1. Introduction

2. Methods

In September/October 2003, a team of 10 Ariaal enumerators administered a questionnaire based on the UNAIDS (1998) document Looking deeper into the HIV epidemic: A questionnaire for tracing sexual networks to 400 respondents in Nasakakwe/Karare. Aimed at reproductive-aged Ariaal men and women, 100 respondents were represented in each of four subgroups: unmarried men, unmarried women, married men, and married women. Unlike in earlier surveys, all respondents had heard of HIV/AIDS, and subsequent focus groups in July 2004 revealed widespread knowledge of HIV transmission and high-risk behavior.

The survey contained questions on sexual behavior, including the number of lifetime sexual partners, number of partners in the past year, degree of mixing among the four major subgroups, duration of relationships, and patterns of condom use. In addition, the survey asked each respondent to evaluate their own personal risk of acquiring HIV, dichotomized as “low” or “high.” These data were used to delineate self-assessment of HIV risk using the methodology developed by Prata et al. (2003). In this approach, risk scores were calculated, with respondents assigned one point for each behavior associated with a high risk of contracting HIV. For our data, these behaviors included having multiple partners, both in the past year and over one's lifetime (Anderson & May, 1988), sexual mixing between age and economically differentiated subpopulations (Gregson et al., 2002), concurrency (defined as partnerships that overlap in time; Morris, 2001), and non-use of condoms (Bracher et al., 2003).

All these risk factors were explained to community participants during the education seminars in Nasakakwe/Karare. Although not all residents attended these seminars, information imparted in them diffused quickly throughout the community. At the time of our 2003 survey, all respondents knew about these seminars and could recognize the general concept of high-risk behavior as well as specific high-risk acts. Assuming the presence of self-deception, we hypothesized that risk scores would not differ directionally with self-assessment status; for example, those who rate their risk of contracting HIV as “low” would not have significantly lower average risk scores than do those with “high” self-assessments.

3. Results

Fig. 1 depicts the percentages of specific reported high-risk behavior arranged by subgroup, clearly showing different patterning of risks associated with each. Also notable is the very low level of condom use for all groups, but especially low rates for both married men and women (respectively, 3% and 1%) relative to those for unmarried men (16%) and women (14%). This mirrors past studies indicating low rates of condom use within marriage throughout sub-Saharan Africa (cf. Caldwell, 1999) due to their association with illegitimate sexual unions and prostitution.

Fig. 1. Percentage at risk by factor and subgroup.

Fig. 2 presents the means and standard deviations for the composite risk scores by subgroup and self-assessed risk status, that is, “low” versus “high.” Results show that for all subgroups, the average risk score for those who rated their risk of contracting HIV as “high” was higher than for those who assessed their risk as “low.” In all but the married men sample the average score differences by response were significantly different (tunmarried men=−2.13, P=.038; tunmarried women=−3.08, P=.003; tmarried men=−0.087 P=.387; tmarried women=−2.60, P=.010). These results suggest that Ariaal subgroups make accurate overall self-assessments of HIV risk based on their reported sexual behavior. This negates our null hypothesis that predicted no difference in risk scores for “low” versus “high” risk self-assessments.

Fig. 2. Composite risk scores (means and standard deviations) derived from self-reported sexual behavior, arranged by “low” and “high” risk self-assessments and subgroup.

4. Discussion

As Caldwell originally pointed out, throughout sub-Saharan Africa, knowledge of HIV transmission does not lead to sexual behavioral change. In the current paper, we hypothesized that populations possessing high knowledge levels of HIV/AIDS while exhibiting little or no individual change in sexual behavior may constitute an example of the evolution of human self-deception. Essential to this hypothesis was self-assessment of HIV/AIDS risk. Using data from a 2003 survey of Ariaal agropastoralists in the sedentary community of Nasakakwe/Karare, we constructed a test of self-assessment under the assumption that inaccurate risk assessment stemmed from self-deception.

Results did not support our hypothesis. Calculated risk scores were always higher for those who evaluated their risk of contracting HIV/AIDS as “High” than those whose self-assessment was “Low.” The failure of these results to support our evolutionary hypothesis for self-deception raises the question of the usefulness of applying evolutionary principles to HIV/AIDS research. However, while we could not demonstrate self-deception in these analyses, an equally important finding is that the population continues high-risk sexual behavior, even in the face of accurate self-assessment of high risk of contracting HIV.

One possible cause for the misfit between accurate risk self-assessment and high prevalence of risk behavior in the Ariaal community may be poor accessibility to condoms. However, this is unlikely for two reasons. First, condoms' association with prostitution and promiscuity poses difficulties in motivating and negotiating condom use, particularly for women within marriage or steady relationship (Chimbiri, 2003). Second, condom use is linked to perceptions of diminished sexual sensation, especially for men (McPhail & Campbell, 2001).

On a more general scale, Ariaal are certainly not alone in exhibiting such a misfit. The interview data of Bailey and Aunger (1995) from farming and foraging groups in the Ituri region of Central Africa show that all groups clearly understand the relationships between early sexual debut, increased risk of STDs, and reduced fertility, yet maintain risky sexual cultural practices, for example, early age of sexual initiation, multiple partners, and extramarital partners. In explanation, Bailey and Aunger argue that humans are designed by evolution to actively seek sexual contact and gratification and as such:

An outcome of our remarkable capacity for sexual pleasure and desire for sexual contact is an apparent inability or unwillingness to detect and avoid the pathogens transmitted by the frequent sexual contacts we seek. While we have the cognitive processes that enable us to be conscious of the risks associated with sexual contacts, the evolved psychological mechanisms driving our desires may be overcome only under extreme conditions. (Bailey & Aunger, 1995, p. 218)

These evolved mechanisms are not limited to sub-Saharan Africa. The “extreme conditions” alluded to above were exemplified by Thailand's “100% Condom Program.” Begun in the early 1990s, this government sponsored program provided free condoms to commercial sex establishments. As a result, condom use in these establishments went from less than 20% in 1989 to over 90% by 1992, the number of new STD cases for men treated at government clinics dropped from 200,000 in 1989 to less than 20,000 in 1992, and HIV prevalence rates among Thai army conscripts declined from 4% in 1993 to less than 2% in 1996. The program's success has been attributed to “mass media campaigns, education and skill-building in workplaces and schools, and peer education” (World Bank, 1999, p. 159).

An equally viable alternative explanation is widespread government coercion. Health officials, government workers, and police were authorized to enforce condom use. Government officials posed as clients in commercial sex establishments to check on condom availability and use. Uncovering noncondom use in sexual acts made individual offenders liable for fines, and commercial sex houses faced loss of government licensing and/or immediate closure. Recognizing this coercive pressure to conform, the 100% Condom Program can also be understood as the evolution of group morals through punishment, as modeled by Boyd and Richerson (1992) in their article aptly entitled “Punishment allows the evolution of cooperation (or anything else) in sizeable groups.” Notably, however, despite the threat of penalties, or perhaps as a result of them, Thailand now has “invisible brothels,” that is, restaurant bars and karaoke pubs where sex workers serve as waitresses, and private massage parlors where “condom use is more difficult” (Im-Em, 1999, p. 168) to enforce.

In the Thai, Ituri, and Ariaal examples, despite accurate understanding of high-risk behavior, there is still strong reluctance to modify such behavior. Self-deception is not present in these cases, but rather the most important variable may be what Watkins (2005, p. 681) terms “the joys of sex.” We suggest that these joys represent the evolved capacity for human sexual pleasure. Certainly, there is evidence for strong selection for sexual gratification in humans, with many of the psychological, morphological, and behavioral characteristics differentiating humans from other species linked to sexual concerns, for example, greater prominence of female secondary sex organs, lack of female estrus, concealed ovulation, greater prominence of orgasms, and so forth (Abramson & Pinkerton, 1995). Given the human evolved drive for sexual gratification, populations react differently to the new threat of HIV/AIDS, with many failing to fully adopt safe sexual practices if these are inconsistent with the prospect of sexual gratification. Some groups, exemplified here by the survey of Canadian adolescents, and worldwide findings from homosexual communities, do employ self-deception to continue or resume high-risk behavior. Others, represented by the Thai, Ituri, and Ariaal, do not resort to self-deception, but still continue high-risk behavior even in the face of accurate self-assessment of risk.

The major challenge to HIV/AIDS behavioral research is to initiate sexual behavior change. If indeed evolved traits are underlying causes for our reluctance to reduce risk behavior, evolutionary theory provides important insights for the rationale for such reluctance. The important question now is can it also provide directions for overcoming this reluctance and motivating behavior change.