James E. Kennedy
Published on the internet in pdf and HTML at
http://jeksite.org/research/bv.pdf and http://jeksite.org/research/bv.htm.
Copyright 1999 James E. Kennedy
ABSTRACT: Guilt and fear associated with religious beliefs have received little research attention. In a convenience sample of 100 adults, fear of punishment by God was significantly higher for fundamentalists protestants than for liberal protestants, or for those with a personal faith not associated with a religious organization. Feelings of guilt for not living up to their religious ideals were approximately the same for fundamentalist and liberal protestants. The pattern of correlations among well-being, importance of religion, and religious guilt and fear was consistent with the hypothesis that religious guilt and fear contribute to suppressing a positive relationship between subjective well-being and importance of religious faith.
Negative affect from guilt and fear associated with religious beliefs is an important topic that has received little research attention. Freud (1961/1930) and Ellis (1980), use guilt as the foundation of their arguments that religion is neurotic. However, their views appear to be based more on interpretations of certain cases in their clinical practices than on research. A recent review of research found neither a positive nor a negative relationship between neuroticism and religion (Francis, 1992). In addition, literature reviews find weak, positive correlations between religious faith and subjective well-being (McFadden, 1995; Witter, Stock, Okun, & Haring, 1985). These results do not support the views of the critics, but the results also do not support the strongly beneficial effects of religion that might be expected based on testimonials and case studies reported in the religion literature (e.g., Koenig, Smiley, & Gonzales, 1988; Maton & Rappaport, 1984).
The available evidence suggests that religion can have both beneficial and detrimental effects on well-being. The different effects of religion and the conditions under which they occur have not yet been delineated in research.
Several observers have reported anecdotal evidence that members of fundamentalist religions tend to have high levels of religious guilt and fear (e.g., Barr, 1980; Hartz & Everett, 1989; Strozier, 1994). The usual explanation is that guilt and fear of punishment result from an inability to meet the very high, perhaps unrealistically high, standards of conduct and thought set by the religion. However, as Hood, Spilka, Hunsberger, and Gorsuch (1996, p. 427) note, these anecdotal reports have not been investigated with quantitative comparisons of different religious groups. In addition, religion may possibly provide mechanisms to counteract feelings of fear and guilt.
Watson, Morris, and Hood (1989) suggested that religious grace or Divine forgiveness may counteract religious guilt. With a combined sample of 1,397 college students, they found a slightly negative association between self-guilt and depression (r=-.08, p<.01) and a slightly positive partial correlation between guilt and depression (rp=.06, p<.05) when grace was controlled. They interpreted this as evidence that religious grace counteracted the otherwise positive relation between guilt and depression.
However, reliable measures of religious fear and guilt have not yet been developed. Watson, Morris, and Hood (1988) noted that the reliabilities of their guilt scale were too low (.48 to .66) for general use and one of the four self-guilt items clearly had elements of grace as well as guilt. The scale also dealt with beliefs more than feelings. For example, a typical item was "I, like everyone else, am sinful."
The primary purpose of the present study was to develop items that reliably measure negative affect resulting from religious fear and guilt. The fact that construct validity is difficult to evaluate because factors that are related to religious fear and guilt have not yet been established is a revealing indication of the exploratory stage of research on this topic. At present, evidence for a relationship between degree of religious conservatism and religious fear and guilt may provide evidence for the validity of the measure as well as for the anecdotal observations noted above. Therefore, exploring differences in religious guilt and fear among religious groups was also a major purpose of this study.
In addition, this study provided an opportunity to explore whether religious fear and guilt contribute to the low correlation between religious faith and well-being. Religious fear and guilt were expected to be negatively correlated with well-being and positively correlated with the importance of religion for a person. This would result in a suppressor effect where controlling for religious guilt and fear should increase the association between well-being and importance of religion.
Items were drawn from standard questionnaires when possible. The religion items were adapted to be applicable for people who view themselves as being spiritual rather than religious (Zinnbauer et al, 1997). The scores for all scales were linearly adjusted so that 0 was the minimum possible score and 100 was the highest possible score. Except for cases noted below, all items had six response options that were numbers 1 through 6 with anchors of "Not at all true for me" and "Completely true for me." Items were reverse scored as appropriate.
Well-being was measured with seven items (alpha reliability .89) from the Medical Outcomes Study (Stewart & Ware, 1992). The respondent indicated how much of the time during the past month s/he had each of the following feelings: "felt calm and peaceful," "felt depressed," "felt tense or 'high strung,'" "been a happy person," "had difficulty trying to calm down," "felt downhearted and blue," "felt cheerful, lighthearted." The six response options ranged from "None of the time" to "All of the time."
Importance of religion was measured with six items (alpha reliability .85), three of which were taken from a standard intrinsic religiosity scale (Genia, 1993) and were modified to include spiritual beliefs. The items were: "My religious or spiritual beliefs are what really lie behind my whole approach to life," "I try hard to carry my religious or spiritual beliefs over into all other dealings in life," "My religious or spiritual faith is very important to me," "Many other things are more important to me than my religious or spiritual beliefs," "I normally do not think about religious or spiritual matters during my daily activities," "My religious or spiritual beliefs have little role in my everyday life."
Religious fear and guilt was measured with five items "I fear punishment by God or a higher power," "I often feel guilty for not living up to the ideals of my religious or spiritual beliefs," "I worry that I will not be worthy of God's love and redemption," "I worry that I will suffer when God or a higher power makes me pay for my mistakes," "I fear God's anger."
Religious category categorized respondents as fundamentalist protestant, Catholic, moderate protestant, liberal protestant, Jewish, personal faith, or unknown. The category was determined from the respondent's answer to the open ended question: "What is your religion or spiritual philosophy? (Include specific denomination or sect, if applicable, or you may write "personal faith" if you do not identify with a religion or spiritual group.)"
Protestant denominations were classified as liberal, moderate, or fundamentalist using the classifications developed for the General Social Survey (Smith, 1987). The GSS fundamentalist category is very similar to what some other researchers describe as conservative protestants (e.g., Roberts, 1995). The liberal protestant category appears to be relatively consistent across researchers. The personal faith category was included to explore the 35 to 45 percent of the population who believe in God but do not regularly participate in a religious organization (Davis & Smith, 1994).
Questionnaires were filled out by a convenience sample of 100 adults of diverse ages and religious affiliations. People known to the author were asked to fill out the questionnaire and to ask others to fill out the questionnaire. Fourteen respondents knew the author personally and the other 86 were contacted by others.
Of the 100 respondents, 61 were female. Eighty-seven were caucasian, six were black, three were asian or asian-american, and two were hispanic. The respondents were predominantly professional. Five had a high school education, 16 had some college or technical school, 31 had a four-year college degree, 39 had some graduate education, and 8 had a Phd. Mean age was 50 and ages ranged from 23 to 80. Fifty of the respondents lived in North Carolina, 13 in Colorado, and the others were distributed over 14 other states.
Twenty-two of the respondents were categorized as fundamentalist protestant, 2 were moderate protestants, 12 were Catholic, 30 were liberal protestants, 1 was Jewish, 15 described a personal faith, 18 either did not answer the question or gave an answer that could not be classified (e.g., "Christian" or "protestant"), and 1 did not believe in God.
Evaluation of the fear and guilt questions indicated that guilt and fear of punishment should not be considered as part of one scale. The item analysis methods of multitrait scaling described in Stewart, Hays, and Ware (1992) are useful when developing a few items to measure a relatively narrow construct. In the present data, the guilt item correlated only .27 with the mean of the four other items, whereas, a correlation of at least .30 and preferably .40 or higher is desirable for scale construction. Therefore, the single guilt item was treated separately and the four other items were used in a religious fear scale (alpha reliability .77). The variation of religious fear and guilt among the religious categories confirmed that fear and guilt were distinct factors in this sample.
The mean religious fear score for fundamentalist protestants was over two times higher than for liberal protestants and over four times higher than those reporting personal faith (see Table 1). Because of the exploratory nature of the study, the Catholic group was also included in the analysis.
The ANOVA for differences among the four groups was highly significant (F(3, 75) = 8.30, p = .0001). Post hoc comparisons are noted in the table.
On the other hand, the mean religious guilt scores were approximately the same for fundamentalist protestants and liberal protestants. As shown in Table 1, the guilt scores for those reporting personal faith was about half that of fundamentalists and liberals. The ANOVA for differences among groups was suggestive, but not significant (F(3, 74) = 2.41, p = .07). T-tests of the differences between the personal faith group and the fundamentalist and liberal groups both gave p < .03, but these differences were not significant when adjusted for multiple post-hoc analyses. The importance of religion also varied significantly among groups with fundamentalist protestants having the highest values and the personal faith category the lowest (F(3, 75) = 5.30, p = .002). There was no evidence that well-being differed among the groups.
Table 1. Mean Values by Religious Category (Standard Deviations in Parenthesis) Fundamentalist Liberal Personal Protestant Catholic Protestant Faith Unknown ____________ ________ __________ _______ ________ N 22 12 30 15 18 Religious Fear** 38.4 30.8 14.9 7.8 26.5 (24.1) (30.1) (16.9) (16.4) (27.9) [a] [b] [b] Religious Guilt 41.8 30.0 40.0 21.3 30.7 (27.5) (18.1) (25.6) (28.8) (32.8) Importance of Religion* 88.2 83.1 78.6 64.6 71.5 (15.6) (12.4) (17.7) (24.7) (25.3) [a] [b] Well-Being 74.5 75.3 70.5 75.2 72.5 (15.2) (8.9) (16.1) (16.3) (13.8) _______________________________________________________________________________ Note. The probability values are for a one-way analysis of variance for differences among the four known religious categories. Means in the same row that do not have the same letter code differ by p < .05 using the Bonferonni method to compare means. The two moderate protestants and one Jewish respondent are not included in the table or analysis of variances. Ns for each mean sometimes vary by one or two cases due to missing values. *p < .01. **p < .001.
The pattern of correlations was consistent with religious fear and guilt suppressing the relationship between well-being and importance of religion. The correlation between well-being and importance of religion increased from r = .22 (p < .05) to rp = .30 (p < .01) when religious fear and guilt were controlled. For comparison, a meta-analysis of the relationship between well-being and religious faith found the median correlation to be .17 (Witter, et al., 1985). As expected, the direction of correlations for religious fear and guilt were negative with well-being (r = -.18, N = 98, p = .08 and r = -.31, N = 96, p < .01, respectively) and positive with importance of religion (r = .14, N = 98, and r = .09, N = 96, both not statistically significant).
The finding that religious guilt and fear of punishment appeared to be distinct factors was not expected and may have important methodological and theoretical implications. Previous discussions of fundamentalism (Barr, 1980; Hartz & Everett, 1989; Strozier, 1994), and the religious guilt scale used in previous research (Watson, Morris, and Hood, 1988) assumed that guilt and fear were part of one construct. The low reliabilities with that scale may have been due in part to combining distinct factors. The four item religious fear scale used here had good reliability. Further work developing measures of religious guilt appears warranted and the present findings from the one-item guilt measure must be interpreted with caution until further research is done.
The present data supported the anecdotal observations that fear of punishment was higher for fundamentalist protestants than for liberal protestants. These results are evidence for the validity of the religious fear scale and also suggest that religious fear may be useful in distinguishing between religious fundamentalists and religious liberals. On the other hand, religious guilt was the same for fundamentalist and liberal protestants, and was suggestively lower for those reporting a personal faith not associated with a religious organization. Further research may find that religious guilt is an important factor distinguishing those who consider themselves spiritual but not religious (Zinnabauer et al, 1997).
Based on personal observations of fundamentalist religions, Strozier (1994) suggested that people who are attracted to fundamentalism tend to have high guilt, and then develop reduced guilt about past (pre-conversion) behavior and increased guilt and fear about current behavior and thought. These speculations about the causal mechanisms for the association between fundamentalism and religious fear merit investigation. The possibility that people participate in religious groups for reasons not related to fear or guilt, and that the unresolved religious fear and guilt are basically side effects for certain susceptible individuals should also be considered. That model is supported by the evidence that social factors appear to have a dominant role in many peoples' decisions about participating in religious groups (Hood, Spilka, Hunsberger, & Gorsuch, 1996; Roberts, 1995). Similarly, the possibility that religious fear and guilt may have some beneficial effects, such as helping certain people control their dispositions for self-destructive behavior merits study. For example, the negative relationships between religious participation and substance abuse are well established (Hood, Spilka, Hunsberger, & Gorsuch, 1996).
Research focusing on how people change as they become involved in religious organizations may be the best way to investigate these alternative models. A good working assumption would be that each of these alternatives, as well as others, apply to some degree.
In the present study, the pattern of correlations among well-being, importance of religion, and religious fear and guilt was consistent with the hypothesis that religious fear and guilt contribute to suppressing a positive relationship between religious faith and well-being. Further work developing measures of religious guilt may sharpen this relationship.
Several additional improvements to the methodology used here may be noted. Replacing the open ended question on religious affiliation with a forced choice question that includes the option of personal faith may reduce the missing and ambiguous responses. In addition, defining and distinguishing fundamentalism is problematic. It is likely that several dimensions of religious faith, including religious fear and guilt, must be better understood before fundamentalist religions can be properly characterized. And, of course the present results need to be confirmed in other populations.
Barr, J. (1980). The Scope and Authority of the Bible. Philidelphia: Westminister Press.
Davis, J.A., & Smith, T.W. (1994). General social surveys, 1972-1994: Cumulative codebook. Chicago: National Opinion Research Center.
Ellis, A. (1980). Psychotherapy and atheistic values: A response to A.E. Bergin's "Psychotherapy and religious values." Journal of Consulting and Clinical Psychology, 48, 642-645.
Francis, L.J. (1992). Religion, neuroticism, and psychoticism. In J. Schumaker (Ed.), Religion and mental health (pp. 149-160). New York:Oxford University Press.
Freud, S. (1961). Civilization and its discontents (J. Stachey, Trans.) New York: W.W. Norton (Original work published in 1930).
Genia, V. (1993). A psychometric evaluation of the Allport-Ross I/E scale in a religiously heterogeneous sample. Journal for the Scientific Study of Religion, 32, 284-290.
Hartz, G.W, & Everett, H.C. (1989) Fundamentalist religion and its effect on mental health. Journal of Religion and Health, 28, 207-217.
Hood, R.W., Spilka, B., Hunsberger, B., & Gorsuch, R. (1996). The psychology of religion. New York: Gilford Press.
Koenig, H.G., Smiley, M., & Gonzales, J. (1988). Religion, health, and aging. Westport, CT: Greenwood Press.
Maton, K.I., & Rappaport, J. (1984). Empowerment in a religious setting: A multivariate investigation. Prevention in human services, 3, 37-72.
McFadden, S.H. (1995). Religion and well-being in aging persons in an aging society. Journal of Social Issues, 51, 161-175.
Roberts, K.A. (1995). Religion in sociological perspective. New York: Wadsworth.
Smith, T.W. (1987). Classifying protestant denominations. GSS Methodological Report No. 13. Chicago:National Opinion Research Center [producer]. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor].
Stewart, A.L., Hays, R.D., & Ware, J.E. (1992). Methods of constructing health measures. In Stewart, A.L., & Ware, J.E., (Eds.). Measuring functioning and well-being: The medical outcomes study approach (pp.67-85). Durham, NC: Duke University Press.
Stewart, A.L., & Ware, J.E., (Eds.), 1992. Measuring functioning and well-being: The medical outcomes study approach. Durham, NC: Duke University Press.
Strozier, C.B. (1994) Apocalypse: On the psychology of fundamentalism in America. Boston: Beacon Press.
Watson, P.J., Morris, R.J., & Hood, R.W. (1988). Sin and self-functioning, Part 1: Grace, guilt, and self-consciousness. Journal of Psychology and Theology, 16, 254-269.
Watson, P.J., Morris, R.J., & Hood, R.W. (1989). Sin and self-functioning, Part 4: Depression, assertiveness, and religious commitments. Journal of Psychology and Theology, 17, 44-57.
Witter, R.A., Stock, W.A., Okun, M.A., & Haring, M.J. (1985). Religion and subjective well-being in adulthood: A quantitative synthesis. Review of Religious Research, 26, 332-342.
Zinnabauer, B.J., Pargament, K.I., Cole, B., Rye, J.S., Butter, E.M., Belavich, T.G., Hipp, K.M., Scott, A.B., & Kadar, J.L. (1997). Religion and spirituality: Unfuzzying the fuzzy. Journal for the Scientific Study of Religion, 36, 549-564.