component 2 = ATLG-R second
component (negative affective reaction and ocial concerns); Component 3
= ATLG-R third component (privacy rights); Component 4 = ATLG-R fourth
component (antidiscrimination).
*p < .05, two-tailed. **p < .001, two-tailed.

Table 2 Summary of Hierarchical Regression Analysis for Variables
Predicting RCI Scores (N = 151)

Predictor B SE B [beta]

Step 1
 Gender 3.45 1.71 .17
 Age -111.62 170.89 -.05
Step 2
 Gender 2.10 1.59 .10
 Age 53.57 161.12 .03
 Component 2 .27 .06 .40*
 Component 3 -.39 .25 -.09
 Component 4 .27 .26 .09
Step 3
 Gender 1.22 1.46 .06
 Age -19.48 147.83 -.01
 Component 2 .00 .07 .00
 Component 3 -.10 .23 -.03
 Component 4 .06 .24 .02
 Component 1 2.30 .42 .57*

Note. [R.sup.2] = .17 for Step 1; [DELTA][R.sup.2] = .19 for Step 2 (p <
.001); [DELTA][R.sup.2] = .32 for Step 3 (p < .001). RCI = Religious
Commitment Inventory; Component 1 = ATLG-R primary component (Herek's
"Condemnation-Tolerance" component); Component 2 = ATLG-R second
component (negative affective reaction and social concerns)t; Component
3 = ATLG-R third component (privacy rights); Component 4 = ATLG-R fourth
component (antidiscrimination).
*p < .001, two-tailed.

Table 3 Summary of Hierarchical Regression Analysis for Variables
Predicting Intrinsic Scores (N = 151)

Predictor B SE B [beta]

Step 1
 Gender .02 .07 .02
 Age -7.41 7.52 -.07
Step 2
 Gender -.05 .07 -.06
 Age 3.35 6.95 .04
 Component 2 .01 .00 .47*
 Component 3 .02 .01 .10
 Component 4 .00 .01 .03
Step 3
 Gender -.09 .06 -.10
 Age .38 6.45 .00
 Component 2 .00 .00 .10
 Component 3 .02 .01 .16
 Component 4 .00 .01 -.03
 Component 1 .09 .02 .53*

Note. [R.sup.2] = .01 for Step 1; [DELTA][R.sup.2] = .22 for Step 2 (p <
.001); [DELTA][R.sup.2] = .12 for Step 3 (p < .001). Component 1 = ATLG-
R primary component (Herek's "Condemnation-Tolerance" component);
Component 2 = ATLG-R second component (negative affective reaction and
social concerns); Component 3 = ATLG-R third component (privacy rights);
Component 4 = ATLG-R fourth component (antidiscrimination)
*p < .001, two-tailed.

Table 4 Summary of Hierarchical Regression Analysis for Variables
Predicting Biblical Authority Scores (N = 151)

Predictor B SE B [beta]

Step 1
 Gender .00 .07 .00
 Age 9.06 6.62 .11
Step 2
 Gender .50 .06 .06
 Age .97 5.97 .01
 Component 2 .01 .00 -.51*
 Component 3 .02 .01 .12
 Component 4 .00 .01 .02
Step 3
 Gender .08 .06 .10
 Age 3.15 5.67 .04
 Component 2 -.01 .00 -.21
 Component 3 .01 .01 .08
 Component 4 .01 .01 .07
 Component 1 -.07 .02 -.42*

Note. [R.sup.2] = .01 for Step 1; [DELTA][R.sup.2] = .26 for Step 2 (p <
.001); [DELTA][R.sup.2] = .08 for Step 3 (p < .001). Component 1 = ATLG-
R primary component (Herek's "Condemnation-Tolerance" component);
Component 2 = ATLG-R second component (negative affective reaction and
social concerns); Component 3 = ATLG-R third component (privacy rights);
Component 4 = ATLG-R fourth component (antidiscrimination).
*p < .001, two-tailed.