Effects of news representations on attitudes toward people living with HIV/AIDS: a cross-cultural comparison of Taiwan and Hong Kong.
Stigma and discrimination have been considered a major obstacle to effective HIV/AIDS prevention and care. Reports of discrimination against people living with HIV/AIDS (PLWHA) are prevalent, and the negative consequences call for further corrective measures and scholarly investigation (Blendon & Donelan, 1988; UNAIDS, 2003). Protecting human rights and promoting public health are perceived as mutually reinforcing by the Joint United Nations Programme on HIV/AIDS (UNAIDS), the main worldwide advocate for accelerated, comprehensive and coordinated global action on the HIV epidemic. Nevertheless, elimination of HIV/AIDS-related stigma and discrimination continues to pose challenges for the public health policy makers and implementers in many parts of the world. For years, research has documented the barriers hindering the local stakeholders from respecting, protecting, and fulfilling the legitimate human rights of PLWHA due to various social, legal, political and economical concerns (e.g., Dodds, 2006; Herek, Widaman & Capitanio, 2005; Kohi, Makoae, Chirwa, Holzemer, Phetlhu, et al., 2006).
One of the major sources of HIV/AIDS-related stigma and discrimination is related to how the disease and PLWHA have been represented in popular discourse. News coverage plays a crucial role in portraying how such stigma and discrimination have been treated in society, especially in cultures or regions where the prevalence of HIV/AIDS is relatively low and most of their people have not had direct contact with PLWHA. Multiple studies have identified news media as important sources of HIV/AIDS related issues and information for both individual citizens and policy makers around the world (Edgar, Noar, & Freimuth, 2008; Freimuth, Greenberg, DeWitt, & Romano, 1984; Fuller, 2003; Simpkins & Brenner, 1984; Wallack, 1990; Weiss, 1974). News media, however, can serve both positive and negative functions in surveillance and social integration. Due to their prevalence in modern society, news media can have a stronger impact on public cognition, attitudes and preventive behaviors than media campaigns. On the other hand, the accuracy and amount of information provided by the media may contribute to misguided individual and pubic choices for action against the epidemic. Such misguidance and misinformation are particularly salient in representations of HIV/AIDS and PLWHA in the news.
This study thus aims to explore whether different ways of representing PLWHA in the news, specifically the transmission routes, can affect how individuals perceive the related issues. Additionally, we are interested in how gender and cultural background influence such perceptions. In this case, we chose Taiwan and Hong Kong, two societies with similar cultural backgrounds but different HIV/AIDS policies and intervention measures, for the comparison.
HIV/AIDS Epidemics and PLWHA-related Issues: Taiwan vs. Hong Kong
As two major Chinese communities in Asia, Taiwan and Hong Kong share a substantial degree of cultural commonalities. They both have low-prevalence rates of HIV infection. The first HIV cases in Taiwan and Hong Kong were both found in the homosexual Westerners in late 1984. In Taiwan, the foreign connection of AIDS tended to represent Westerners as promiscuous in the public discourse during the early years of the epidemic (Hsu, Lin, & Wu, 2004). Such a sentiment, however, was less evident in Hong Kong as it was under the British sovereignty until 1997 (Chan & Lee, 2004). As the spread of HIV among heterosexuals became more dominant, a new stigmatized group of female sex workers emerged in both regions in the late 1980s (Hsu, Lin, & Wu, 2004; Lau & Tsui, 2002). Starting from the 1990s, the concern about human mobility with imported foreign laborers from South and Southeast Asia in Taiwan (Hsu, Lin, & Wu, 2004) and the increasing human traffic between Hong Kong and mainland China after 1997 (Chan & Lee, 2004) has subjected these two regions to higher potential risk of HIV transmission.
Sexual contact had been the main transmission route in Taiwan until 2004 when injecting drug use (IDU) cases mushroomed as the highest HIV infected group. A pragmatic harm reduction plan implemented in mid-2006 quickly brought the IDU transmission under control (Center for Disease Control, Taiwan, 2007). At present, sexual contact (64.24%) still remains the major transmission route in Taiwan (Center for Disease Control, Taiwan, 2010). In Hong Kong, sexual contact (74.41%) has continued to be the major mode of HIV transmission since report of its first cases in 1984 (Department of Health, Hong Kong, 2010). Heterosexual transmission still is predominant, but a rapid surging epidemic among men who have sex with men (MSM) is witnessed lately. The HIV infection in IDU population, though still relatively low, has also become more apparent since 2006 (Hong Kong Advisory Council on AIDS, 2010). Table 1 displays the respective epidemiological profiles of HIV/AIDS transmission in Taiwan and Hong Kong.
Over the years, PLWHA's rights to medical care, education, work, marriage, and so on remain controversial in the public forums in Taiwan and Hong Kong. Health authorities in both regions have collaborated with local non-government organizations (NGOs) and/or legislators to eliminate the HIV/AIDS-related stigmas. Comparatively speaking, the Taiwanese government's efforts in this regard focused more on the medical arena, whereas its Hong Kong counterpart started to promote the acceptance of PLWHA in the public domain as early as the 1990s. For example, medical care services, outpatient and hospitalization for AIDS patients are all free in Taiwan. The government offers free highly active antiretroviral treatment for AIDS patients (Twu, Huang, Lai, & Ou, 2004), AIDS tests for pregnant women and plan of reducing injuries from AIDS (Center for Disease Control, Taiwan, 2008). However, dedicated efforts to change the public's negative responses to PLWHA are made mostly by local NGOs (Hsu, 2004). Although a formal regulation was announced by Taiwanese government's AIDS Advisory Committee to protect the confidentiality of PLWHA in 1990, it was not until 2007 that protection of PLWHA's human rights became legalized when a full-scaled statute was officially amended (Center for Disease Control, Taiwan, 2007).
In Hong Kong, legal and government efforts to eliminate HIV/AIDS-related stigmas were rigorous and extensive. Homosexuality was decriminalized in 1990. In 1995, the Disability Discrimination Ordinance was enacted to protect people against discrimination, harassment, and vilification on the ground of their disabilities. Its coverage was extended to PLWHA and their associates (Virtual AIDS Office of Hong Kong, 2005). To minimize marginalization of PLWHA, the Committee on Promoting Acceptance of PLWHA was formed in late 1999 to enable a comprehensive control of the epidemic. Media campaigns appealing to the public to accept PLWHA have been launched over the years as an ongoing endeavor (Virtual AIDS Office of Hong Kong, 2010).
Theoretical Models Explaining HIV/AIDS-related Stigmas
Stigma refers to a type of socially constructed identity that is devalued in a particular context (Archer, 1985; Crocker, Major, & Steele, 1998). According to Albert (1986), what is underlying and fuelling the stigmatization of HIV/AIDS risk groups is the fear that the social distinctions for protection will be breached, leaving the general population open to the fatal infections of HIV/AIDS. Thus, the maintenance of social distance is crucial, and this results in blaming an individual as responsible for his/her own disability.
Public attitudes toward stigmatized groups, including PLWHA, are often complex, conflicted ambivalent, and ultimately mixed with sympathy and fear (Herek & Glunt, 1991). Schulte (2002) proposed two distinctive models to help explain how disease-related stigmatization occurs, specifically the behavioral model and the cultural conflict model.
The Behavioral Model
This model explains variation in stigmas by examining the differences between stigmatized and non-stigmatized diseases and/or statuses, giving particular attention to the behavioral culpability of the afflicted. More specifically, the behavioral model of stigma focuses on the physical aspects of the illness, such as behaviors of the patient and causes of the disease (Schulte, 2002). If a person is afflicted with a disease defined as "onset-controllable," he/she will receive more blame. From this perspective, it can be argued that those PLWHA contracting HIV/AIDS via "morally incorrect" behaviors such as MSM, heterosexual contact with multiple partners, or IDU will receive more stigmatizing treatment than those perceived "innocent" victims who contracted the disease via perinatal transmission or blood transfusion. Devine, Plant and Harrison (1999) used the term "controllable causality" to further explain why the public is more sympathetic with and willing to extend help to PLWHA due to perinatal or blood transmission, for those PLWHA are not presumed to have behavioral culpability for their HIV infection.
The Cultural Conflict Model
Such a model explains the observers' attribution of stigma as a result of attitudes acquired via acculturation or socialization. While the behavioral model assumes the reactions to stigmatized illnesses are relatively consistent across observers, the cultural conflict model of disease-related stigma focuses on differences among observers, especially observers' views of the afflicted (Schulte, 2002). This approach is consistent with the research tradition on the formation of symbolic attitude, where individual attitudes are perceived as resulting from general principles and moral philosophies.
In this sense, HIV/AIDS-related stigmas may be based on individual projections of different perceptions, namely symbolic and instrumental (or pragmatic) attitudes (Grandall & Glor, 1997). The latter is related to personal anxiety over the threat of the disease, which is affected by perceived severity, susceptibility, and preventability of the disease. In other words, the instrumental attitude related to HIV/AIDS would be the extent to which infection is feared. In contrast, the symbolic attitudes are associated with individual stereotypes of high-risk groups (e.g., homosexuals, IDUs or specific ethnic groups). When it comes to HIV/AIDS, the cultural conflict model associates attitudes toward PLWHA with individuals' prejudiced attitudes toward homosexuality, as has been evidenced in various research studies (Brecher, 1988; D'Angelo, McGuire, Abbott, & Sheridan, 1998; Rogers, Dearing, & Chang, 1991; Beharrell, 1998; Lupton, 1992; Traquina, 1996; Johnson, Flora, & Rimal, 1997; Triplet & Sugarman, 1987).
The cultural conflict model also proposes that stigmas may be affected by individual differences, including differences in gender. Research on medical care workers found that female workers were more willing to invest their time and effort in providing care for PLWHA than their male counterparts. They also had fewer negative stereotypes of PLWHA (Heaven, Connors, & Kellehear, 1990; Ladany & Stern, 1998; Rezler & Ten Haken, 1984; Stern, Norman, & Komm, 1993). In comparison, male medical care workers were more willing to render their services when the AIDS patients were heterosexual and contracted the virus via blood transfusion (Ladany & Stern, 1998).
HIV/AIDS News Story Formats: Base-rate Information vs. Exemplars
Upon comparing the explanatory power of the symbolic and instrumental attitudes, this study also employs a story format that is different from traditional news reporting about HIV/AIDS issues. Hsu (2001) content-analyzed the HIV/AIDS related news stories covered by eleven Taiwanese newspapers in 1995 and 1996 and found that the majority of coverage was related to intervention policies and activities. Sources featured in the stories were primarily experts and officials. The unique voices of PLWHA were surprisingly under-represented. According to Zillmann and Brosius (2000), base-rate information such as statistical parameters featured in a news story is less likely to be fully processed by the audience when compared with an exemplar that illustrates concrete anecdotal about individual experiences or accounts related to a specific problem. The press tendency to report PLWHA-related issues in forms of base-rate information is likely to blur the individual perspective of PLWHA and consequently marginalize their presences and exacerbate existing stigmatizations toward them. This study, therefore, employed a news story in forms of exemplar to examine the effect of transmission routes (with implied behavioral culpability) on news readers' stigmatic attitudes toward PLWHA.
To summarize, the formation of HIV/AIDS-related stigmas is a complex process that needs further scholarly investigation. This study, employing an integrated perspective of the behavioral and cultural conflict models, by taking account of news framing, aims to explore the effect of the perceived behavioral culpability associated with the transmission routes on perceptions of PLWHA and their rights. Furthermore, the possible influences of instrumental attitudes in perceiving the threat of the disease and symbolic attitudes toward a typical high-risk group are also taken into account. Finally, the roles of gender and societal differences are investigated to provide a more holistic picture of the topics under study.
Following the research literature on the behavioral model (Schulte, 2002), it is easy to predict that individuals contracted HIV via less controllable routes, such as perinatal transmission or blood transfusion, will induce more supportive attitudes from the public than those via controllable routes, such as MSM or heterosexual contacts. Nevertheless, when we present the story in a format of "exemplar" instead of "base-rate information," news representations may exert different effects on the readers. Therefore, we only hypothesize that there are differences in individual attitudes toward PLWHA in terms of news framing, rather than predict a specific direction of the attitude strength. It is hypothesized that:
H1: There are differences among news readers in their attitudes toward PLWHA regarding how controllable the HIV/AIDS transmission route is framed in a news story.
According to the cultural conflict model (Schulte, 2002) and studies of symbolic attitudes (Herek, 1987; Grandall & Glor, 1997), we hypothesize that:
H2: The more negative attitudes news readers have toward homosexuality, the less supportive attitudes they will have toward PLWHA.
According to the literature on instrumental attitudes (Grandall & Glor, 1997), we posit that:
H3: The more fearful news readers are about HIV infection, the less supportive attitudes they will have toward PLWHA.
As the gender effects found in studies on attitudes toward PLWHA are based on medical workers (e.g., Ladany & Stern, 1998), it is questionable whether such findings can be observed in those who do not have direct contact with PLWHA. In fact, international studies on gender differences at the level of discriminatory attitudes against PLWHA have yielded mixed results (e.g., Davis, Noel, Chan, & Wing, 1998; Herek & Capitanio, 1993; Kaplan & van den Worm, 1993; Lau, Tsui, & Chan, 2005). Therefore, we only hypothesize that there are gender differences in individual attitudes toward PLWHA, rather than predict a specific direction of the attitude strength. Thus,
H4: Men and women are different in their supportive attitudes toward PLWHA.
As Hong Kong government has established more concrete preventive/educational measures against discrimination of PLWHA over the years than its Taiwanese counterparts, it is expected that such efforts may have generated a measurable positive impact on its residents. Thus, the hypothesis is posited as follows:
H5: Hong Kong news readers have more supportive attitudes toward PLWHA than their Taiwanese counterparts.
To examine the causal relationship between the variables under study, an experiment was conducted. A 4 (news framing of the transmission route) x 2 (gender) x 2 (place of residence) factorial design was employed. Among the three between-subject factors, gender was inherent and place of residence enabled the cross-cultural comparison while news representation of the transmission route was manipulated to include four experimental conditions. These conditions consisted of four versions of a news story varying in terms of how an AIDS patient contracted the disease, including via MSM, heterosexual sex with multiple partners, blood transfusion, and an unspecified route. The first three versions vary in terms of the behavioral culpability for the infection and thus how stigmatic they may seem to news readers. The final version served as a control as no infection route was specified.
Pretests were conducted in both Taiwan and Hong Kong to assess clarity and comprehensibility of the experimental news stories (pretest 1) and measurement items (pretest 2). The outcomes provided useful references to revise the news stories and questionnaire. To ensure the consistency of the experimental texts while preserving the unique journalistic integrity employed in these two regions, the Taiwanese version was used as a master version and linguistic adjustments were made in the Hong Kong version to simulate local news and wording customs. All news story versions were revised by professional journalists and proofread by academic staff teaching advanced reporting in Taiwan and Hong Kong. Special effort was made to minimize any possible substantial differences between the Taiwanese and Hong Kong versions except the dialectal idioms and journalistic styles used in the two places.
The experimental stimulus was presented as a printout of a news story from online Chinese newspapers. Four versions of a news story were created by manipulating the transmission route of the AIDS case featured in the news story. To minimize participants' deliberate attention to the experimental stimulus, three other news stories unrelated to HIV/AIDS were used as fillers. The experimental stimulus was placed as the second or fourth news items to be read by the participant. The order of presentation was counterbalanced. Each news story was followed by questions measuring readers' self-assessed interest in, relevance, credibility, and accuracy of the news story. A battery of questions measuring participants' overall attitudes toward PLWHA as well as their symbolic and instrumental attitudes concerning HIV/AIDS and related issues was also posed after the stimulus news story. Demographic data such as age and sex were collected as well.
A total of 984 college students from two major public universities in Taiwan (N = 480) and Hong Kong (N = 504) were recruited to participate in the experiment. A 3: 2 ratio of female to male participants was observed in both regions, which reflected the gender distributions in these two universities. The age of participants was comparable in both regions as well (M = 20.01, SD = 1.52 for Taiwan; M = 20.34, SD = 1.11 for Hong Kong). Participants were randomly assigned to the four different experimental conditions. They were told that the purpose of the study was to investigate their preferences for and responses to specific news topics and issues. Each data collection session ranged from 20 to 30 minutes. Upon completion of the experimental session, each participant was given either coupon (Taiwan) or cash (Hong Kong) reward equivalent to US$8.00. All participants were debriefed via email after all of the data collection sessions were conducted.
To further process the battery of questions that measured participants' attitudes toward PLWHA, HIV/AIDS, and related issues, principal components analyses were conducted to extract factors that could explain the greatest amount of the total variance. Only factors with eigen values greater than one were considered for the dependent variables. For each covariate, we extracted items loaded on the first factor as we were more interested in measures of a conceptually single phenomenon. The results of KMO tests and Bartlett's tests of sphericity were noted to ensure the multivariate normality and sampling adequacy of the question items included. To select question items that best measured the variables under study, the results of reliability tests and construct validity were also taken into consideration.
To examine the causal relationship between independent and dependent variables, the data collected were primarily analyzed using multivariate covariate analysis (MANCOVA). MANCOVA examines whether there are differences among the dependent variables simultaneously, which reduces the chance of experiment-wise Type I error (Hurlburt, 2006), and makes statistical adjustment possible by controlling the influence of covariates on the outcomes. Pearson correlation tests and Bartlett's test of sphericity were conducted to ensure that the covariates and dependent variables were correlated and the use of MANCOVA was indeed appropriate. Subsequent univariate F-tests were conducted to examine the pattern of changes in the four dependent variables. The Bonferroni correction therefore was used to safeguard against Type I error, with the univariate significance criterion being set at a = .05/4 = .0125.
Dependent Variables. The dependent variables under study were under two general dimensions: individual views on PLWHA's rights, and attitudes toward PLWHA. Measurement items were first created based on the relevant literature on HIV/AIDS stigmatization and intervention (e.g., Dodds, 2006; Herek, Widaman, & Capitanio, 2005; Kohi, Makoae, Chirwa, Holzemer, Phetlhu, et al., 2006), with revisions made to take local situations into account, and were then pretested on convenience samples (pretest 2: Taiwan N=34; Hong Kong N=43) for modification. Typical Likert scales were employed to measure participant attitudes, with one meaning "strongly disagree" and five meaning "strongly agree." Means were computed after summing the scores for specific measurement items for variables under study. The higher the mean, the more positive the attitude.
Fourteen measurement items were originally included to assess participant views on PLWHA's rights. A principal components analysis using varimax rotation yielded three factors (accounting for 47.08% of the total variance explained; KMO = .88, Bartlett's test [chi square] = 2787.06, p < .001). The three items (i.e., right to medical care, right to work, and right to anonymous screening) loaded on the third factor do not appear to be conceptually unidimentional, and the reliability coefficient is low (Cronbach's [alpha]= .35). We thus retained items loaded on the first two factors for further analyses. The first factor, named as attitudes toward PLWHA's rights to work, mobility, and education, consists of seven items related to PLWHA's rights that may involve casual contacts with other people (accounting for 29.78% of the total variance explained, Cronbach's [alpha]= .80). The second factor originally consists of four items (accounting for 9.43% of the total variance explained). After both theoretical and statistical considerations, we selected the two items addressing PLWHA's intimate contacts with other people, and named the variable attitudes toward PLWHA's rights to marriage and birth-giving. It should be noted that even although the reliability level is not statistically satisfactory (Cronbach's [alpha]= .48), these items are deemed conceptually related. (1) The upper half of Table 2 displays the corresponding measurement items to and the item loadings on each factor.
Similarly, among the twelve items that measured participant attitudes toward PLWHA, three factors were extracted (total variance explained = 56.14%; KMO = .85, Bartlett's test [chi square] = 3214.37, p < .001). Measurement items loaded on the two factors were selected to form the other two variables: willingness to interact with PLWHA (five items; total variance explained = 34.04%) and intention to advocate for PLWHA (five items; total variance explained = 12.44%; Cronbach's [alpha]= .78). The two items composing the third factor are more related to the AIDS disease than PLWHA, and were thus not selected. A reliability test also reveals that three of the five willingness to interact with PLWHA items yielded the greatest alpha value (Cronbach's [alpha]= .82). They were thus retained for further analyses. The middle portion of Table 2 displays the corresponding measurement items to and the item loadings on each factor.
Covariates. Two covariates, fear of infection and attitudes toward homosexuality, were selected to represent news readers' instrumental and symbolic attitudes toward HIV/AIDS and related issues. The operationalization of the covariates was as follows:
Fear of infection. Six items were originally employed to measure the participants' fear of contracting HIV/AIDS. A principal components analysis was performed to select items loaded on one factor (26.42% of total variance explained; KMO = .55, Bartlett's test [chi square] = 317.92, p < .001). Two out of five selected items were used to construct the covariate, fear of infection. Although the reliability level is not high (Cronbach's [alpha]= .56), as seen in the middle portion of Table 2, they are considered to measure the same dimension conceptually and were thus retained for further analyses. The higher the score, the less was the concern about infection.
Attitudes toward homosexuality. Eleven items were used to assess participant attitudes toward homosexuality (see the lower portion of Table 2). The principal components analysis result shows that all eleven items loaded on the same factor (accounting for 47.25% of the total variance explained; KMO = .99, Bartlett's test [chi square] = 4255.09, p < .001; Cronbach's [alpha]= .88). The higher the score, the more positive was the attitude toward homosexuality.
A Pearson correlation test shows that there was no significant correlation between these two covariates (r = -.02, n.s.). They were however found significantly correlated with two and four dependent variables, respectively: fear of infection was positively related to rights to work, mobility, and education (r = .10, p < .01) and willingness to interact (r = .09, p < .01); attitudes toward homosexuality was positively related to rights to work, mobility, and education (r = .30, p < .001), rights to marriage and birth-giving (r = .19, p < .001), willingness to interact (r = .29, p < .001), and intention to advocate (r = .33, p < .001). The results indicated that it was appropriate to use these two covariates in the MANCOVA test.
Order Effect. The experiment news story was placed with three filler news stories either as the second or fourth item. The two positions were randomly assigned and counterbalanced. T-tests were conducted to compare participant ratings on the items measuring self-assessed attention to, quality, and credibility of the news story presented in either order. No significant differences, meaning no order effect, were found. Data collected from the two order conditions were collapsed and treated as one sample in subsequent analyses.
Issue Relevance. In addition to randomly assigning subjects to groups, the study was designed to avoid any possible confounding sources that may contaminate the experimental treatment effects (Bausell, 1994), including the news stimulus that may induce a high degree of relevance or irrelevance from the participant. A five-point item measuring participant self-reported issue relevance (with one being "highly irrelevant" and five being "highly relevant") shows that most people considered the AIDS issue covered in the news story to be moderately relevant to them (M = 3.17, SD = 1.04 for Taiwan; M = 2.93, SD = 1.03 for Hong Kong).
Analyses on Dependent Variables
Interrelatedness of the Dependent Variables. Four dependent variables were under study in this experiment. To ascertain whether it was appropriate to use combined dependent variables for analysis in the MANCOVA test, Pearson correlation tests and Bartlett's test of sphericity were conducted to examine the interrelatedness of the dependent variables. Significant bivariate correlations were found between pairs of dependent variables, with r ranging from .22 to .57 (p < .001). The result of Bartlett's test of sphericity ([chi square] = 857.85, p < .001) confirmed the interrelatedness of the four dependent variables.
Overall, experiment participants (after controlling the two covariates) tended to express positive attitudes toward the following dependent variables: rights to work, mobility, and education (M = 4.05, SD = .63), willingness to interact (M = 3.58, SD = .89), and intention to advocate (M = 3.48, SD = .64). Their attitudes were, however, more reserved in terms of rights to marriage and birth-giving (M = 2.88, SD = .99).
The means of these attitudes toward PLWHA by experimental conditions such as transmission route, region, and gender are presented in Table 3.
Influence of Covariates. The MANCOVA results (see Table 4) also yielded significant associations between the combined dependent variables and two covariates: Wilks' [lambda] = .98, F(4, 948) = 14.59, p < .001, [[eta].sup.2] = .02 with fear of infection, and Wilks' [lambda] = .82, F(4, 948) = 53.53, p < .001, [[eta].sup.2] = .18 with attitudes toward homosexuality. Therefore, both H2 and H3 were supported.
The univariate analyses of the influence of fear of infection presented in Table 5 did not result in statistically significant differences in any dependent variable. Nevertheless, the difference in willingness to interact with PLWHA was close to the significance level set at [alpha] = .0125, F(1, 968) = 5.55, p = .019, [[eta].sup.2] = .01, showing that those who expressed less fear of infection tended to be more willing to interact with PLWHA.
As for the influence of attitudes toward homosexuality on the combined dependent variables, differences can be observed on all four dependent variables (see Table 5): rights to work, mobility, and education, F(1, 968) = 137.13, p < .001, [[eta].sup.2] = 13, rights to marriage and birth-giving, F(1, 968) = 47.07, p < .001, [[eta].sup.2] = .06, willingness to interact, F(1, 968) = 84.63, p < .001, [[eta].sup.2] = .12, and intention to advocate, F(1, 968) = 36.39,p < .001, [[eta].sup.2]= .10. Those participants who expressed more positive attitudes toward homosexuality were more likely to support PLWHA's rights to work, mobility, education, marriage and birth-giving, and more likely to interact with and advocate for them.
Effects of Independent Variables. Main effects of all three independent variables on the combined dependent variables were observed in the result of MANCOVA. Specific results of MANCOVA and subsequent univariate analyses for each independent variable were as follows.
News representation of the transmission route. Different conditions of transmission routes were found to have significant impact on the combined dependent variables as shown in Table 4, Wilks' [lambda] = .94, F(12, 2508) = 5.12, p < .001, [[eta].sup.2] = .02. Results of univariate analyses indicated that the difference could be attributed to two dependent variables (see Table 5): rights to marriage and birth-giving, F(3, 968) = 12.63, p < .001, [[eta].sup.2] = .05; and intention to advocate, F(3, 968) = 2.09, p < .01, [[eta].sup.2] = .02. Post-hoc tests using Sheffe's method were conducted to examine the difference patterns between groups. The results show that participants exposed to the MSM condition (M=3.29) held significantly more positive attitudes than those in the other three conditions (unspecificed M= 2.78; heterosexual M= 2.90; blood transfusion M= 2.88; all ps < .001) regarding rights to marriage and birth-giving. The former also held more positive attitudes (M=3.61) than their counterparts in the heterosexual (M=3.44, p < .01) and unspecified (M=3.37, p < .05) conditions regarding intention to advocate, and than those in the unspecified condition regarding willingness to interact (MSM M = 3.72; unspecificed M= 3.48, p < .05). H1, which tested the behavioral culpability model, was thus partially supported.
Gender. An effect of gender difference on the combined dependent variables was also observed, Wilks' [lambda] = .98, F(4, 948) = 4.43, p < .01, [[eta].sup.2] = .02 (see Table 4). H4, which predicted a gender difference between male and female attitudes toward PLWHA, was therefore supported. The results of univariate analyses (see Table 5) show that the effect was mainly due to the differences in rights to work, mobility, and education, F(1, 951) = 8.35, p < .01, [[eta].sup.2] = .01, and in willingness to interact, F(1, 951) = 11.33, p < .01, [[eta].sup.2] = .01. The difference in rights to marriage and birth-giving was close to the significance level set at a = .0125, F(1, 951) = 4.09, p = .024, [[eta].sup.2] = .01. In terms of group means, female participants revealed significantly more intention than their male counterparts to advocate for PLWHA. The former also tended to be more positive than the latter regarding willingness to interact, and rights to work, mobility, and education, although such differences were not statistically significant.
Place of residence. A significant difference between Taiwanese and Hong Kong participants was found in the combined dependent variables, Wilks' [lambda] = .91, F(4, 948) = 23.87, p < .001, [[eta].sup.2] = .09 (see Table 4). The results of univariate analyses (see Table 5) indicate that significant differences could be found in all dependent variables, rights to work, mobility, and education, F(1, 951) = 60.19, p < .001, [[eta].sup.2] = .06; rights to marriage and birth giving, F(1, 951) = 52.33, p < .001, [[eta].sup.2] = .05; willingness to interact, F(1, 951) = 21.96, p < .01, [[eta].sup.2] = .02, except in intention to advocate. Specifically, Hong Kong participants expressed greater support for rights to work, mobility, and education and rights to marriage and birth-giving and stronger willingness to interact than their Taiwanese counterparts. H5 was mostly supported.
Finally, we also found a significant effect of the two-way interaction between gender and place of residence, Wilks' [lambda] = .99, F(4, 948) = 2.55, p < .05, [[eta].sup.2] = .01 (see Table 4). Nevertheless, subsequent univariate analyses did not find statistically significant differences in any dependent variable.
Discussion and Conclusion
Heated debates have waged about whether news media representations of PLWHA and/or specific risk groups have been discriminative and have contributed to the formation of HIV/AIDS-related stigmas. The issue has been greatly complicated by the news media's paradoxical roles in surveillance and in covering deviance, audience pre-existing values and inherent differences, and the contextual influences on information processing. From the humanistic perspective, it is an urgent task to eliminate stigmas and discrimination associated with HIV/AIDS. To achieve this end, it will be necessary to discover the mechanism contributing to the formation of such stigmas and discrimination to understand how to tackle the problem. Therefore, this study has both practical and academic significance.
The results show that the participants generally tended to express positive attitudes toward the following three dependent variables: PLWHA's rights to work, mobility and education, willingness to interact with PLWHA, and intention to advocate for PLWHA. They were, however, more reserved in the other dependent variable measuring their perceived PLWHA's rights to marriage and birth-giving. It appears that individuals tended to be politically correct in terms of supporting PLWHA's basic human rights. However, when it came to such intimate relationships as marriage and birth-giving, they were more likely to take the social impact into consideration and thus less likely to show support. In other words, the less personal the issue, the more likely individuals were to express supportive attitudes toward PLWHA.
Furthermore, those participants who expressed less fear of infection and more positive attitudes toward homosexuality indicated more positive attitudes toward PLWHA's rights to work, mobility, and education, and stronger willingness to interact with PLWHA. In addition, participants' positive attitudes toward homosexuality were correlated with a stronger tendency to support PLWHA's rights to marriage and birth-giving and to advocate for PLWHA. In other words, symbolic attitudes appeared to be a better predictor of attitudes toward PLWHA in less abstract and distant domains, particularly in domains where more intimate relationships and public advocacy were specified. This again shows that attitudes toward PLWHA are more than health-related issues.
In terms of the effect of news representation of the transmission route, we found that those participants who read the AIDS news story emphasizing the MSM transmission route expressed more positive attitudes than their counterparts in the heterosexual-with-multiple-partners and unspecified conditions in all dependent variables except for intention to advocate for PLWHA. This could be due to the overly stigmatizing representation of HIV/AIDS in the news media, such as an emphasis on the MSM transmission route. Such bias tends to induce more sympathy with PLWHA from the individuals, especially from young adults who may have less conservative attitudes toward sex and sexuality-related issues. In addition, the findings show those exposed to the blood-transfusion condition were less likely to express support for PLWHA's rights to marriage and birth-giving in comparison to those in the MSM condition.
A possible explanation for such findings is that participants exposed to the MSM condition were less likely to associate homosexual contacts with the rights to marriage and birth-giving, especially the right to marriage is still illegal in both Taiwan and Hong Kong. Such rights may thus be considered less relevant to same-sex relationships. Therefore, participants in this condition were less likely to deny these rights to PLWHA. Conversely, those who were exposed to the blood transfusion condition may have felt a lack of control in preventing such a transmission. They were thus less supportive of PLWHA's rights to marriage and birth-giving--rights that involve intimate interactions. Blood transfusion may also remind them of perinatal transmission via blood transfusion, so they approved less of PLWHA's rights to marriage and birth-giving. As this study did not collect experiment participants' induced thoughts in processing the news story, these speculations need to be explored in future research.
In addition, the journalistic style of the news story employed in this experiment may have some implications for readers' perceptions of and attitudes toward HIV/AIDS-related issues. The stimulus story, presented in a human-interest format, featured an exemplar with a remorseful/painful source who disclosed his guilty feelings over passing the virus to his then deceased partner or wife. The impact may have induced more sympathy from the readers, particularly for socially disadvantaged groups (e.g., the homosexual) that are subject to over-stigmatization in real life. Future studies on the effect of news representation of the transmission route may want to take the journalistic style such as the use of exemplars and presentation tone into consideration when exploring the influence of behavioral culpability on individual attitudes toward HIV/AIDS-related issues.
As for the influence of gender differences, female participants were more likely to show positive attitudes toward PLWHA than their male counterparts. This echoes with the results of previous studies conducted on health care workers (e.g., Heaven, Connors, & Kellehear, 1990; Ladany & Stern, 1998; Rezler & Ten Haken, 1984; Stern, Norman, & Komm, 1993). It implies that more campaign efforts need to target at men in order to increase their positive attitudes toward PLWHA.
In terms of places of residence, the Hong Kong participants, in comparison to the Taiwanese participants, expressed a higher degree of willingness to interact with PLWHA and a stronger likelihood to support PLWHA's rights to marriage and birth-giving as well as to work, mobility, and education. Such findings reinforce the substantive efforts made necessary in the legal and legislative domains in addressing the elimination of stigma and discrimination against PLWHA, as can be seen in Hong Kong's HIV/AIDS related policies and campaigns. Again, the foregoing speculations should be further investigated in future research.
Empirical results comparing the explanatory power of the behavioral and cultural conflict models indicate that both had a certain degree of influence on individual attitudes toward PLWHA's rights. The cultural conflict model seemed to have stronger explanatory power. This finding shows that individual attitudes toward PLWHA and their entitled rights were not simply based on disease-related issues. The stereotypes projected on the high-risk groups by the individual made a greater difference. The explanatory power of the behavioral model, as mediated by news representation of the transmission route, appeared less influential than the pre-existing beliefs/values as evidenced by the effects of the cultural conflicts model. Nevertheless, reading HIV/AIDS stories describing various HIV transmission routes did make a difference to individual attitudes toward PLWHA when symbolic and instrumental attitudes were controlled. If one-shot reading of a manipulated news story can induce PLWHA-related attitudes different from those generated by reading other manipulated news stories, it implies that long-term news representations of the HIV/AIDS-related issues may have cultivated much stronger attitudes supporting or opposing PLWHA's rights, depending on how the issues were accumulatively framed. Thus, the role of news media in representing HIV/AIDS and PLWHA cannot be overlooked.
As an exploratory study with a cross-cultural message design, some limitations of the measurement need to be noted when making further generalizations. Although both statistical and theoretical considerations were taken to construct scales measuring variables, the relatively low scale reliability coefficients for the dependent variable rights to marriage and birth-giving (Cronbach's [alpha]= .48) and the covariate fear of infection (Cronbach's [alpha]= .56) warrant cautions in conceptualizing the dimensionality of the variables. Some explanations can be acknowledged here. To begin with, marriage and birth-giving are conceptually related according to traditional Chinese values. They may however be regarded as separate and inconsequential matters nowadays, especially for young adults who have more liberal social views. Moreover, participants' responses to the birth-giving item may have been influenced by their beliefs about the injustice of transmitting HIV to innocent babies who have no choice in the matter. On the other hand, participants may have presumed that people who marry someone with HIV/AIDS are making an informed decision and preventing this would be infringing on people's rights. Therefore, their responses to these two items may have based on different rationales. Similarly, worrying about self infection of HIV/AIDS and the belief that HIV/AIDS is so pervasively dangerous that few people can be exempted may involve more than one conceptual dimension. The former seems to have more to do with self perception, whereas the latter concerns perception of others. Future research could elaborate the conceptual differences about fear of infection between the first-person and third-person perceptions.
To conclude, findings of this experiment may be limited in providing direct suggestions for the improvement of HIV/AIDS intervention or journalistic practice in representing PLWHA. However, they do point to some promising directions for future studies. It is important to note that the topic of PLWHA's rights will become even more complex as new moral considerations associated with it progress with development of the epidemic and bio-technologies. Therefore, future research should be more comprehensive when conceptualizing and operationalizing attitudes toward PLWHA's rights by including such factors related to social, political, economic, and cultural realities. Attitudes induced by gender differences, religious beliefs, homophobic tendency, media influences, scientific and medical advances, and perceptions of PLWHA's rights to medical care should all be taken into account.
Albert, E. (1986). Illness and deviance: The response of the press to AIDS. In D. A. Feldman & T. M. Johnson (Eds.), The social dimensions of AIDS: Method and theory (pp. 163-178,). New York: Praeger.
Archer, D. (1985). Social deviance. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (pp. 743-804). (3rd ed., Vol. 2). New York: Random House.
Bausell, R. B. (1994). Conducting meaningful experiments: 40 steps to becoming a scientist. Thousand Oaks, CA: Sage.
Beharrell, P. (1998). News variations. In D. Miller, J. Kitzinger, K. Williams & P. Beharrell (Eds.), The circuit of mass communication (pp. 46-67). Thousand Oaks, CA: Sage.
Blendon, R. J., & Donelan, K. (1988). Discrimination against people with AIDS: The public's perspective. New England Journal of Medicine, 319, 1022-1026.
Brecher, E. M. (1988). Straight sex, AIDS, and the mixed-up press. Columbia Journalism Review (March/April) , 46-50.
Carmines, E. G., & Zeller, R. A. (1973). Reliability and validity assessment. Beverly Hills, CA: Sage.
Centers for Disease Control, Taiwan. (2007, December 7).Taiwan government's responses to HIV/AIDS. Retrieved May 5, 2009 from http://www. cdc.gov.tw/content.asp?CuItem=7424.
Center for Disease Control, Taiwan (2008, November 5). Wan-wan urges public to express love, concern for HIV/AIDS patients (press release). Retrieved July 5, 2009 from http://www.cdc.gov.tw/ct.asp?xItem =20109&ctNode=960&mp=5.
Centers for Disease Control, Taiwan. (2010, March 31). HIV/AIDS (1984-2010/3/31). Retrieved April 21, 2010 from http://www.cdc.gov.tw/public/Data/0481 4534571.doc.
Chan, M. K., & Lee, S. (2004). Can the low HIV prevalence in Hong Kong be maintained? AIDS Education and Prevention, 16 (Supplement A), 18-26.
Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78, 98-104.
Crocker, J., Major, B., & Steele, C. (1998). Social stigma. In D. T. Gilbert, S. T.
D'Angelo, R. J., McGuire, J. M., Abbott, D. W., & Sheridan, K. (1998). Homophobia and perceptions of people with AIDS. Journal of Applied Social Psychology, 28, 157-170.
Davis, C., Noel, M. B., Chan, S. F., & Wing, L. S. (1998). Knowledge, attitudes and behaviors related to HIV and AIDS among Chinese adolescents in Hong Kong. Journal of Adolescence, 21, 657-665.
Department of Health, Hong Kong (2010, March 2). Hong Kong STD/AIDS update: A quarterly surveillance report. Retrieved April 24, 2010 from http://www.info.gov.hk/aids/archives/backissuestd/a ids09q4.pdf.
Devine, P. G., Plant , E. A., & Harrison, K. (1999). The problem of "us" versus "them" and AIDS stigma. American Behavioral Scientist, 42, 1212-1228.
Dodds, C. (2006). HIV-related stigma in England: Experiences of gay men and heterosexual African migrants living with HIV. Journal of Community & Applied Social Psychology, 16, 472-480.
Edgar, T., Noar, S. M., & Freimuth, V. S. (Eds.) (2008). Communication perspective on HIV/AIDS for the 21st century. New York: Lawrence Erlbaum.
Freimuth, V. S., Greenberg, R. H., DeWitt, J., & Romano, R. M. (1984). Covering cancer: Newspapers and the public interest. Journal of Communication, 34, 62-73.
Fuller, L. K. (Ed.) (2003). Media-mediated AIDS. Chesskill, NJ: Hampton.
Grandall, S. C., & Glor, J. (1997). AIDS-related stigmatization: Instrumental and symbolic attitudes. Journal of Social Psychology, 27, 95-123.
Heaven, P. C., Connors, J., & Kellehear, A. (1990). Structure and demographic correlates of attitudes towards AIDS sufferers. Journal of Psychology, 124, 245-52.
Herek, G. M. (1987). Can functions be measured? A new perspective on the functional approach to attitudes. Social Psychology Quarterly, 50, 285-303.
Herek, G. M., & Capitanio, J. P. (1993). Public reaction to AIDS in the United States: A second decade of stigma. American Journal of Public Health, 83, 574-577.
Herek, G. M., & Glunt, E. K. (1991). AIDS-related attitudes in the United States: A preliminary conceptualization. Journal of Sex Research, 28, 99123.
Herek, G.. M., Widaman, K. F., & Capitanio, J. P. (2005). When sex equals AIDS: Symbolic stigma and heterosexual adults' inaccurate beliefs about sexual transmission of AIDS. Social Problems, 52(1), 15-37.
Hong Kong Advisory Council on AIDS. (2010). Recommended HIV/AIDS strategies for Hong Kong: 2007 to 2011. Retrieved April 23, 2010 from http://www.info.gov.hk/aids/pdf/g178.pdf.
Hsu, M. (2004, August). HIV/AIDS media campaigns and their effectiveness in Taiwan. Paper presented at the 2004 bi-annual conference of International Association for Media and Communication Research, Porto Alegre, Brazil.
Hsu, M. (2001). AIDS and media. Taipei: Chuliu.
Hsu, M., Lin, W., & Wu, T. (2004). Representations of "us" and "others" in the AIDS news discourse: A Taiwanese experience. In E. Micollier (Ed.), Sexual cultures in East Asia: The social construction of sexuality and sexual risk in a time of AIDS (pp. 183-222). UK: Routledge-Curzon.
Hurlburt, R. T. (2006). Comprehending behavioral statistics. Belmont, CA: Thomson Wadsworth.
Johnson, D., Flora, J. A., & Rimal, R. N. (1997). HIV/AIDS public service announcements around the world: A descriptive analysis. Journal of Health Communication, 2, 223-234.
Kaplan, M. E., & van den Worm, Y. (1993). The relationship between South African adolescents' knowledge and fear of AIDS and their attitudes toward people who have AIDS. Journal of Social Psychology, 133, 581-583.
Kline, P. (1999). The handbook of psychological testing. London: Routledge.
Kohi, T. W., Makoae, L., Chirwa, M., Holzemer, W. L., Phetlhu, D. R., Uys, L., Naidoo, J., Dlamini, P. S., & Greeff, M. (2006). HIV and AIDS stigma violates human rights in five African countries. Nursing Ethics, 13(4), 404-415.
Ladany, N., & Stern, M. (1998). Medical students' perceptions of and willingness to treat patients who are HIV infected. Journal of Applied Social Psychology, 28, 1031-1050.
Lau, H. T. F., & Tsui, H. Y. (2002). Surveillance of HIV/AIDS related attitudes and perceptions among the general public in Hong Kong from 1994 to 2000. AIDS Education and Prevention, 14, 419-431.
Lau, J. T. F., Tsui, H. Y., & Chan, K. (2005). Reducing discriminatory attitudes toward people living with HIV/AIDS (PLWHA) in Hong Kong: An intervention study using an integrated knowledge-based PLWHA participation and cognitive approach. AIDS Care, 17, 85-101.
Lupton, D. A. (1992). From complacency to panic: AIDS and heterosexuals in the Australian press, July 1986 to June 1988. Health Education Research, 7, 9-20.
Rezler, A., & Ten Haken, J. (1984). Affect and research in medical education. Medical Education, 18, 331-338.
Rogers, E. M., Dearing, J., & Chang, S. (1991). AIDS in the 1980s: The agenda-setting process for a public issue. Journalism Monographs, 126, 1-47.
Schmitt, N. (1996), Uses and abuses of coefficient alpha. Psychological Assessment, 8(4), 350-353.
Schulte, A. (2002). Consensus versus disagreement in disease-related stigma: A comparison of reactions to AIDS and cancer patients. Sociological Perspectives, 45, 81-104.
Simpkins, J. D., & Brenner, D. J. (1984). Mass media communication and health. In B. Dervin & M. J. Voigt (Eds.), Progress in communication sciences (pp.275-297). Norwood, NJ: Ablex.
Stern, M., Norman, S., & Komm, C. (1993). Medical students' differential use of coping strategies as a function of stressor type, year of training, and gender. Behavioral Medicine, 18, 173-180.
Traquina, N. (1996, August). Portuguese journalism and HIV/AIDS: A case study in news. Paper presented at the annual conference of Association for Education in Journalism and Mass Communication, Anaheim, CA.
Triplet, R. G., & Sugarman, D.B. (1987). Reactions to AIDS victims: Ambiguity breeds contempt. Personality and Social Psychology Bulletin, 13, 265-74.
Twu, S., Huang, Y., Lai, A., & Ou, N. (2004). Update and projection on HIV/AIDS in Taiwan. AIDS Education and Prevention, 16(supplement A), 53-63.
UNAIDS (2003). World AIDS Campaign 2002-03. Retrieved January 23, 2004 from http:// www.unaids.org/en/events/campaigns.asp.
United Nations (2000, December). The universal declaration of human rights. Retrieved July 1, 2005 from http://www.un.org/Overview/rightrs.html.
Virtual AIDS Office of Hong Kong (2005). About AIDS-AIDS and STD information. Rights of people with HIV/AIDS. Retrieved July 1, 2005 from http://www.info.gov.hk/aids/english/hotline/03_6.ht m.
Virtual AIDS Office of Hong Kong (2010). HIV/AIDS situation in 2009 reviewed. Retrieved April 24, 2010 from http://www.info.gov.hk/aids/English/press/2010/100 0302.htm.
Wallack, L. (1990). Mass media and health promotion: Promise, problem, and challenge. In C. Atkin & L. Wallack (Eds.), Mass communication and public health (pp. 41-50). Newbury Park, CA: Sage.
Weiss, C. H. (1974). What America's leaders read. Public Opinion Quarterly, 38, 1-21.
Zillmann, D., & Brosius, H. (2000). Exemplification in communication: The influence of case reports on the perception of issues. Mahwah, NJ: Lawrence Erlbaum.
(1.) Schmitt (1996) argued that there is no sacred level of acceptable or unacceptable level of alpha. In some cases, measures with (by conventional standards) low levels of alpha may still be quite useful. Carmine and Zeller (1973) also indicated that reliability coefficients based on factor analysis are not as restrictive as those methods for estimating reliability that assume parallel items. In addition, alpha values below .7 can be expected when dealing with psychological constructs (Kline,
1999) such as those items measured in the present study. The value of alpha is positively related to the number of items on the scale as well (Cortina, 1993). As there are only two items in the "rights to marriage and birth-giving" index, we considered such a factor tentatively acceptable in the study.
Mei-Ling Hsu, National Chengchi University
Hao-Chieh Chang, Hong Kong Baptist University
Dr. Mei-Ling Hsu
Department of Journalism
National Chengchi University
Wenshan, Taipei, Taiwan 11605
Table 1. A comparison of HIV/AIDS epidemics in Taiwan and Hong Kong Category Taiwan Total population 23 million Total number of PLWHA 18,687 First reported case December, 1984 Major transmission routes Sexual contact (MSM: 41.65%; heterosexual: 22.59%), IDU: 33.98% Major infected sex Male (91.6%) Major infected age group 30-39 (32.8%) Local or Chinese ethnic PLWHA 98.8% (local Taiwanese) Category Hong Kong Total population 6.8 million Total number of PLWHA 4,443 First reported case November, 1984 Major transmission routes Sexual contact (MSM: 44.83%; heterosexual: 44.83%), IDU: 6.17% Major infected sex Male (80.46%) Major infected age group 30-39 (37.59%) Local or Chinese ethnic PLWHA 66.49% (Chinese ethnicity) Sources: HIV/AIDS statistics, by Center for Disease Control, Taiwan (2010, March 31); HIV/AIDS statistics, by Department of Health, Hong Kong (2010, March 2). Table 2. Measuring items for dependent variables and covariates Attitudinal Measuring items Item variables loadings on the variable PLWHA's rights 1. PLWHA should be allowed to .67 to work, stay in the school dormitory. mobility, and 2. Universities should not admit .67 education PLWHA. * 3. PLWHA should be allowed at .73 public places. 4. Schools should not deprive a .66 student of the right to attend classes even though he/she is diagnosed of HIV/AIDS. 5. PLWHA should be quarantined. * .68 6. Employers should fire those .63 employees who are diagnosed of HIV/AIDS. * 7. Expatriates who are diagnosed .45 of HIV/AIDS should be deported. * PLWHA's rights 1. Restrictions should be posed ,67 to marriage and on PLWHA's reproduction rights birth-giving to avoid perinatal transmission through pregnancy. 2. PLWHA should not get married. .58 Willingness to 1. I can be with my friends who .82 interact with are diagnosed of HIV/ AIDS PLWHA as I normally do. 2. I am willing to make friends .78 with PLWHA. 3. I will attend classes with .77 fellow students who are diagnosed of HIV/AIDS. Intention to 1. I will attend to his/her needs .52 advocate for if someone among my peers is PLWHA diagnosed of HIV/AIDS. 2. I will speak out for PLWHA in .59 public to correct misconceptions about them. 3. I will volunteer in a HIV/AIDS .78 intervention campaign. 4. I will spend time with PLWHA to .80 learn more about them. 5. I will petition to promote .76 more caring for PLWHA. Fear of 1. I am not worried about .67 infection contracting AIDS. 2. HIV/AIDS has posed a tremendous .71 threat to the local community so that few people can be exempted. * Attitudes toward 1. I find homosexuality .81 homosexuality unacceptable whether it is inherent or socialized. * 2. Homosexual contacts are as .78 normal as heterosexual ones. 3. I am afraid of becoming gay .77 while being with homosexuals. * 4. I am particularly afraid of .72 being harassed by homosexuals. * 5. There is nothing to be ashamed .71 of being homosexual. 6. The term homosexuality gets .67 on my nerves. * 7. I do not care about my friend's .67 sexual orientation. 8. I am afraid of being taken for .63 homosexual when being with gay friends. * 9. I am okay to room with one .60 whose sexual orientation is different from mine. 10. I will shun a homosexual .58 relative. * 11. Those who contract HIV/AIDS .58 via man-to-man sex deserve the disease. * Attitudinal Scale reliability variables (Cronbach's [alpha]) PLWHA's rights 1. .80 to work, 2. mobility, and 3. education 4. 5. 6. 7. PLWHA's rights 1. .48 to marriage and 2. birth-giving Willingness to 1. .82 interact with 2. PLWHA 3. Intention to 1. .78 advocate for 2. PLWHA 3. 4. 5. Fear of 1. .56 infection 2. Attitudes toward 1. .88 homosexuality 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. * These items were reversely recoded to observe the unidirectionality of the variable. Table 3. Means of various attitudes toward PLWHA by experimental conditions Region Taiwan (N = 480) Attitudes toward PLWHA Male Female Rights to work, mobility, and education Unspecified 3.76 3.91 (0.74) (0.58) Heterosexual 3.80 3.99 (0.63) (0.65) MSM 3.99 4.06 (0.71) (0.55) Blood transfusion 4.00 3.87 (0.65) (0.70) Rights to marriage and birth-giving Unspecified 2.65 2.62 (0.97) (0.78) Heterosexual 2.46 2. 85 (0.75) (0.87) MSM 3.14 3.07 (0.96) (0.81) Blood transfusion 2.77 2.72 (0.82) (1.00) Willingness to interact with PLWHA Unspecified 3.47 3.39 (1.06) (0.74) Heterosexual 3.42 3.41 (0.81) (0.96) MSM 3.74 3.51 (0.75) (0.92) Blood transfusion 3.47 3.46 (0.83) (0.85) Intention to advocate for PLWHA Unspecified 3.30 3.54 (0.64) (0.57) Heterosexual 3.35 3.59 (0.62) (0.66) MSM 3.49 3.71 (0.65) (0.51) Blood transfusion 3.54 3.45 (0.64) (0.59) Region Hong Kong (N = 504) Attitudes toward PLWHA Male Female Rights to work, mobility, and education Unspecified 4.12 4.23 (0.58) (0.50) Heterosexual 4.20 3.99 (0.54) (0.64) MSM 4.16 4.19 (0.72) (0.56) Blood transfusion 4.19 4.23 (0.55) (0.55) Rights to marriage and birth-giving Unspecified 2.83 2.97 (1.03) (0.86) Heterosexual 3.16 3.01 (1.08) (0.96) MSM 3.57 3.44 (0.89) (0.81) Blood transfusion 3.03 3.00 (1.02) (1.06) Willingness to interact with PLWHA Unspecified 3.38 3.65 (0.92) (0.85) Heterosexual 3.65 3.64 (0.86) (0.97) MSM 3.74 3.86 (0.76) (0.84) Blood transfusion 3.69 3.73 (0.81) (0.86) Intention to advocate for PLWHA Unspecified 3.26 3.34 (0.60) (0.66) Heterosexual 3.28 3.45 (0.66) (0.81) MSM 3.52 3.67 (0.61) (0.60) Blood transfusion 3.50 3.47 (0.50) (0.62) Notes: (1) Figures in the parenthesis ( ) indicate standard deviations. (2) Typical Likert scales were employed to measure participant attitudes, with one point meaning "strongly disagree" and five meaning "strongly agree." Table 4. MANCOVA results of attitudes toward PLWHA Wilks' F Hypothesis [lambda] df Covariates Attitudes towards .82 53.53 4 homosexuality Fear of infection .98 14.59 4 Main effects Transmission route .94 5.12 12 Gender .98 4.43 4 Place of residence .91 23.87 4 Two-way interactions Transmission route X gender .99 1.22 12 Transmission route X place .98 1.24 12 of residence Gender X place of residence .99 2.55 4 Three-way interaction Transmission route X gender X .99 1.18 12 place of residence Error df p .[eta]. sup.2] Covariates Attitudes towards 948 .000 .18 homosexuality Fear of infection 948 .001 .02 Main effects Transmission route 2508 .000 .02 Gender 948 .001 .02 Place of residence 948 .000 .09 Two-way interactions Transmission route X gender 2508 .266 .01 Transmission route X place 2508 .250 .01 of residence Gender X place of residence 948 .038 .01 Three-way interaction Transmission route X gender X 2508 .288 .01 place of residence Table 5. Univariate results of attitudes toward PLWHA Sources Dependent variables df [Mean. sup.2] Attitudes Rights to work, mobility, and 1 45.08 toward education homosexuality Rights to marriage and birth-giving 1 47.07 Willingness to interact with 1 84.63 PLWHA 1 36.39 Intention to advocate for PLWHA Fear of Rights to work, mobility, and 1 .92 infection education Rights to marriage and birth-giving 1 2.54 Willingness to interact with 1 3.67 PLWHA 1 .65 Intention to advocate for PLWHA Transmission Rights to work, mobility, and 3 .58 route education Rights to marriage and birth-giving 3 12.63 Willingness to interact with 3 2.30 PLWHA 3 2.09 Intention to advocate for PLWHA Gender Rights to work, mobility, and 1 2.75 education Rights to marriage and birth-giving 1 4.09 Willingness to interact with 1 7.50 PLWHA 1 .01 Intention to advocate for PLWHA Place of Rights to work, mobility, and 1 19.79 residence education Rights to marriage and birth-giving 1 41.74 Willingness to interact with 1 14.55 PLWHA 1 .11 Intention to advocate for PLWHA Sources Dependent variables F p Attitudes Rights to work, mobility, and 137.13 .000 toward education homosexuality Rights to marriage and birth-giving 59.02 .000 Willingness to interact with 127.79 .000 PLWHA 102.03 .000 Intention to advocate for PLWHA Fear of Rights to work, mobility, and 2.79 .096 infection education Rights to marriage and birth-giving 3.19 .074 Willingness to interact with 5.55 .019 PLWHA 1.83 .176 Intention to advocate for PLWHA Transmission Rights to work, mobility, and 1.76 .154 route education Rights to marriage and birth-giving 15.83 .000 Willingness to interact with 3.47 .016 PLWHA 5.87 .001 Intention to advocate for PLWHA Gender Rights to work, mobility, and 8.35 .004 education Rights to marriage and birth-giving 5.13 .024 Willingness to interact with 11.33 .001 PLWHA .04 .841 Intention to advocate for PLWHA Place of Rights to work, mobility, and 60.19 .000 residence education Rights to marriage and birth-giving 52.33 .000 Willingness to interact with 21.96 .000 PLWHA .30 .587 Intention to advocate for PLWHA Sources Dependent variables [[eta]. sup.2] Attitudes Rights to work, mobility, and .13 toward education homosexuality Rights to marriage and birth-giving .06 Willingness to interact with .12 PLWHA .10 Intention to advocate for PLWHA Fear of Rights to work, mobility, and .00 infection education Rights to marriage and birth-giving .00 Willingness to interact with .01 PLWHA .00 Intention to advocate for PLWHA Transmission Rights to work, mobility, and .01 route education Rights to marriage and birth-giving .05 Willingness to interact with .01 PLWHA .02 Intention to advocate for PLWHA Gender Rights to work, mobility, and .01 education Rights to marriage and birth-giving .01 Willingness to interact with .01 PLWHA .00 Intention to advocate for PLWHA Place of Rights to work, mobility, and .06 residence education Rights to marriage and birth-giving .05 Willingness to interact with .02 PLWHA .00 Intention to advocate for PLWHA