The Attitudes Toward Lesbians and Gay Men Scale (ATLG) : Adaptation and Testing The Reliability and Validity in Chile

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 8

The Attitudes Toward Lesbians and Gay Men Scale (ATLG): Adaptation and

Testing the Reliability and Validity in Chile

Manuel C􏰁ardenas and Jaime Eduardo Barrientos Psychology School, Universidad Cat􏰁olica del
Norte

Previous researchers have used the attitudes toward lesbians and gay men (ATLG) scale (Herek, 1988) and
reported their own reliability coefficients and validity measures, but research on this subject is scarce in Chile.
In order to determine if ATLG scale was a reliable and valid instrument in our country, we adapted and
examined the psychometrics properties using a sample composed of 142 psychology and economics
undergraduate students. This study found the ATLG scale to be reliable (a=.90) and valid for the Chilean
population and is a recommended instrument for measuring attitudes toward gay men and lesbians.

Negative attitudes toward gay men and lesbians have become an important topic and research area in social
sciences in the last 20 years. Although the issue regarding gay men and lesbian rights is increasingly present
in the Chilean public agenda and the Chilean population currently seems to be more tolerant, prejudice and
stereotypes toward sexual minorities have not changed much in the last few years, and neither have the
discriminating actions deriving from them. It must be remembered that article 365 of the Penalty Code that
penalized sexual relations between men was repealed in 1999 during the democratic government post military
dictatorship. The unequal treatment toward gay men and lesbians and lack of social protection for homosexual
persist in Chilean society when, for example, trying to get a job and using public utility services (MUMS,
2006).

On the other hand, studies about population attitudes toward gay men and lesbians in Chile are scarce and
only recently have been studied via very general public opinion surveys in which questions about
homosexuality are confined to two or three items or scales, for which reliability and validity indicators were
not reported. Therefore, they can be used as illustrations only with due precautions. Data show that there is a
particularly strong social intolerance and homophobia regarding homo and bisexuality in Chile. For example,
data from a questionnaire administered by Fundacion IDEAS in 1997 showed that there is a value charge of
intolerance and discrimination considered risky on different issues, particularly on homosexuality, which
were considered ‘‘very serious’’, with a figure of 60.2% (100% was the maximum, showing very strong
prejudice). The second report of this foundation showed that 45.2% of the subjects think that homosexuality
should be forbidden since it is against human nature (Fundacio􏰁n IDEAS, 2001). Most recent results reported
by IDEAS show that 50% or more of the population could be regarded as homophobic. IDEAS (2001) defines
homophobia as a form of personal or institutional prejudice toward homosexuals that manifests itself as fear
or rejection toward physical closeness with them.

Results also indicated that people believed homosexuality should be an important issue in medical research in
order to avoid the birth of homosexuals in the future. Moreover, in their last report (Fundacion IDEAS, 2003),
43% of the sample thought that homosexuals should not become school teachers. Other data from a Chilean
sexual behavior survey regarding intolerance levels toward homoand bisexuality showed that, concerning
sexual relations between people of the same sex, there are very restrictive judgments, even stricter than for
other sexual practices (CONASIDA & ANRS, 2000). Thus, data show that only 5% of the Chilean population
approve of men or women who have sexual relations with people of the same sex. Most of the population
reports disapproval of homosexuality, both male and female.

Previous studies have examined other factors associated with attitudes toward homosexuality. For example,
studies report that negative attitudes toward homosexuality closely correlate with authoritarianism (Dumbar,
Brown, & Amoroso, 1973) and political conservatism (Herek, 1988; Hayes, 1995; Steffens & Wagner, 2004).
Also, a close relationship is observed between negative attitudes toward homosexuality and traditional
attitudes toward gender or sex roles (Herek, 1988; Kite & Deaux, 1986; Morrison & Morrison, 2002). If the
subject violating the sex role is a man, transgression is viewed as more serious (Herek, 1984; La Mar & Kite,
1998). Prejudice against gay men and lesbians is related to variables like religiosity (Gentry, 1987; Herek,
1988; Morrison & Morrison, 2002) and attitudes toward feminism (Kite & Deaux, 1986). It also is connected
with variables like age: the older the subject, the more negative the attitude; educational level: the higher the
educational level, the less the prejudice (Herek, 1988; 1994); and sex: men are more prejudiced than women
(Herek, 2000; La Mar & Kite, 1998; Lingiardi, Falanga, & Augelli, 2005).

In a different sense, when a heterosexual population thinks that homosexuality is a choice by subjects
themselves, rejection figures tend to be higher (King, 2001; Sakalli, 2002).

Studies also support the idea that people with homosexual relatives, or who have established close
relationships with homosexuals, hold more favorable attitudes toward them (Bowen & Bourgeois, 2001;
Hinrichs & Rosenberg, 2002; Steffens & Wagner, 2004), and those without these contacts hold more
prejudiced attitudes (Lingiardi, Falanga, & Augelli, 2005). Among the latter, men feel more rejection and put
more social distance from gay men than from lesbians (King & Black, 1999; Kite & Whitley, 1996; La Mar &
Kite, 1998), whereas women report feeling more at ease with gay men than with lesbians (Herek, 1994).
Whatever the case, there is a positive correlation between attitudes toward gay men and lesbians (Herek &
Capitanio, 1996), and both are evaluated better than bisexuals of their same sex (Steffens & Wagner, 2004).
Finally, even in cases in which prejudice is low and attitudes toward homosexuals are rather favorable,
subjects do not translate this tendency into a defense of homosexual rights (Ellis & Kitsinger, 2003).

There is no thorough and useful information to learn about attitudes toward gay men and lesbians and their
connection to other relevant psychosocial and social variables due to the unavailability of Chilean scales to
measure attitudes toward homosexuality consistently and validly.

Therefore, this article is aimed at reporting the phases of the process for adapting and validating the ATLG
scale (Herek, 1988). This scale has been translated and validated for samples from various populations and
countries, but no studies have been done for its adaptation to the Chilean context. Only recently a study with a
Spanish version was done among U.S. residents of Mexican origin (Herek & Gonzalez-Rivera, 2006), but
using a 4-item version of the attitudes toward gays (ATG) and attitudes toward lesbians (ATL) scales.

So, the objective of this study was to determine if the ATLG scale is a suitable device to measure prejudice
toward gay men and lesbians in our particular social context and to examine the association between those
attitudes and theoretically relevant sociodemographic, psychological, and social variables. The following
hypotheses were tested:

H1: Chilean men hold more negative attitudes toward homosexuality than do Chilean woman. This difference is greater
for attitudes toward gay men than toward lesbians.

H2: Chilean people from various socioeconomic levels express different attitudes toward gay men and lesbians. As this
hypothesis is exploratory, the significance of the above-mentioned differences is not given.

H3: Highly religious Chilean population holds more negative attitudes toward homosexuality.
H4: Chilean people politically identified with the Right Wing will have more negative attitudes toward homosexuality.

H5: Chilean people express less negative attitudes toward homosexuality if they have had personal contact with gay men
and lesbians.
H6: The Chilean people express more positive emotions toward homosexuals when their prejudice levels are lower.
Individuals who hold more positive attitudes toward homosexuals express significantly less negative emotions toward
homosexuals.

H7: Chilean people belonging to an ethnic minority express more positive attitudes toward homosexuality, for both gay
men and lesbians.
H8: Chilean people who report more negative attitudes toward homosexuals are those who believe that homosexuals have
too much power. Those that hold more positive attitudes toward homosexuals report that homosexuals have little power.

H9: People who show more positive attitudes toward homosexuals express their agreement with giving more rights to
them. Those with more unfavorable attitudes will prefer restricting or eliminating rights.
H10: Students from most liberal majors (psychology) have attitudes more positive than those that are from traditionally
more conservative majors in our country (economics).

Method

Participants and Procedures

The participants in this study were 152 volunteers recruited from university introductory undergraduate
psychology and economics courses. The present study used a convenience sample consisting of 142 subjects,
50 men (35%) and 92 women (65%) whose ages ranged from 18 to 31 years (M 1⁄4 19.92, SD 1⁄4 1.99); 10
subjects who reported themselves as homosexual (n 1⁄4 5) or bisexual (n 1⁄4 3), or who did not answer the
questions related to this issue (n 1⁄4 2) were eliminated from the analysis. All of them were students majoring
in psychology (n 1⁄4 105) and economics (n 1⁄4 37) at Universidad Catolica del Norte, Chile.

The study was presented to the subjects as part of a study on group relationships and related representations to
different social groups because when studying this kind of subject in Chile, we avoid explaining the study
topic directly to avoid negative responses. Therefore, the words ‘‘representation’’ frequently are used for this
purpose, referring to intergroup relationships. All respondents were assured anonymity. The undergraduate
students completed a battery of paper-and-pencil measures which included independent variables. The
questionnaire was completed by individual respondents. Volunteers received extra credit points for their
course grade. The ethical criteria of the Chilean’s Research Commission (CONICYT) were followed.

Measures

Social and demographic measures. The questionnaire included items to ascertain respondent sex, age,
socioeconomic level, religious denomination, political self-categorization, sexual orientation, and ethnic
identification.

Socioeconomic level. This variable was divided into three socioeconomic levels according to participant’s
reports on perception of belonging to a certain socioeconomic class or level—upper, middle, and low class.

Religion. This variable consisted of a self-report on the answer to the question, ‘‘Do you consider yourself a
religious person?’’ Then, subjects had to identify their religious group: Catholic, Evangelist, Muslim, Jeishw,
and other.

Sexual orientation. Three categories were assessed: homosexual, bisexual, and heterosexual. Homosexual and
bisexual people were not included in the analysis.

Political self-categorization. This measure asked about political preference, from right to left. The scale
consisted of six alternatives, from extreme right to extreme left, recoded in three options (right, center, and
left). Respondents were asked, ‘‘In political matters people frequently speak of left and right. Where you
would place your ideas in the following scale?’’

Ethnic minority. Respondents were asked, ‘‘Do you consider yourself as belonging to some ethnic minority?’’
Two response alternatives were provided: ‘‘Yes’’ or ‘‘No.’’

Other relevant variables.

Power. This measure consisted of a 10–100 graded scale, options being ‘‘no’’ (10) or ‘‘full power’’ (100).
Respondents were asked, ‘‘How much power (e.g., capacity to make decisions, move freely, economic power,
influences on administration, etc.) would you say that homosexuals have as compared with heterosexuals in
our context?’’
Degree of intimacy. This means the degree of intimacy with homosexuals, ranging from ‘‘no closeness’’ to
‘‘very close.’’ Besides, a measurement of pleasure with the degree of intimacy was added, from ‘‘not
agreeable’’ to ‘‘very agreeable.’’ In both cases, the response options ranged from 1 to 6.

Measures related to homosexuality.

ATLG. Students completed the ATLG scale (Herek, 1988) translated and adapted for its validation (see the
Appendix). As in other studies using this scale (Herek & Gonzalez-Rivera, 2006), and since the extent to
which the term ‘‘gay’’ is understood by nongay Chileans has not been documented, we used the term
‘‘homosexual’’ in the Spanish translation to refer to gay men. The ATLG translation used in this study was
not the same as that used by Herek and GonzalezRivera (2006); that version has only four statements for ATL
and ATG scales. The ATLG Spanish version in this study had not been used previously with Chilean
populations. The scale originally was developed in English, and their psychometrics properties are well
established (Herek, 1994). The ATLG scale consists of 20 statements, 10 about gay men (ATG subscale) and
10 about lesbians (ATL subscale). Respondents show their level of agreement or disagreement using
Likerttype items (from ‘‘strongly disagree’’ to ‘‘strongly agree’’). The original scale has only 5 points, but a
6-point scale was used to encourage subjects to make decisions. High scores, close to 6, showed greater
prejudice, and low scores reported favorable attitudes toward homosexuals.

HATH. The subjects also reported on their attitude toward lesbians and gay men using the heterosexual
attitudes toward homosexuals (HATH) scale Larsen, Reed, & Hoffman, 1980). The HATH scale consisted of
20 Likert-type items interpreted in the same fashion. The scale used in this questionnaire consisted of 5-point
Likert-type response categories ranging from strongly agree to strongly disagree. The HATH scale was
translated into Spanish and adapted to the Chilean population (Barrientos & C􏰁ardenas, 2007). In the adapted
scale, the reliability levels obtained by means of the Cronbach alpha were .90, and all the items showed an
item–total correlation higher than .25. This reliability index reveals high internal consistency that., is even
slightly higher than that resulting from original studies (Larsen, Reed, & Hoffman, 1980) in which it reached
a reliability index of .86.

Homosexual rights. Respondents were asked, ‘‘I think that homosexual rights should be . . .’’ options

being multiple-choice: extended, restricted, supported, or eliminated.

Emotions toward homosexuals. Respondents were asked, ‘‘Has a homosexual ever made you feel any of these
emotions?’’ Emotions follow: frustration; hope, inspiration, hate, attraction, displeasure, fear, envy,
congeniality, inconvenience, repugnance, sorrow, comprehension, pride (for what that individual has done),
and respect. Responses were coded ‘‘Yes’’ (1) or ‘‘No’’ (0). Participants also were asked, ‘‘As a whole, how
would you say your feelings and emotions are toward homosexuals?’’

Results Sample Characteristics

The descriptive statistics for the different sociodemogrphic variables and their corresponding mean scores in
the ATL, ATG, and HATH scales can be observed in Table 1.

Reliability

The reliability levels obtained for the overall ATGL scale were .90, the scale measuring attitudes toward
lesbians (ATL) was .81, and that toward gay men (ATG) .84. In the original study, overall ATLG a 1⁄4 .90,
ATG a 1⁄4 .89, and ATL a 1⁄4 .77 (Herek, 1988). The item–total correlation of all the items was higher
than .25, so indexes shown can be considered remarkable and would reveal high internal test consistency.

Factor Analysis
Factor structure analysis of the ATLG and subscales was attempted and is shown in Tables 2, 3, and 4. For the
ATLG scale, an extraction procedure was used for main components with VARIMAX command; however,
the best explanation for the existence of the predicted factors occurred with a rotate model. Sample adequacy
measures indicated good data agreement for the factor analysis of the full scale (Kaiser–Meyer–Olkin
measures 1⁄4 .88 and Barttlet sphericity test, p 1⁄4 .000) and both the ATL scale (Kaiser–Meyer–Olkin1⁄4.80
and Barttlet sphericity test p 1⁄4 .000) and the ATG scale (Kaiser–Meyer–Olkin 1⁄4 .86 and Barttlet sphericity
test, p 1⁄4 .000).

The ATLG scale factor structure illustrated the presence of five factors explaining 62.50% of the total
variance, but previous studies showed two factors: ‘‘condemnation–tolerance’’ and ‘‘beliefs.’’ The first factor
in previous research accounts for 35%–45% of the total common variance in responses, which are similar for
male and female respondents and for questionnaires concerning lesbians and gay men. The second factor
accounts for another 5% of the total variance (Herek, 1984).

In this study, three factors found referred to the ATL subscale (Table 3) and two of them to the ATG subscale
(Table 4). Their relationship with these scales will be commented upon because they seem to be related to
similar components but directed to the case of either lesbians or gay men. Therefore, the three factors of the
ATL subscale would explain 61.40% of the variance. Factor 1 would explain 30.68% of the variance with
items 3, 5, 6, 8, 9, and 10, which could refer to ‘‘traditional values.’’ Factor 2 would explain 15.45% of the
variance and includes items 4 and 7, which would deal with ‘‘social sanction.’’ Finally, factor 3 would
explain 15.26% of the variance and includes items 1 and 2, which would refer to ‘‘social rights.’’ The third
factor is found in other studies (La Mar & Kite, 1998) in which it is not regarded as a separate factor.

For the ATG subscale, the factor structure shows two main factors that would explain 54.29% of the variance.
The first would explain 32.61% of the variance and includes items 1, 2, 3, 4, 6, 9, and 10. These refer to the
condensation of those items dealing with gay mens rights and the stereotypes that relate gay men to
perversion, which were labeled ‘‘beliefs.’’ The first ATG factor is similar to the ‘‘morality’’ dimension in
other papers (La Mar & Kite, 1998). Factor 2 would explain 21.68% of the variance and includes items 5, 7,
and 8. These would refer to the evaluation of gay men’s ‘‘natural=antinatural’’ dimension.

Validity Indicators

ATLG validity was tested using a series of Student t tests on results. Differences regarding prejudice levels
toward homosexuals between men (n 1⁄4 50, M 1⁄4 2.28, SD1⁄41.02) and women (n1⁄492, M1⁄41.91,
SD1⁄4.71) were confirmed. Women reported significantly lower prejudice levels than men [t(141) 1⁄4 􏰀2.07; p
< 0.05], suggesting more positive attitudes toward gay men and lesbians. On the other hand, when comparing
men’s scores and women’s scores on the ATL subscale (M 1⁄4 1.88 and M 1⁄4 2.12, respectively) and the
ATG subscale (M 1⁄4 2.59 and M 1⁄4 3.08, respectively), we verify the above-mentioned differences. Both
men (t(49) 1⁄4 11.14; p < .05) and women (t(91) 1⁄4 12.07; p < .05) express significantly higher levels of
prejudice toward gay men than toward lesbians.

In analyzing subscale data, ATG differences confirmed the differences found above: Women (M 1⁄4 2.24, SD
1⁄4 .85) showed attitudes significantly more positive than men (M 1⁄4 2.59, SD 1⁄4 1.02) [t(140) 1⁄4 􏰀2.56; p <
0.05]. For the ATL scale these differences did not appear [t(140) 1⁄4 􏰀1.41; ns], meaning that there are no
attitudinal differences between men (M 1⁄4 2.11, SD 1⁄4 .99) and women (M 1⁄4 1.88, SD 1⁄4 .85) toward
lesbians. This would confirm the fact that men hold more negative attitudes toward gay men and more
favorable views of lesbians (Kite & Whitley, 1996; King & Black, 1999; La Mar & Kite, 1998). These sex
differences are equally demonstrated by the number of negative emotions reported [t(139)1⁄4􏰀2.90; p<0.005],
women expressing fewer negative feelings.

No significant differences were found among subjects from different socioeconomic levels (neither in the full
scale, nor in the subscales).
Differences were detected between subjects who regarded themselves as religious and those who did not.
Those who reported religious beliefs showed attitudes significantly more negative than those who did not,
both for the full scale [t(140) 1⁄4 􏰀2.02; p < 0.05] and ATG [t(140) 1⁄4 􏰀3.51; p < 0.001] and ATL [t(140) 1⁄4
􏰀2.72; p < 0.005] subscales.

The test for political self-categorization effects suggested significant differences for ATLG scale scores [F (2,
135) 1⁄4 4.70; p < 0.05] among subjects in the left categories (n 1⁄4 31, M 1⁄4 1.92, SD 1⁄4 .66), as compared
with those in the central (n 1⁄4 86, M 1⁄4 2.52, SD 1⁄4 .97) and right (n1⁄421, M1⁄42.24, SD1⁄4.90) ones. ATL
[F (2, 135) 1⁄4 6.21; p < 0.05] and ATG [F (2, 135) 1⁄4 6.67; p < 0.05] subscale scores were also significant,
subjects in the central and right categories being more prejudiced in both cases.

Differences also were detected in the measure of pleasure that contact with gay men and lesbians had
produced in the subjects, participants with negative attitudes being those who had felt the least pleasure [t(123)
1⁄4 7.35; p < 0.001]. Differences also were found in general feelings and emotions toward gay men and
lesbians [t(123) 1⁄4 5.05; p < 0.001].

Another set of variables showed significant differences between subjects highly (n 1⁄4 53; 37%) and slightly
(n 1⁄4 88; 62%) prejudiced toward gay men and lesbians divided by the theoretical middle point of the scale.
Highly negative subjects reported a significantly greater number of negative emotions [t(138) 1⁄4􏰀2.98; p <
0.05] and a significantly smaller number of positive emotions [t(138) 1⁄4 2.71; p < 0.005] toward gay men and
lesbians.

No significant differences were found in the attitude toward gay men and lesbians of subjects regarding
themselves as belonging to a minority ethnic group and those who did not identify themselves with a minority
ethnic group (neither in the full scale, nor in the subscales).

No differences were found between subjects scoring high and low in the ATLG scale as to their perception of
homosexuals’ relative power. Both groups think that gay men and lesbian people lack social power in our
environment.

Regarding rights, differences were found among subjects highly and slightly prejudiced toward homosexuals.
Participants with positive attitudes express agreement with the extension of homosexual rights (n 1⁄4 81;
92%). Subjects with more negative attitudes show more unfavorable attitudes and prefer restricting rights or
keeping them as they are (n 1⁄4 24; 46%). This reveals that negative attitudes toward gay men and lesbians
were more strongly associated with restricting homosexual rights.

To determine if attitudes toward homosexuals are related to students’ majors, psychology students (n 1⁄4 105,
M 1⁄4 2.19, SD 1⁄4 .80) were compared with economics students (n 1⁄4 37, M 1⁄4 2.85, SD 1⁄4 1.08). A
significant effect was found for the full scale [t(140) 1⁄4 􏰀3.33; p 1⁄4 0.002], ATL scale [t(140) 1⁄4 􏰀4.18; p
1⁄4 0.000], and ATG scale [t(140) 1⁄4 􏰀2.35; p < 0.05]. Psychology students report more positive attitudes
toward homosexuals than did economics students.

The correlation between ATGL and heterosexual attitudes toward homosexuals (HATH) scale Larsen, Reed,
& Hoffman, 1980) was expected to show a high correlation because they are different measures of a similar
construct (convergence validity). ATLG (M 1⁄4 2.36; SD 1⁄4 0.92) and HATH (M 1⁄4 2.04; SD 1⁄4 0.85)
scores were correlated significantly in the expected directions (r(150) 1⁄4 87; p 1⁄4 .000). The correlations
between ATG and ATL with HATH also were the expected ones (r(150) 1⁄4 .82; p < 0.01 and r(150) 1⁄4 .83;
p < 0.01, respectively). Finally, a correlation among ATL and ATG subscales was found to be positive and
high (r(150) 1⁄4 .82; p < 0.01). These correlations would illustrate that both scales do measure the same
construct (prejudice toward gay men and lesbians). Therefore, subjects who score high in prejudice in one
scale also do so in the other scale.

Discussion
This article reports on the reliability and validation process of the ATLG full and subscales with the Chilean
population. The scales were found to have high internal consistency, and the subscales were well correlated.
The full and subscales were also highly correlated with the HATH scale, showing that these devices are
measuring the same phenomenon. Finally, the ATLG scale validity indicators were correlated to attitudes
toward gay men and lesbians.

The original scale provides only information on the factor structure for four separate samples of
undergraduates using a common factor model with oblique rotation. Herek (1984) found a bipolar
condemnation= tolerance factor and a beliefs factor and, in another study, La Mar and Kite (1998) found four
factors: condemnation=tolerance, morality, contacts, and stereotypes. In our study, the ATLG scale factor
structure shows five factors, two of them belonging to ATG and three of them belonging to ATL. The two
relevant ATG factors refer to ‘‘beliefs’’ and ‘‘natural=antinatural atural sexual option.’’ The ATL factors
refer to the threat to ‘‘traditional values,’’ ‘‘social sanction,’’ and ‘‘lesbian rights.’’ The factor ‘‘traditional
values’’ is similar to the factor proposed as a measure of new forms of prejudice by Pettigrew and Merteens
(1995) in their scales of subtle prejudice. The factor ‘‘social sanction’’ could be equivalent to the
condemnation=tolerance factor proposed by Herek (1984). The novel factor is the third one, ‘‘lesbian rights,’’
that is not included by Herek (1984) or by La Mar and Kite (1998), as expected for the latter. This is a
distinctive factor that should be further studied. The fourth ATG factor could be similar to that named
‘‘beliefs’’ by La Mar and Kite (1998). Finally, the fifth factor, ‘‘natural=antinatural sexual option,’’ is a new
factor that is not found in previous studies (but that is addressed directly by Herek when asking some
questions). Therefore, three of the ATLG scale factors belong in the ATL subscale and two in the ATG
subscale. In futures studies, each subscale eventually could include questions for the five factors found in this
research.

As expected, the attitudes toward gay men and lesbians are affected by respondents’ sex. Women are more
favorable and not more tolerant toward homosexuals than men, and men are less tolerant than women,
particularly when they evaluated gay men. Women also are more unfavorable when they evaluated gay men;
this last data are not consistent with the literature. In fact, the ATL subscale was the only one where no
significant sex differences were found. This is probably due to the fact that males think that the relationship
between women is an erotic fantasy frequently observed in pornographic films (Louderback & Whitley,
1997). These data are consistent with the literature and demonstrate that negative male attitudes are especially
high for gay men. But, under no circumstances can a statement be made to pose that lesbians are accepted by
heterosexual men; they simply are less rejected than gay men. Results suggest that it is important to measure
prejudice toward gay men and lesbians separately, because the literature shows different prejudice patterns for
each group, which is the same as in this study.

Moreover, in future studies, sex differences in attitudes toward gay men and lesbians should be understood by
considering cultural constructions of gender and sexuality in Chile. Although in the Western world and in
Chile an important transformation is taking place in both sexual and gender values (Inglehart & Baker, 2000),
previous studies in Latin America and Chile have shown the great importance of social context and social
factors in sexuality, gender, and masculinity (CONASIDA & ANRS, 2000; Viveros, 2001). Future studies
should explore this subject and include some measures of it.

Our results reveal that people with more unfavorable attitudes toward gay men and lesbians are also the most
religious (Gentry, 1987; Herek, 1988; Morrison & Morrison, 2002). The growing secularization in Chile is
related to the spread of social and sexual changes, but this state of affairs has generated negative reactions
among those who cherish traditional values. Values about individual rights and citizen equality produced
tension in Chile with family and gender models sustained by the Catholic Church, patriarchal traditions, and
political conservatism.

The study also suggests that when subjects meet homosexuals and evaluate mutual contact positively, they are
less prejudiced, and subjects who express more negative emotions toward sexual minorities are also those
who express fewer positive emotions toward them, thus being consistent with the literature (Bowen &
Bourgeois, 2001; Hinrichs & Rosenberg, 2002; Steffens & Wagner, 2004) Other results confirm that
belonging to an ethnic minority does not seem to be related to attitudes toward gay men and lesbians, and it
does not seem to be connected with a particular socioeconomic group. A though in this study the variable
‘‘socioeconomic status’’ is not correlated with attitudes toward gay men and lesbians as reported in other
studies (Sandfort, 1998), its association with attitudes toward gay men and lesbians could be due to other
more psychological relations they are confused with, like general political conservativeness (Steffens, 2005).

Finally, the response patterens found in this study tend to confirm the prejudice that men, economics students,
religious people (mainly Catholics), and right-wingers feel toward gay men and lesbians. These attitudinal
characteristics regarding antigay men and lesbians could be based on a more global conservative perspective.
So, negative attitudes toward gay men and lesbians would play a defensive role for conservative values and
ideas, since the ATLG scale allows describing sociodemographic homophobic patterns in Chile more
appropriately. Future studies should include this subject.

Later studies will make it necessary to focus on other limitations, that is, the existence of some forms of
prejudice that can remain unnoticed because they are limited by social desirability. As reported in previous
studies (Sandfort, 1998) in an intolerant environment toward homosexuality, the expression of a socially
penalized behavior such as homoand bisexuality may be symbolically equivalent to a confession.

There are other variables that may influence attitudes toward homosexuals that must be controlled. These
include prejudice linked to personality traits and defense mechanisms and subjects’ adhesion to gender
stereotypes and sex roles, along with other variables that have not been measured directly, such as
conservatism and educational level. It also would be relevant to analyze how ways of expressing this behavior
have changed since hostile emotions traditionally expressed may have turned into other expressions such as
fear, discomfort, insecurity, or disgust, which are proper of new forms of prejudice (Dovidio & Gaertner,
1986; Pettigrew & Merteens, 1995; Rueda & Navas, 1996), together with the difficulty of expressing more
positive emotions toward the exogroup.

In the same way, the study has another limitation. The questionnaire analyzed does not lack the common
problems occurring in self-reported measurements. These problems refer to the facts that people significantly
differ in their capacity to be aware of their own internal states (Nisbett & Wilson, 1977); they are not always
willing to show their attitudes publicly due to their apprehension for evaluation (Rosenberg, 1969); and that
they try to correct their impressions (Tedeschi, Schlenker, & Bonoma, 1971) or judgments as they occur in
order to adjust them (Greenwald & Banaji, 1995, Petty & Wegener, 1998). On the other hand, it is necessary
to recognize that a great part of people’s lives is conditioned by mental processes that automatically are
activated (Devine, 1989) and would manifest themselves without the subject being aware of them (Bargh,
1999). Due to the limitations of these self-reported measurements, further studies should be done to adapt
indirect measurements (nonreactive), allowing access to people’s internal states and attitudes without directly
asking for them. These measurement procedures demand a quicker and less conscious evaluation and are
more difficult to adjust to expectations. Measurements of this type are the ‘‘task of automatic evaluation’’
(Fazio & Olson, 2003, Fazio, Sanbonmatsu, Powell, & Kardes, 1986) and the ‘‘implicit association test’’
(Greenwald, McGhee, & Schwartz, 1998; Greenwald & Nozek, 2001; Steffens, 2005).

Another limitation relates to the convenience undergraduate sample used. The generalizability limits with
such types of samples are documented, but in Chile research on this issue is just starting. Therefore, the data
reported here should be viewed as preliminary. Also, using a Spanish version of ATLG to measure subjects in
a cultural group different from the one for which it was initially developed can be problematic as reported by
Herek and Gonzalez-Rivera (2006). It is likely to find some cultural differences in the meanings assigned to
certain concepts such as ‘‘gay’’ and ‘‘natural.’’ In the future, the properties of the Spanish ATLG version also
will be better understood if it is administered to other Spanish-speaking samples.

As a conclusion, results were consistent with previous studies and the theoretical framework about
homophobia and prejudice.

You might also like