A Second Decade of Stigma:
Public Reactions to AIDS in the United States, 1990-91

Gregory M. Herek, Ph.D. & John P. Capitanio, Ph.D.

Department of Psychology
University of California at Davis

Copyright © 1992 by Gregory M. Herek, Ph.D. All rights reserved

Note: A short version of this paper was published in the American Journal of Public Health, 1993, 83, 574-577

ABSTRACT

Background. This study measured the pervasiveness of stigmatizing attitudes and beliefs concerning AIDS among the American public. Because African Americans have been disproportionately affected by AIDS, stigma also was assessed in a sample of Black Americans.

Methods. In 1990-91, telephone interviews were conducted with a general sample of 538 U.S. adults, and a separate sample of 607 African Americans. Respondents were asked about their feelings toward persons with AIDS (PWAs), support for punitive policies, likelihood of avoiding PWAs, and beliefs about HIV transmission.

Results. Most respondents manifested some stigma. Only 16.5% of Blacks and 22% of Whites did not give any stigmatizing responses. African Americans were more likely than Whites to overestimate the risk of infection through casual contact, but were less likely to hold negative personal feelings toward PWAs. Overall, females were less likely than males to stigmatize PWAs on measures pertaining to punitive policies and avoidant behaviors.

Conclusions. AIDS stigma persists as a problem in the United States. Stigma among African Americans appears to focus on AIDS as a disease that threatens the Black community, whereas Whites' stigma appears to reflect attitudes toward the social groups principally affected by the epidemic. Stigma reduction should be a central goal of AIDS educational efforts

BACKGROUND

Throughout the 1980s, a second epidemic shadowed AIDS in the United States. Many people infected with HIV were socially isolated, fired from their jobs, driven from their homes, and even physically attacked.¹ AIDS-related stigma also posed threats to the physical and psychological well-being of those simply perceived to be at risk. Members of the gay and lesbian community, for example, appeared increasingly to be targets of hate crimes, many of which included references by the perpetrators to AIDS.² In addition, AIDS-related stigma affected public support for government policies, and governmental support for AIDS-education programs.³ And it affected the willingness of individuals and entire communities at risk to acknowledge AIDS as a problem and to initiate prevention programs.⁴

As we enter the second decade of AIDS, the temptation is great to assume that the epidemic of stigma has ended. The federal Americans with Disabilities Act now protects people with AIDS (PWAs) from discrimination. AIDS has become a principal focus for charitable events and programs. And public compassion for PWAs seems to have increased as a result of public disclosures by influential people, such as Earvin "Magic" Johnson and Rock Hudson, that they have AIDS or are infected with HIV.

Despite these developments, other signs suggest that AIDS-related stigma persists. One of the major candidates in the 1992 presidential primaries, for example, asserted that AIDS is "nature's form of retribution" against homosexuals.⁷ Some medical professionals avoid treating patients with AIDS.⁸ And evidence for stigma continues to emerge from survey research and anecdotal reports.^9,10

The research described in the present paper was designed to measure the pervasiveness of stigmatizing attitudes and beliefs concerning AIDS among the American public as the epidemic's second decade began. Through telephone interviews with a probability sample of U.S. adults, we assessed the extent of stigma in a variety of manifestations: negative feelings toward PWAs, beliefs that they deserved their illness, support for punitive AIDS policies, and desires to avoid contact with PWAs. We also assessed misconceptions about HIV transmission, which may contribute to stigma.⁹ This included assessment of perceptions that casual social contact can transmit HIV and beliefs that injecting drug use or male-male sexual intercourse in themselves cause AIDS (rather than acting as a route of transmission for HIV).

In addition to sampling the views of the public generally, the research assessed AIDS-related attitudes and beliefs in a separate sample of African Americans. The Black community in the United States has ample reason to be alarmed about AIDS. In 1991, the rate of AIDS cases per 100,000 population among Blacks was 49.2 – compared to 11.7 among Whites and 31.4 among Hispanics.¹¹ Although Blacks represent only 12% of the U.S. population, they accounted for more than one-fourth of adult males with AIDS in the United States through 1991, and more than half of adult females and children with AIDS.¹¹

Concern about the disproportionate incidence of AIDS in their community could well translate into a high level of stigma among African Americans, which could further impede prevention and treatment efforts.⁴ Alternatively, the Black community's historical experience with prejudice and discrimination might mitigate its willingness to stigmatize people with AIDS. Studying AIDS-related stigma among African Americans, therefore, can help to improve intervention programs in the Black community and can contribute to a better general understanding of public reactions to AIDS

METHODS

General adult sample. A general adult sample was drawn from the population of all English-speaking adults (at least 18 years of age) residing in households with telephones within the 48 contiguous states. Ten-digit telephone numbers were generated using a stratified two-phase procedure for random-digit dialing.¹² First, area codes and prefix combinations on the Bell Communications Research tape were ordered geographically, and a large first-phase sample was selected with systematic random sampling. Four-digit random numbers were appended to the selected area code-prefix combinations to generate 10-digit telephone numbers, which were then compared to numbers on a tape created by Donnelly Marketing Services. The Donnelly tape indicates how many listed residential telephone numbers occur in each series of 100, organized according to the first eight digits. Two strata were then created. Stratum 1 comprised numbers whose first eight digits included at least one listed residential telephone number. Stratum 2 contained numbers for which no corresponding residential listings were found on the Donnelly tape. From the stratified pool of first-phase selections, a second phase was drawn by disproportionately sampling at the ratio of 18:1 for Stratum 1:Stratum 2. This method resulted in the second phase sample in which 48.7% (768/1578) of the selected telephone numbers were found to be households. Of the 768 households, 653 (85.0%) were enumerated. Of these, interviews were completed with 538 (82.4%), yielding a response rate (enumeration rate X completion rate) of 70.1%.

African American sample. The African American sample was selected using telephone numbers purchased from Survey Sampling, Inc. (Fairfield, CT). The list was based on census tracts where the density of Black households is 30% or higher (according to 1980 census data, 13.7% of all U.S. census tracts fit this description). Telephone numbers were taken from telephone directory listings and, in 21 states, were supplemented by motor vehicle registration data. This approach excluded Blacks living in untracted areas (e.g., very rural settings) as well as those living in neighborhoods with fewer than 30% African American households.

Eligibility criteria were that the respondent be a Black, English-speaking household resident at least 18 years of age. Of the 1900 telephone numbers in the sample list, 1523 (80.2%) were found to be residential households. Of these, 1343 (88.2%) were enumerated. Excluding non-Black households left 794 eligible homes, from which 607 interviews (76.4%) were completed. Because one goal of our project is to monitor reactions to AIDS among Black Californians, this group was oversampled, representing 263 of the 607 completed interviews. The response rate for the African American sample was 67.4%.

Procedures

Interviews were conducted by the staff of the Survey Research Center at the University of California at Berkeley between 12 September 1990 and 13 February 1991, using their computer-assisted telephone interviewing (CATI) system. No limit was set on the number of recontact attempts. Upon reaching an adult in the household, the interviewer enumerated the first name and race of each person 18 years or older living in the household. Based on this information, one respondent was selected randomly and, if that person was available, the interview began. If the target respondent was unavailable, the interviewer established a later time for recontact. Once the target respondent was identified, most interviews (62.4%) were completed within one or two attempts. Twenty-four respondents, however, required more than seven attempts before the interview was successfully completed. The maximum number of attempts before completing an interview was 19. Chi-square analyses revealed no consistent response differences according to the number of contact attempts for either sample. The mean duration of the interview was 39 minutes

Measures

We assessed four different manifestations of stigma: negative feelings toward persons with AIDS, support for coercive AIDS-related policies, blame for persons with AIDS, and intentions to avoid a person with AIDS in various situations. We also assessed beliefs about HIV transmission through casual contact and beliefs that male homosexual behavior or injecting drug use in themselves cause AIDS. The items are explained briefly here. The complete text of the items is available in the Appendix.

Negative feelings toward people with AIDS. Respondents were asked to rate the extent to which they felt angry at PWAs, afraid of them, and disgusted by them. Four response alternatives were provided (e.g., very angry, somewhat, a little, not at all angry).

Support for punitive policies. Respondents were asked how much they agreed or disagreed that "people with AIDS should be legally separated from others to protect the public health" and that "the names of people with AIDS should be made public so others can avoid them."⁹ Four response alternatives were provided (agree strongly, agree somewhat, disagree somewhat, disagree strongly).

Blame for persons with AIDS. Respondents were asked whether they agreed or disagreed that "people who got AIDS through sex or drug use have gotten what they deserve."⁹ As with AIDS policy attitudes, four response alternatives were provided ranging from agree strongly to disagree strongly.

Avoidant behaviors. Respondents were asked to predict their own behavior in each of four different situations involving potential contact with a person with AIDS. The situations were (1) having a close friend or relative who developed AIDS; (2) having a child attending a school where another student was known to have AIDS; (3) working in an office where a male coworker developed AIDS; and (4) finding out that the owner of a small neighborhood grocery store had AIDS. For each situation, respondents were offered a variety of response alternatives that represented an avoidant response (e.g., not helping to care for the sick friend, avoiding contact with the coworker) or a supportive response (e.g., caring for the friend, helping the coworker or treating him the same as always).

Beliefs about HIV transmission through casual contact. Respondents indicated their belief about the likelihood "that a person could get AIDS or AIDS virus infection" through five different routes: (1) kissing on the cheek, (2) sharing a drinking glass, (3) using public toilets, (4) being coughed on, and (5) insect bites. Five response alternatives were provided (very likely, somewhat likely, somewhat unlikely, very unlikely, and it is impossible to get AIDS from this activity).

Missing Data

Beliefs about homosexuality and drug use. To assess the belief that male homosexual behavior or injecting drug use in themselves cause AIDS (even in the absence of HIV), respondents were asked to assess the likelihood that an individual could contract AIDS in each of three hypothetical situations. Two of the scenarios involved male homosexual contact, and the third involved injecting drug use. In none of the situations could HIV possibly be transmitted. The scenarios were: (1) two healthy homosexual men, neither of whom is infected with the AIDS virus, have sexual intercourse using condoms; (2) the same two uninfected men have sexual intercourse, but they do not use condoms; (3) someone uses drugs intravenously, but does not share needles and is not a homosexual. In each case, respondents were asked to rate the chances that the person(s) described in the scenario would become infected, using four response alternatives (almost sure to become infected, has a fairly strong chance, very little chance, no chance)

  RESULTS

Sample characteristics. Of the 538 completed interviews in the general adult sample, 247 respondents (45.9%) were male and 291 (54.1%) were female. Racially, the sample included 436 Whites (81%), 56 Blacks (10.4%), 27 Hispanics (5%), 15 Asians (2.8%), and 4 who labeled themselves as "Other" (<1%). The mean age was 43.8 years (SD = 15.97); median annual household income was between $30,000 and $40,000; and the median level of educational attainment was "some college." Slightly more than one-third of the respondents (35.3%) labeled themselves Democrats; 31.6% were Republicans; 24.5% were Independents. Approximately 25% knew someone with AIDS or HIV.

Levels of stigma in the general adult sample. AIDS-related stigma appears to be manifested by a significant minority of the American public. As shown in Table 1, more than one-fourth of the respondents in the general sample felt very/somewhat disgusted or angry toward PWAs, and more than one-third were afraid of them. Roughly one-third agreed that people with AIDS should be quarantined and that their names should be made public. One-fifth agreed that PWAs deserved their illness. Although the majority of respondents predicted that they would be supportive of people with AIDS in various situations, one in eight would not help to care for a friend with AIDS and nearly one-half would avoid shopping at a neighborhood grocery store where the owner had AIDS.

Table 1:

Table 2:

Table 2 describes some of the beliefs that may contribute to stigma. Roughly half of the respondents believed that HIV is likely or only somewhat unlikely to be transmitted through insect bites, sharing a drinking glass, or being coughed on. Slightly fewer respondents believed that HIV is spread through public toilets, and approximately one-fifth believed HIV could be transmitted through a kiss on the cheek. Many respondents appeared to equate male homosexual behavior or drug use with HIV transmission, even in situations where such transmission would be impossible. Almost one-half believed that a healthy, uninfected man has at least a fairly strong chance of becoming infected through sexual intercourse with another uninfected man. One-fifth of the sample believed that HIV could be transmitted between uninfected men even if they used condoms, and one in seven believed that a drug user could become infected by injecting drugs without sharing needles or engaging in homosexual behavior.

African American Sample

Sample characteristics. Of the 607 interviews completed with the African American sample, 219 respondents (36.1%) were male and 388 (63.9%) were female. Their mean age was 48.8 years (SD = 17.9); their median annual household income was between $20,000 and $30,000; and their median level of educational attainment was "high school graduate." Most of the respondents (68%) labelled themselves as Democrats; 8.1% were Republican; and 16% were Independents. More than one-third (37.7%) knew a person with AIDS or infected with HIV.

Levels of stigma among African Americans. Responses to the stigma and beliefs items are presented in Tables 3 and 4. More than one-fifth of the African American respondents reported feelings of disgust or anger toward PWAs, whereas approximately one-third reported feeling afraid of them (see Table 3). Four out of ten Blacks agreed that people with AIDS should be quarantined and that the names of people with AIDS should be made public. Approximately one-sixth of the African American respondents felt that people with AIDS deserved their illness. As shown in Table 3, between 15% and 22% of African American respondents expressed their unwillingness to care for a close friend with AIDS, to allow their own child to attend the same school as a student with AIDS, or to continue to work with a male coworker with AIDS. More than half would not patronize the store of a neighborhood grocer with AIDS.

Table 3:

Table 4:

A significant minority of Blacks believed that HIV transmission is possible through casual contact, ranging from 23% for a kiss on the cheek to 48% for insect bites (see Table 4). Roughly six out of ten Blacks believed that HIV could be transmitted by unprotected male homosexual intercourse when neither partner is infected with HIV. Approximately one-fourth believed transmission between uninfected homosexual men was possible even if condoms were used, and that injecting drug use transmits HIV even when needles are not shared.

Racial Comparisons for Summary Measures

For comparison purposes, the responses from Whites in the general adult sample (n = 436) are presented in Tables 3 and 4.It appears that Blacks were more concerned about possible transmission of HIV whereas Whites held more negative feelings toward persons with AIDS. Blacks expressed greater support for measures that would keep PWAs separate from others (e.g., quarantine, publishing names) and were more likely to say that they would avoid PWAs under various circumstances. Consistent with this pattern, Blacks also were more likely to overestimate the risk of HIV transmission in a variety of situations. Whites, in contrast, expressed more negative feelings toward persons with AIDS and a greater willingness to blame PWAs for their illness.

To highlight the overall trends in these differences, five Likert-type scales¹³ were constructed by summing responses to conceptually related items. For negative feelings toward PWAs and coercive policies, the responses were treated as a 4-point scale, with low scores indicating that the feeling was not reported or that the respondent disagreed with the item. Avoidance scores were computed by summing responses to the four hypothetical contact situations, with a value of 1 assigned to avoidant responses and a value of 0 to supportive or prosocial responses. For casual contact beliefs, responses were scored on a 5-point scale, ranging from a score of 1 for "impossible" to transmit to a score of 5 for "very likely" to transmit. For beliefs about homosexuality and drug use, responses to each of the three items were treated as a 4-point scale, ranging from 1 for "no chance" of infection to 4 for "almost sure" to become infected.

The five scales resulting from this procedure (see Table 5) demonstrated acceptably high reliability:

1. For negative feelings toward PWAs (items 1-3 in Table 3; possible range = 3-21), Cronbach's alpha = .60.
2. For support for coercive policies (items 4 and 5 in Table 3; possible range = 2-8), alpha = .61.
3. For avoidant behaviors (items 7-10 in Table 3; possible range = 0-4), alpha = .72.
4. For casual contact beliefs (items 1-5 in Table 4; possible range = 5-25), alpha = .82.
5. For risk group beliefs (items 6-8 in Table 4; possible range = 3-15) alpha = .69.

In addition, responses to the single item assessing blame for PWAs are also presented in Table 5, scaled on the same 4-point continuum as that used for the coercive policy items.

Scale scores were analyzed using two-way analyses of covariance (ANCOVA), with race (Blacks, Whites) and gender (females, males) as independent variables. Gender comparisons were made because women and men frequently have different relationships to the AIDS epidemic and those affected by it. Because most cases of AIDS in the United States have been traced to male homosexual behavior,¹¹ for example, it seemed likely that heterosexual men would be more likely than heterosexual women to stigmatize persons with AIDS, since heterosexual men generally express greater hostility toward gay men than do women.¹⁴

Respondents' highest level of formal education was entered as a covariate; it was coded on a 6-point ordinal scale ranging from eighth grade or lower to at least some graduate work. This variable was included because the African American sample was more likely than the general adult sample to include respondents from urban areas with high concentrations of Black households. Consequently, its members also were likely to have a lower socioeconomic status. Because including educational level in the ANCOVA only partially offsets the bias introduced by using two different sampling frames, the statistical comparisons that follow must be interpreted with caution.

Table 5:

Comparison of the scale scores (Table 5) confirmed the general pattern observed for the individual items. Blacks scored significantly higher than Whites on the coercive policies scale, whereas Whites scored higher on the negative feelings scale and the individual blame item. Blacks also scored significantly higher on both scales measuring beliefs about HIV transmission. A significant racial difference was not observed for the avoidant behaviors scale.

Gender differences were observed in support for coercive policies and avoidant behaviors. Regardless of their race, men were more likely than women to stigmatize PWAs on these two measures. Significant gender-by-race interactions indicated that White women were the least likely of any group to anticipate that they would avoid PWAs or to overestimate the risks of casual contact. No gender differences were observed for negative feelings toward PWAs, blame, or beliefs about homosexuality and drug use.

Overall Index of Stigma

Was stigma limited to a small core group of respondents, or did most respondents give a stigmatizing answer to at least some of the survey items? To answer this question, we counted the total number of stigmatizing responses each person gave to the items concerning negative feelings, coercive policies, blame, and avoidant behaviors. The distribution of scores on this 10-item "stigma index" were similar for Blacks and Whites. Only 16.5% of Blacks and 22% of Whites did not give any stigmatizing responses. A full 45% of Black respondents and 42% of Whites gave stigmatizing responses on three or more items. And 16% of Blacks and 15% of Whites gave six or more stigmatizing responses.

CONCLUSIONS

The results indicate that AIDS-related stigma remains a serious problem as the United States enters the second decade of the epidemic. Between one-third and one-fifth of the general public holds negative feelings toward PWAs, believes that they deserve their illness, or supports punitive measures to be taken against them. A somewhat smaller proportion would translate these sentiments into avoidance or rejection of a loved one, school child, or coworker with AIDS. Nearly half would avoid shopping at a neighborhood grocery where the owner had AIDS. Based on scores for the 10-point "stigma index," it appears that roughly 40% of Americans manifest AIDS-related stigma to some extent. Approximately one in six manifest what could be considered a high level of stigma.

AIDS stigma is probably fostered in part by erroneous beliefs about HIV transmission ^9,15 Many respondents overestimated the risks posed by various forms of casual contact. Furthermore, a disturbingly large number seemed to believe that male homosexual intercourse or injecting drug use in themselves cause AIDS, even when neither sexual partner is infected or when needles are not shared. Those who hold such beliefs may discount the effectiveness of prevention programs that encourage safer sex or use of clean needles, since they perceive homosexual sex and needle use (not HIV) to be causes of AIDS. Such misconceptions also are likely to increase the stigma experienced by gay men and injecting drug users.

Stigma is also pervasive among African Americans. Comparison of responses from Blacks and Whites, however, suggests racial differences in its manifestations. African Americans appear to be more concerned about HIV transmission and more eager to avoid PWAs than are White Americans. But they also harbor less negative personal feelings toward PWAs than do Whites. As discussed elsewhere,^1,6,9 AIDS-related stigma can result both from perceptions of AIDS as an incurable, progressive, and transmissible disease, and from its prevalence in the United States among members of already-stigmatized groups, especially gay men. The data presented here suggest that AIDS-related stigma among African Americans might derive primarily from the former concern, whereas stigma among Whites might primarily reflect hostility and condemnation toward outgroups.

A detailed analysis of the underlying sources of racial differences in stigma is beyond the scope of the present paper. Three possible explanations for it, however, can be considered here briefly. First, the differences may result from methodological artifacts. Whereas Whites were sampled through random-digit dialing, the African American sample was drawn from geographic areas with high concentrations of Black residents, primarily urban neighborhoods. Blacks living in predominantly non-Black neighborhoods or sparsely populated rural areas were excluded. One consequence of this approach is that middle- and upper-class African Americans living in predominantly White neighborhoods were not sampled. Because such individuals are likely to be more highly educated than others, and because education was observed to be correlated with responses to most of the knowledge and attitudes items, observed racial differences in responses might actually be due to educational differences between the samples. As described above, however, inter-group differences in item responses generally were significant when the effects of education were statistically controlled. We also found that the results reported here were comparable to those obtained by the National Center for Health Statistics for similar questions about AIDS transmission asked of Blacks at approximately the same time.¹⁶ Consequently, although the results described in the present paper may have been affected by sampling procedures, we believe that this effect is small.

At least two substantive factors may have affected differences between Blacks and Whites in their attitudes and beliefs surrounding the AIDS epidemic: (1) racial differences in personal contact with people who have AIDS or are at risk for contracting HIV, and (2) racial differences in trust for scientific and government authorities. Compared to Whites, considerably more African Americans reported that they personally knew someone with AIDS or HIV (37.7% of Blacks, 25.7% of Whites). In addition, almost twice as many Blacks (39.8%) as Whites (20.4%) reported feeling very worried that someone they knew or cared about would develop AIDS or become infected. Many respondents reported that these factors influenced their own feelings about AIDS. Among Blacks, 31.5% felt that knowing a person with AIDS had affected their own feelings about the epidemic "some" or "a great deal" (compared to 18.2% of Whites). And 65% of Blacks (compared to 55.8% of Whites) felt that their concern about a loved one contracting HIV in the future had affected their feelings "some" or "a great deal." Thus, Blacks were more likely than Whites to have had personal experience with a person with AIDS or to anticipate such an experience in the future; and they were more likely to feel that such experiences and expectations had affected their own responses to the AIDS epidemic.

Racial differences in AIDS-related attitudes may also result from differential levels of trust in the authorities responsible for society's response to the epidemic. Blacks were more likely than Whites to believe that "the government is using AIDS as a way of killing off minority groups" (20.1% of Blacks agreed, compared to 4.2% of Whites) and were more likely to disagree with the statement "I believe scientists and doctors who say AIDS is not spread by casual social contact" (27.7% of Blacks disagreed, compared to 14.1% of Whites). This distrustful outlook is not difficult to understand, given the many civil rights setbacks suffered by the African American community in recent years, as well as widespread knowledge about the Tuskegee Syphilis Study.^4,17 Indeed, references to the Tuskegee study were made by some African American participants in focus groups conducted prior to the survey.¹⁸

It is reasonable to infer that distrust for authorities concerning the transmission risks posed by casual contact fosters African Americans' greater willingness to endorse policies such as quarantine and their greater overestimation of the risks posed by casual contact. At the same time, however, Blacks' personal involvement with the epidemic and their personal experience with societal prejudice and discrimination may make them less willing to express blame or negative feelings toward PWAs. Thus, whereas Whites' support for restrictive policies might reflect underlying hostility toward PWAs, such support among African Americans may reflect their perception that the epidemic poses an immediate threat to their own community, one that must be faced without reliance upon White governmental and scientific authorities. Consequently, they adopt a conservative approach to HIV transmission that is coupled with compassion for persons already infected.

The results presented here underscore the importance of reducing stigma and fostering compassion toward persons with AIDS in conjunction with providing information about risk. Such efforts should be focused especially on men, who are more likely than women to support coercive policies and to avoid contact with persons with AIDS. This gender difference may result from a variety of factors, including the more nurturant role assigned to women in American society and women's generally lower levels of antigay prejudice.¹⁴

The results also highlight the importance of providing information about HIV through sources that are credible to the target audience. For African American audiences, this may well require that messages be formulated and presented by members of the Black community. Overestimation of the risks of HIV transmission through casual contact may reflect African Americans' disbelief and distrust of White-identified authorities more than a lack of knowledge.¹⁷

As the second decade of the epidemic progresses, the need to understand AIDS-related stigma will become greater than ever before. Increasing numbers of uninfected people will be confronted with the need to reduce their own risk for HIV infection. And for every person newly-diagnosed with HIV disease, there will be many others – family members, friends, neighbors, coworkers, caregivers – who must respond to her or him in an informed and compassionate way. Voters will be faced with the challenge of electing candidates and endorsing policies for preventing HIV transmission and providing medical care to those already infected. Society also will have to absorb the epidemic's considerable economic costs. Because AIDS-related stigma will hamper the ability of individuals and society to respond effectively to the epidemic, understanding its social and psychological underpinnings is of critical importance.