Perceived Threat of Homosexuals in Indonesia: Construct, Measurement, and Correlates

In Indonesia, homonegativity still exists and acts of violence against homosexual groups still happen, but there has not been much research on aspects of this homonegativity beyond inquiring into general attitudes towards homosexuals. This study has several aims: to qualitatively explore dimensions of perceived threat, to develop a measure of perceived threat of homosexuals, and to study some correlates and perceived threat of homosexuals in Indonesian samples. We used a mixed-method approach (qualitative-quantitative) to gather the data. Study 1 is a qualitative study that seeks to identify expressions and dimensions of perceived threat of homosexuals using thematic analysis of responses to an open-ended questionnaire and of online media articles. Study 2 develops a measure of perceived threat of homosexuals based on the results of Study 1, then establishes the factor structure and reliability of the measure, and explores some correlates of the measure. developed. The implications of using the scale to further study homonegativity in Indonesia and other Asian societies is discussed. Persepsi Ancaman atas Kehadiran Kelompok Homoseksual di Indonesia: Konstruk, Alat Ukur, dan Korelasi


Introduction
In the last five decades, there has been greater acceptance of homosexuality in many part of the world (Ahuja, 2017). Campaigns for public awareness for Lesbian, Gay, Bisexual, and Transgender (LGBT) rights have emerged in many places. Despite increasing LGBT visibility globally, there are still many places in the world where the legal status of homosexuals is inferior to that of the heterosexual majority (Gulevich, Osin, Isaenko, & Brainis, 2018). A multi-country survey conducted by the Pew Research Center (2013) showed that there is an equally widespread rejection for homosexuality in predominantly Muslim nations and in Africa as well as in parts of Asia and in Russia despite of broad acceptance of homosexuality in North America, the European Union, and Latin America. In line with this, homonegative attitudes persist in Southeast Asia, with survey data indicating that on average, four out of 10 Southeast Asians reject neighbors who are lesbian or gay (Manalastas et al., 2017), and that Indonesia and Malaysia being the two countries with the highest levels of homonegativity. Indeed, in some Southeast Asia countries e.g. Malaysia, Singapore, Brunei Darussalam as well as the South Sumatra and the Aceh province in Indonesia, same-sex sexual acts are criminalized under the law (Carroll, 2016).
In Indonesia, prejudice, hostility, and violence against homosexuals are somewhat severe. A 2018 survey found that 87.6% of Indonesians viewed LGBT people as a threat, 80% reject a neighbor who is homosexual, around 90% cannot accept leader who is homosexual, and 81.5% believed homosexual behavior was forbidden by whichever religion they adhered to (Merigo, 2019;Wibawa, 2018). It is not uncommon for some academic conferences with LGBT topics to be cancelled by Islamic organizations, or book seminars to be cancelled because the author's sexual orientation. The existence of homonegativity (homophobia, heterosexism, or anti-gay prejudice) in Indonesia points to a need to understand factors associated with negative attitudes held by the majority toward the minority group. Understanding those predictors may provide important suggestions for improving attitudes toward homosexual group (Tartakovsky & Walsh, 2015) and thus promoting good intergroup relations and enhancing wellbeing for both homosexual and heterosexual groups. In this study we focus on one factor -perceived threat of the minority group.
Intergroup Threat Theory. Threat and fear are said to be among the fundamental causes of negative intergroup attitudes (Stephan, Diaz-Loving, & Duran, 2000). That is, negative intergroup attitude is influenced by the extent to which one is perceived threatening by another group. Perceived threats have real consequences, regardless of whether or not the threat perceptions are accurate. Defensive and destructive aggression can occur in facing the feeling of being threatened. Given that perceptions of threats can predict negative attitudes and violence against outgroups that are perceived as threatening, then exploring the perception of intergroup threats is important in the case of homonegativity in Indonesia, particularly as not much is known about how LGBT groups and individuals are perceived as threats in Indonesian society.
Integrated threat theory (Stephan & Melay, 2011;Stephan, Ybarra, & Morrison, 2016) is a social psychological theory that primarily concerned with threat perceptions. The theory is a recent theoretical postulation on this subject and provides a useful and integrated framework for understanding prejudice and negative attitudes towards outgroups or minorities (Ngwayuh, 2017;Scheibner & Morrison, 2009;Wolsko, Park, & Judd, 2006). Integrated threat theory (henceforth, ITT) is focuses on perceived threat, and not with the actual threat posed by outgroups, and according to ITT, an intergroup threat is experienced when members of one group perceive that another group is in a position to cause them some form of harm.
The most recent revisions of ITT have focused on two basic threats namely realistic and symbolic threats (Ngwayuh, 2017;Stephan, Ybarra, & Morrison, 2009;Stephan et al., 2016). In addition to these two basic threats, they added a new dimension namely individual and group threat. So in essence there are four basic types of threats that vary along two dimensions. The first-dimension concerns whether the threat involves realistic (tangible) or symbolic (intangible) harm to the ingroup. Realistic threats include perceived threats to ingroup's power, resources and general welfare such as territorial threats, threats to political power, economic threats, and threats of physical harm. Symbolic threats include perceived threats that are considered harmful or attack of ingroup's religion, values, identity, belief system, norms, ideology, morality, worldview, or way of life. The second-dimension concerns whether the threats are perceived to be directed at the ingroup as a whole or individual member. Individual threats are threats that perceived to be directed at individual members. Meanwhile, group threats are threats that perceived to be directed at the whole group members. In combination, the two-dimension result in four types of threat: realistic group threat, symbolic group threats, realistic individual threats, and symbolic individual threats (Stephan et al., 2016). Stephan and Melay (2011) stated that to be able to make an impact, threats from outgroup must be evaluated as credible. Regardless the threat is real or not, as long as the ingroup perceived it as credible sources that potentially harmful for the existence of the ingroup, then the perception is able to lead ingroup responses to the December 2019 ½Vol. 23 ½ No. 2 outgroup. There are number of consequences of perceived threat include cognitive (stereotyping, ethnocentrism, dehumanization, etc.), affective (fear, anger, disgust, despair, etc.), and behavioral (discrimination, lying, cheating, harassment, etc.) outcomes (Gulevich et al., 2018;Riek, Mania, & Gaertner, 2006;Stephan et al., 2016). Responses to intergroup threat are determined by whether the perceived threats are symbolic or realistic. Symbolic threats are likely to elicit dehumanization, delegitimation, moral exclusion, and lessen empathy to the outgroup. Realistic threats are more to lead pragmatic responses like withdrawal, avoidance, and aggression. Given the seemingly intense and varied responses associated with homonegativity, it is interesting to explore what types of threats are perceived by Indonesians to be associated with LGBT.

Antecedents of Perceived Threat of Homosexuals.
Research in other parts of the world have identified a range of factors that may be antecedents of perceived threat, such as the relative power of the groups (Stephan et al., 2016). In this regard, Gulevich et al. (2018) suggested that if sexual orientation is viewed as innate, this belief constitutes no reason to expect that the proportion of homosexuals in the society or their power as a group should increase; but if sexual orientation is viewed as a result of social influence (e.g., some kind of "gay propaganda" that can "turn" heterosexuals into homosexuals), there might be a perception that the proportion of homosexuals or their relative power will be increased, possibly threatening the status of heterosexual. Their study found that attribution of homosexuality to social causes was a positive predictor of perceived threat of homosexuals, whereas biological causes showed an inverse effect.
The strength of ingroup identity, amount and type of contact, and outgroup knowledge are other factors associated with perceived threats (Stephan et al., 2016). Individuals with higher ingroup identity are likely to experience more perceived threats than individuals with lower ingroup identity. And because homosexuals often do not conform to the rigid standards set by traditional gender roles, they provoke anxiety in males who strongly identify with heterosexual gender roles (Gulevich, Osin, Isaenko, & Brainis, 2016).
Ingroup members who are relatively unfamiliar with outgroup tend to be prone to experience threats than who knows each other (Gulevich et al., 2016;Stephan et al., 2016). If the ingroup has more positive experience with the outgroup, the outgroup is likely to be evaluated as threatening (Ngwayuh, 2017). Ingroup members who have less personal contact with outgroups also tend to experience threat compare with who have more contact. However, this amount of contact depends on type of contact; if ingroup members experience negative interaction with outgroup, they tend to more easily feel threatened by outgroup. Consistent with these principles, the negative effect of religious fundamentalism, authoritarianism, and heterosexual identity on perceived threat of homosexuals tend to be weaker in individuals who have positive contact experience with homosexuals, even as the level of religiosity moderated the effects between both types of threats and prejudice (Makashvili, Vardanashvili, and Javakhishvili 2018).
Individual difference variables such as social dominance orientation (SDO), right-wing authoritarianism (RWA), individual self-esteem and collective self-esteem as antecedents of perceived threat (Stephan et al., 2016). Low individual self-esteem makes people tend to experience threats more; similarly, people with high collective self-esteem experience feelings of threat more than those with low collective self-esteem. SDO is proposed to elicit realistic threats, whereas RWA elicits more symbolic threats. On the other hand, an individual's traits could affect the extent to which others will perceive that individual as threatening. For example, characteristics such as laziness, irresponsibility, or unfairness (Lazarus, 1991;Shaver, Schwartz, Kirson, & O'Connor, 1987;Weiner, Perry, & Magnusson, 1988) are likely to trigger feelings of anger and threat on the part of people who are irritated by such traits.
Type of outgroup is also the one of important antecedents of perceived threat (Stephan et al., 2016). For instance, economically competitive outgroups may pose realistic threats related to ingroup's potential losses of resources; outgroups that carry diseases may pose realistic threats related to ingroup's fear of contamination; and outgroups that are perceived as socially deviant, may more easily elicit ingroup's symbolic threats. And it may be that perception of LGBT groups as deviants are related to perceptions of threat of homosexuals.
As the preceding paragraphs indicate, that are a range of factors that might shape how individuals may perceived LGBT groups as threats, and there are no published studied that explore such factors might shape perceived threat of homosexuals in Indonesia. But to study such factors would require the use of valid and reliable measures of perceived threat that are appropriate to the Indonesian context, and that take into consideration dimensions of the perceived threat of homosexuals among Indonesians. There exist perceived threat scales but they are not directly relevant to studying homonegativity in Indonesia, as they mainly focus on perceived threat of immigrants and of ethnic groups (e.g., Stephan & Stephan, 1996;Tartakovsky & Walsh, 2015). A perceived threat of homosexuals scale that have been developed by Clifton (2011), but this scales focusses only perceived threat of same-sex marriage. Another related scale was developed by Gulevich et al. (2016;, but the scale was developed to measure such perceptions among Russians, and may not be appropriate December 2019 ½Vol. 23 ½ No. 2 to capture perceived threats as experienced by Indonesians. We did not find any related scale developed and used in similar Asian societies, and the study was motivated by the understanding that the availability of such a scale might help better understand negative attitudes towards homosexuals in the Indonesian region.
The Current Study. To summarize, this study has several interrelated aims, all of which relate to developing a scale to measure perceived threat of homosexuals in Indonesians. The first aim is to identify the specific perceived threats of homosexuals from the perspective of Indonesians to be used in the scale. The other aims of the study relate to exploring and validating the factor structure of the scale, and to exploring some correlates of the scale.
To address these various aims, we adopted a mixed methods approach, first using qualitative thematic analysis to identify specific forms of perceived threat related to homosexuals. The results of the qualitative study were used to develop the scale which was studied using a series of quantitative analyses of data generated from the administration of the scale. The details of these mixed method approach are described in Study 1 and Study 2 are described below.
According to ITT, perceived threat is one of the predictors to explain negative attitudes towards certain minority groups and/or outgroups, and is therefore, important to consider in efforts to understand the reasons why individuals have a negative attitude towards homosexuals in Indonesia. In the long run, a better understanding of the reasons will help guide efforts to reduce homonegativity in persons and in social groups in Indonesia, and the ability to credibly measure perceived threat of homosexuals is an important component of this effort. A credible scale of perceived threat of homosexuals can be used by researchers to explore predictors and antecedents of homonegativity comprehensively, and point to factors that can be addressed in policy, social and group intervention programs targeting the reduction of homonegative and its consequences in Indonesian society, and even in other similar societies.

Study 1
The aim of study 1 was to develop items for a perceived threat of homosexuals in Indonesians based on two sources of public discourse on homosexuals: responses to open-ended question and opinion posts in online media. A sample of Indonesians were asked to respond to open-ended questions that inquired about their perceptions about their general opinions and attitudes towards LGBT persons and also experiences that relate to feeling uncomfortable or threatened by the same persons. Online media posts expressing opinions related to the same topics were also gathered, in line with the proposition that public opinion is one component of the social ecology that contribute to LGBT persons' experiences of prejudice (Herek, 2007;Herek & McLemore, 2013). Moreover, public opinion provides important basic descriptive information about how LGBT groups are perceived by a particular society (Manalastas et al., 2017) and also has been used as a ultimate component in popular metrics that measure a country's level of friendliness to LGBT people (Lemke, Tornow, & PlanetRomeo.com, 2015). These two sources of qualitative data were analyzed to generate depictions and categories of perceived threat involving homosexuals.

Methods. Participants.
For the open-ended question, 346 undergraduate students (mostly from a private Catholic university in Yogyakarta) were recruited. Their mean age was 19.62 years (SD = 1.69), predominantly female, heterosexual, and Catholic (for more details information see Appendix Table 1). The predominance of students and Christian participants in the open-ended survey is intended to balance religious background of the qualitative data sources, given that the posted social media opinions mostly come from the Muslim adults (see table 1). We delivered the open-ended question through paperand-pencil test along with the demographic questions. The participants wrote their responses for both questions and we analyzed participants' responses verbatim. We did not ask opinion of gays and lesbians separately due to public opinion in mass media more addressed to homosexuals in general.

Corpus on online media posts.
For the second data corpus, online articles posted in two popular websites were collected: Kompas.com and Kompasiana.com; the two are online websites both under the auspices of the Kompas Daily newspaper. The goal of using public opinion documents or discourses posted on these two websites is to broaden the range of individuals whose opinions were thematically analyzed, beyond those of the predominantly Christian students who answered the two questions above. However, information on the authors of the online posts were not gathered nor summarized. Kompas Daily was chosen because Kompas is a national newspaper that most widely read since its first publication (Wardhana, 2014). Kompas's readers are approximately 2.25 million and are spread throughout Indonesia's areas, and widely assumed to represent Indonesian perspectives (i.e., instead of Western media perspectives).
The article search was carried out in these two websites by using the keyword "attitudes toward homosexuality" or "attitudes toward homosexuals." The word "threat" was not used as a search word because no articles used the term "threat" explicitly. The articles that were selected for analysis were evaluated to the extent that they contain the target topic, which includes the threats felt by the presence of homosexual groups. From Kompas.com website, 45 articles from five webpages were retrieved, but only one article directly contained statements on the target topics. From Kompasiana.com website, 137 articles from seven webpages were retrieved, and 37 of these had statements related to the target topic. Only the 38 articles with statements about perceived threat were included in the final analysis.
Qualitative data analysis. The two data corpuses were analysed to identify main themes related to perceived threat of homosexuals, with the first dataset representing mainly opinions of university students who were predominantly Christian, and the second dataset representing opinions of a wider range of Indonesian persons, including journalists and nonprofessional writers from various groups of people, cultural backgrounds, religious and also included community leaders, experts, and professionals.
The two datasets were separately analyzed using thematic analysis following analytic procedures based on the six stages suggested by Braun and Clarke (2006). Data were analyzed into meaning units, then each meaning unit was categorized based on the similarity of its meaning, and finally, each subtheme was grouped based on the similarity of the theme. If the respondent's answer on the responses to open-ended question and online articles contained more than one meaning units, then the researcher analyzed all the meaning units that appear on data. Thus, the number of meaning units or the number of references obtained can be more than the number of respondents and the number of online articles obtained.
The coding of themes in the thematic analysis was conducted by two researchers (the first two authors) independently to ensure the credibility of the codes identified. The first coder is a doctor in clinical psychology, with an interest in clinical psychology research in a socio-cultural context, while the second researcher. The second coder has a master's degree in social psychology. After the independent coding, the two researchers compared their codes to discussed discrepancies in order to achieve inter-coder consensus. After comparisons the two coders refined the coding frame, and following an iterative process, reviewed the initial codes and interpretations until final agreement was achieved.
Although the thematic analyses were conducted separately, similar categories or themes were identified in the students' answers to the two questions and in the articles retrieved form the two websites. In other words, the data analysis had reached saturation as no new themes that emerged by looking at the other dataset. The themes with sample quotations for the two data sets are summarized below.    Table 2). Our findings showed that homosexuals were perceived bringing realistic threats (31.10%) and symbolic threats (48.30%) to the community. In general, homosexuals are considered more threatening society symbols and values, for example, they are potentially damaging the moral values of society and young generation (15.50%), to fade Indonesia's customs and culture (10.4%), and to influence others to imitate their sexual orientation (10.8%). The presence of homosexual groups is also seen as bringing a realistic threat, such as causing conflict and friction in the community (7.7%), spreading sexually transmitted diseases (5.5%), and reducing the population (5.3%). However, a small number of participants (10.2%) stated that homosexual groups did not pose any threat; instead they were the ones who received threats from the community (4.2%).

Makara Hubs-Asia
December 2019 ½Vol. 23 ½ No. 2 Thematic analysis of online articles resulted in 12 subthemes of threats that categorized under two main themes entitled realistic threat and symbolic threat (see Table 3). Symbolic threats were more common than realistic threats; homosexuals were considered to be in conflict with religious values (29%) and influencing others to imitate them (14.5%). Realistic threats also emerge in the themes of online articles, such as disturbing the security of the community, spreading sexually transmitted diseases, or reducing the population.
Comparison of subthemes between the responses to open-ended questions and the online media articles indicated eight similar subthemes namely "triggering criminality", "emerging sexual crimes", "spreading sexually transmitted diseases", "causing mental disorders", "against/weakening religion and human nature", "damaging morality", "against society/ Indonesian's customs and cultures or values", and "influencing others to imitate". Subthemes that came up from mass media articles but not from the open-ended questions are "endangering family values" and "homosexuality is illegal-unconstitutional". Meanwhile, subthemes that only emerged from the open-ended questions are "damaging Indonesia's reputation", "making people insecure", "reducing opportunities of getting partner", and "decreasing population". Based on this comparison, it can be concluded that there was similar pattern between public opinion from mass media articles who represent mostly Muslim and adult (middle and late adult) sample and participants' answers from the openended questions who represent mostly Christian and early adult sample.
The results of Study 1 revealed specific expressions of perceived threat of homosexuals in Indonesia. These findings were then used to construct a scale of perceived threat with two main subscales (realistic and symbolic threat), which is described and further analyzed in Study 2 below.

Study 2
Study 2 focuses on the development of the perceived threat of homosexual's scale for Indonesians and studies on its factor structure and some correlates. The organization of this section of the paper will reflect the different components of the study.
For the scale development, item selection was firstly based on subthemes of participants' answer from the open-ended questions that achieved more than 5%. The mass media subthemes used to confirm the subthemes that selected from the open-ended questions. After selection based on percentage and confirmation process, the researchers develop seven items that represented realistic and symbolic threats. The seven final items and its categorization are presented in Table 4.
The empirical and analytic aspects of Study 2 refer to the following: an exploratory factor analysis, a confirmatory factor analysis, the test of internal consistency, and exploration of correlates with demographic characteristics, and finally, test of correlations with other measures of intergroup attitudes to examine convergent and discriminant validity. The data for these analyses were derived from a new sample of participants. The different aspects of the methods, analysis, and results are described below.

Methods. Participants. Participants involved in Study 2
were 410 students from the Psychology Department of a private Catholic University in Yogyakarta. None of the participants were part of the sample of participants in Study 1. The mean age of participants was 19.6 years (SD = 1.39) and further details on characteristics of the participants are presented in Table 5. For the exploratory and confirmatory factor analysis procedures, the total sample was randomly split into two; data from 195 were used in the exploratory factor analysis, and 215 were used in the confirmatory factor analysis (215 participants).

Instruments. The Perceived Threat of Homosexuals
Scale. This is the seven-item scale developed from results of Study 1 and summarized in Table 4.
Participants were asked to indicate their degree of agreement for each statement using a Likert scale from 1 (strongly disagree) to 5 (strongly agree). Evaluative Bias Scale (Wolsko et al., 2006). This scale consists of six items that assess attitudes about the degree to which individuals view their ingroup more positively than their outgroups. Participants responded using a Likert scale from 1 (strongly disagree) to 7 (strongly agree). The internal consistency computed for the current sample was α = 0.77.

Tablel 4. Items of Perceived Threat of Homosexuals
The Scale of Ethnocultural Empathy (Wang et al., 2003). Two subscales of this scale were used: empathic perspective taking and empathic awareness. The emphatic perspective taking subscale consists of seven items, while the emphatic awareness subscale consists of four. For all items, participants responded using a Likert scale from 1 (strongly disagree) to 6 (strongly agree). Internal consistency was computed for the current samples, and these were α = 0.65 and α = 0.81, respectively.

Results. Exploratory Factor Analysis.
Before a factor analysis is performed, the researchers conducted a data normality testing. The results showed that the scale has a normal distribution that was indicated from its nullity and z value of the Kolmogorov-Smirnov testing was under 1.97 with p > 0.05 (z = 0.84, p = 0.49) or D value = 0.06, p = 0.49 (p > 0.05).
The test of the assumptions for principal axis factoring (Leech, Barrett, & Morgan, 2005) indicated that all items were adequate for each factor (Kaiser-Meyer-Olkin = 0.86). Based on theoretical framework that the types of threats are orthogonal, the varimax orthogonal rotation was applied in the principal axis factoring. The results showed that PTHS consists of only one factor that accounted for 43.85% of the variance, with factor loading of each item around 0.46 to 0.81. The factor loading was presented in Table 6.
Confirmatory Factor Analysis. The factor of the scale was further studied using confirmatory factor analytic procedures. Confirmatory factor analysis was conduct using R 3.6.1 using robust maximum likelihood estimator with Yuan-Bentler (Yuan & Bentler, 1998) and using fitness index criteria of Hu and Bentler (1999). Note: Model A = one-factor model; Model B = two-factor model-realistic-symbolic Two model of PTHS were compared; model A is the one-factor model suggested by the exploratory factor analysis and Model B is a two-factors model that is suggested by the theoretical distinction between realistic and symbolic forms of perceived threat. As shown in Table 7, the model comparison indicated that Model A (one-factor model) had better fit with the data compared to Model B (two-factors model). Thus, the confirmatory factor analysis result supported the results of the exploratory factor analysis.
Internal consistency. Reliability testing on seven items of The Perceived Threat of Homosexual Scale (PTHS) was conducted using data from the complete sample. The analysis showed that the PTHS has a good internal consistency (α = 0.84) and no items have a total item correlation below 0.30. The complete analysis results can be seen in Table 8.

Correlations with Demographic Antecedents.
Data from the complete sample were used in a regression analysis; but because of missing data from eight participants, only 402 were included in the final analysis. In the regression analysis, the perceived threat of homosexual scale score was regressed to different demographic variables, and the results indicated that there is no association between age, sex, and sexual orientation with perceived threat of homosexual (note: sex and sexual orientation were coded as dummy variables for the regression analysis). Meanwhile, religion identity (β = 0.15, p < 0.01), and having relatives who are homosexuals (β = -0.18, p < 0.01) related to perceived threat of homosexuals. As predicted by previously studies and literatures review, the stronger the people identifying to their religion, the more they feel threatened by homosexuals. The result also showed when individuals have relatives who are homosexuals, they tend to perceived homosexuals are unthreatening. The overall regression model was significant [F(5, 396) = 5.97, p < 0.01], and total R 2 = 0.07

Convergent and Discriminant Validity.
To test the discriminant and convergent validity of the developed scale, we hypothesized that the scale scores would shows moderate correlation with Evaluative Bias Scale and a weak correlation or have no association with The Scale of Intergroup Empathy. The Evaluative Bias Scale assesses attitudes about the degree to which they view their ingroup more positively than their outgroups (Wolsko et al., 2006), and studies show that perception of threats influences the evaluation of others and vice versa (Stephan et al., 2009 , 2003). We assumed that those who report higher levels of group empathy would show report lower leves of perceived threat of homosexuals. As such, the scale is assumed suitable for testing the discriminant validity of the perceive threat of homosexuals scale.
Before running the analysis, normality testing was conducted using unstandardized residuals for all the variables. Normality analysis results in z values less than 1.97, suggesting that all unstandardized residuals are normally distributed. The descriptive statistics and correlations of key variables presented in Table 8.
Pearson's correlation showed that PTSH was significantly correlated with the Evaluative Bias Scale, the Empathic Awareness subscale, and the Perspective Taking Subscales. As we hypothesized, there was a positive correlation between PTSH and the Evaluative Bias Scale indicated convergent validity of these two scales.
Meanwhile, the small but statistically significant correlations between PTSH and both the Empathic Awareness and Perspective Taking Subscales is evidence of the discriminant validity of these two scales.

Discussion
Previous studies of intergroup relations show that perceived threat of outgroups regardless its real or not, if considered credible, tend to increase ingroup negative attitudes toward outgroups. This study intended to understand public opinion of perceived threat of homosexuals and to develop a useful measure for this perceived threat. The results of the qualitative study found that the main concern of the society was the bad influence caused by the homosexuals. People viewed homosexuals are deviant from noble and religious values so that they are perceived as a threat to moral and religion values. Homosexuals were also seen actively influencing others, so people are concerned if homosexuals influence heterosexuals, especially young generations becoming homosexuals. This is considered as a salient threat in damaging the morality of the Indonesian people.
The presence of homosexuals that deviates from social norms was also viewed threatening of social harmony as their presence might lead to rejection and violence from non-homosexuals. In Indonesia and other Asian countries, maintaining social harmony is important, so anything that potentially trigger conflict is viewed as a threat to the society.
Both of these concerns may be able to explain Manalastas et al. (2017) study why people in Indonesia as well as in Malaysia rejected homosexuals as their neighbors.
Rejection of homosexuals' neighbors may be also related to fears of contracting sexual diseases. In Indonesia, HIV/ AIDS is always attached to homosexuals. So, the presence of homosexuals was perceived as a threat of individuals' and society' health safety.
People perception of homosexuals threats in this study has some similarities with Gulevich et al. (2016) findings in Russian context. Russian people also considered homosexuals as a threat to the morality, deviating from the social norm, cannot be morally justified, and challenges the values and culture of society. Homosexuals were also viewed aggressively influencing children and heterosexuals to turn into homosexuals. As well as Indonesian people, Russians also afraid of contaminating of the transmission of sexual diseases and concerned with declining population. The main difference between Gulevich's and this finding is the perception of source of threat. In Russia homosexuals were viewed as product of Western culture, so the main threat was about westernization. While in Indonesia, the source of threats is the violation and destruction of moral and religious values that was perceived more threatening compared with westernization.
Considering that the types of perceived threats will lead to different negative attitudes, then exploring and developing a perceived threat scale that contextual with the society will greatly helpful in understanding intergroup attitude and its association with the perceived threat. Therefore, the development of PTSH scale based on public opinion in Indonesia is expected to be beneficial for other studies that intending to explore perceived threat as a predictor of homonegativity. This scale was also expected can be used by researchers from other countries that have similar characteristics with Indonesia such as Malaysia, Vietnam and Singapore.
The Perceived Threat of Homosexuals Scale (PTSH) developed in this study is the first scale of perceived threat of homosexuals in Indonesia. This scale was constructed based on public opinion through bottom-up approach, therefore this scale is expected to represent lay people perceptions of homosexuals' threat. Based on EFA and CFA, seven items of this scale that firstly be categorized in realistic and symbolic threat, grouped in only one factor. This result is a bit different from previous literatures of intergroup threats. According to Stephan et al. (2009), intergroup threats are divided into two, namely, realistic and symbolic threats. The results of the qualitative analysis of this study also indicated that there are two types of intergroup threat. However, from PAF, this scale is better categorized in one factor. This can be understood considering that homosexuality is more prominent when it is associated with the issue of morality, which in threat theory is included in symbolic threats. Although the results of the thematic analysis found two types of intergroup threat, the frequency of participants' answers centered on the theme of morality. In other words, the salient threat of homosexuals perceived by lay people is a symbolic threat. Many December 2019 ½Vol. 23 ½ No. 2 rejections of homosexuals are based on the assumption that homosexuals are against moral and religious values. Thus, it might explain why PAF analysis only showed one factor.
Study 2 indicated that this scale has a adequate discriminant and convergent validity. A positive correlation between PTSH and Evaluative Bias Scale and negative correlation with both Empathic Awareness and Perspective Taking subscale indicated that PTSH is able to measure perception of homosexual threats exclusively, not overlapping with other concepts in intergroup area. PTSH also has a good internal consistency and adequate construct validity.
The results of exploration of association between some antecedents and perceived threats of homosexuals showed that sex and age were not significantly related to perceived threats of homosexuals. This nonsignificant association between age and perceived threats may be due to the age homogeneity of the participants who were all undergraduate students. Another finding of this study that indicated no relationship between sex and perceived threat of homosexuals seems to be similar with Manalastas et al. (2017) findings which showed that there was no difference in homonegativity by sex in Indonesia, Singapore, and Thailand. For the future research, it will be more interesting to use gender identity rather than sex as a predictor, because previous studies have shown that prejudice against homosexuals is generally found in men whose traditional masculine identification is very high (Gulevich et al., 2016). Homosexuals that are not in harmony with the standard role of masculinity will trigger anxiety and feeling threatened.
Some findings of this study are also in line with study of Gulevich et al. (2016) that showed perceived threat of homosexuals was more common in people who identify themselves as heterosexual than partially or completely homosexuals. Previous studies have shown that familiarity with outgroups reduces feeling threatened (Gulevich et al., 2016;Ngwayuh, 2017;Stephan et al., 2016), therefore why in this study having family who homosexuals also a predictor of perceived threat of homosexuals.
The combined results of our qualitative and quantitative study indicated that The Perceived Threat of Homosexuals Scale is a good scale to measure the perception of homosexual threats of Indonesian people. However, the scale can be further validated in a wider Indonesian sample that includes non-University students from different regions of Indonesia. But with the findings from this study, so far, there is good preliminary evidence this scale expected can be used for further studies that intended to explore of people's perception of homosexuals' threats. This scale might be used by policy makers in understanding of perceived threat of homosexuals of Indonesian society. By knowing people's perception of homosexuals' threat, professional helpers, the government or policy makers can develop programs or policies for correcting the misleading perception. Besides, PTSH can also be applied in the education setting to understand students' perception of homosexual threat, so the schools can develop materials or activities that can reduce threatened feelings and enhance positive intergroup attitudes.

Conclusion
Given the severity of homonegativity in Indonesia, it is very important that all professionals (governmental officials, social workers, teachers, clinical psychologists, counselors, etc.) find ways to try to reduce homenegativity, and one means is through correcting misleading perception of homosexuals among Indonesians. This study explored perceived threats of homosexuals among a sample of Indonesians and used the findings to construct a scale that was also shown to be a reliable and valid indicator of homonegativity.
Although there is more work that can be done to further support the validity of the scale. The scale makes a small but useful contribution as a tool for further research on intergroup relations, not only Indonesia, but also in other countries that have similar characteristics.