top of page

Applications to HIV Prevention

For two decades, I have obtained NIH funding to apply the theoretical principles uncovered in my lab to curb disease in areas related to HIV and lifestyle change. This work encompasses several projects, going from modeling intervention effects, to randomized controlled trials, to social media work. Below are some illustrations of our findings.

Guidelines about Appropriate Messages and Contents for HIV-Prevention Interventions Targeting Diverse Groups

My work on analyzing the impact of different intervention contents has led to the design of best practices in the area of HIV. The following is an example (Albarracín et al., 2005; Albarracín & Glasman, 2016) of how systematic social psychological research can be used to delineate public health guidelines with respect to the types of interventions necessary for different genders, ethnicities, ages, and risk groups (e.g., men who have sex with men, or injection drug users).

Guidelines about Appropriate Communicators for Women and Ethnic Minority Recipients of Behavioral Interventions

We have also investigated the types of communicators (message source, counselor) that are most effective for privileged and disenfranchised groups (Durantini, Albarracín, et al., 2006; Albarracín & Glasman, 2016). Contrary to common beliefs that non-expert community members are better to reach relatively marginalized or powerless groups, we found that women and African-Americans actually increase condom use to a greater extent when experts rather than community members appeal to them. This does not mean, however, that any expert will do. Instead, the experts must be similar to the audience in whatever characteristic makes that group distinct. Women respond better to female experts, African Americans respond better to African-American experts, and so forth. These findings should have a strong impact on health policy and may also illuminate decisions with respect to affirmative action in admissions to professional and graduate schools.

Evidenced-Based Interventions in the HIV Domain


An important intervention now being used in Florida health departments involves an invitation to the program based on the notion of defensive confidence, which I defined and investigated years ago. Defensive confidence entails the perceived ability to defend attitudes and behaviors when they come under attack (Albarracín & MItchell, 2004). People with low defensive confidence feel vulnerable when faced with counter-attitudinal information and, as a result, avoid exposure to uncongenial messages. As a result, we reasoned, it would be possible to increase openness to counter-attitudinal information by increasing defensive confidence, a pattern that would have major implications for behavioral-change programs.


Behavioral change interventions such as those to increase condom use or healthy lifestyles often face the dilemma of “preaching to the choir.” Practitioners attempt to enroll clients who don’t want to do what the program proposes, with the goal of getting them to do what they oppose. My defensive confidence work gave me the idea of anchoring enrollment decisions on clients’ resistance by offering the program as an opportunity to validate what they are actually doing. I found that an empowering, defensive confidence introduction to the intervention as “opening doors” for the client produced higher rates of enrollment than promises that the intervention would increase preventive behaviors (Albarracín, Durantini, et al., 2011). In a second randomized controlled trial completed by my team (Albarracín, Wilson, et al, 2017), I also used a recipient-centered approach and obtained increases in retention in HIV-prevention counseling programs.

Social Media Impacts


A large number of social interactions currently unfold in the online world. To be able to study these interactions, their effects, and the adequacy of prior theory, we mine social media data in relation to structural, survey, and health data. Tracking behavior through social media may provide an unobtrusive, naturalistic means of predicting HIV outbreaks and understanding the behavioral and psychological factors that increase communities’ risk. We have used bottom-up (“Big Data”) approaches to predict HIV, gonorrhea and chlamydia from Twitter, achieving very good results. Below appear actual rates along with our online-risk index showing predicted rates.

Further, some of our work on social media has been top-down. Specifically, general action goals (Act Now, Just do it!) may support protective health behavior (e.g., using condoms) but also encourage activity indiscriminately (e.g., risky sex), resulting in mixed health effects (Ireland, …, & Albarracín, 2014). We found that, controlling for demographic and structural predictors of HIV, more active language was associated with lower HIV rates. In addition, we have examined community norms by conducting hybrid, top-down bottom-up topic analyses. Because gay men of color are a priority group from the point of view of HIV prevention, testing, and treatment, we have investigated normative predictors of disease in counties with higher ethnic-minority representation and higher proportion of same-sex households. We found that traditional, risk seeking, hyper-masculinity notions (see figure) are more associated with HIV in counties that are high on these demographics. By leveraging language used on social media to improve existing predictive models of geographic variation in HIV, future targeted HIV-prevention interventions may have a better chance of reaching high-risk communities before outbreaks occur.

bottom of page