List randomization for eliciting HIV status and sexual behaviors in rural KwaZulu-Natal, South Africa: a randomized experiment using known true values for validation

Nonverbal Response Cards Reduce Socially Desirable Reporting of Violence Among Adolescents in Rural Burkina Faso: A Randomized Controlled Trial

PURPOSE

Accurate measures of violence are difficult to obtain from self-reported data because of stigmatization and social undesirability of the topic. Most methods that attempt to reduce such biases require literacy and either remove the benefits of interviewer guidance or do not give individual-level results. We tested a low-tech nonverbal response card that avoids revealing interviewees' responses to interviewers while retaining interviewer support among adolescents in communities with very low educational attainment.

METHODS

As part of a broader health questionnaire, we asked a sample of 1,644 adolescents, aged 12-20 years, in northwestern Burkina Faso about their experiences of physical and sexual violence. We randomized participants to either a conventional verbal response arm or a nonverbal response card arm where respondents' answers were unspoken and not displayed to interviewers. We first evaluated response validity and reliability in each arm, then compared prevalence rates across arms and evaluated whether any differences varied by respondent characteristics using regression models.

RESULTS

The level of internal reliability of responses among nonverbal respondents was similar to or greater than that of verbal respondents. Nonverbal respondents reported similar patterns of physical assault and sexual debut as verbal respondents but significantly higher levels of sexual assault and forced sex. These differences were broadly similar across sample subgroups defined by age, gender, proneness to social desirability, and mental health.

CONCLUSIONS

Nonverbal response cards offer a practical and beneficial method for reducing underreporting of stigmatized and traumatic experiences while maintaining data quality in low-literacy populations.

HIV infection risk and condom use among sex workers in Senegal: evidence from the list experiment method

Social desirability bias, which is the tendency to under-report socially undesirable health behaviours, significantly distorts information on sensitive behaviours that is gained from self-reports. As a result, self-reported condom use among high-risk populations is thought to be systematically over-reported, and it is impossible to identify the determinants of condom use. The main objective of the article is to elicit unbiased information on condom use among female sex workers (FSWs) using the double list experiment method to analyse the role of HIV infection and exposure to HIV prevention methods in condom use. More specifically, the difference in levels of condom use between HIV-positive and HIV-negative FSWs is estimated. In addition, the role of FSWs' registration and participation in a pre-exposure prophylaxis (PrEP) demonstration project in condom use is considered. A list experiment was designed to elicit condom use information from 786 FSWs in Senegal who were surveyed in 2015 and 2017. Using the list experiment method, participants were randomly assigned to one of two groups (treatment or control) and were asked to report the number of statements they agreed with. Respondents assigned to the control group were presented with three non-sensitive items, whereas those allocated to the treatment group were presented with the same three statements plus the sensitive item (e.g. 'I used a condom during my last intercourse with a client'). Comparing the average number of sentences that were agreed with in both groups provides an estimation of the condom use rate in the treatment group and estimating such prevalence for several sub-groups allows the role of HIV infection risk in condom use to be identified. The percentage of FSWs using condoms in their last sexual intercourse with a client was 80% in 2015 and 78% in 2017, which was significantly lower than the 97% obtained in the face-to-face surveys in both waves. When estimating condom use among sub-groups with the list experiment method, we found that condom use among HIV-positive FSWs was only 34%, which was 47 percentage points lower than condom use among HIV-negative FSWs. We also found that registered FSWs are more likely to use condoms than clandestine FSWs. However, we did not find any difference in condom use between FSWs who were enrolled in the PrEP demonstration project and those who were not enrolled. Health policies should therefore aim to increase condom use among HIV-positive FSWs.

Nothing but the truth: Consistency and efficiency of the list experiment method for the measurement of sensitive health behaviours

RATIONALE

Social desirability bias, which is the tendency to under-report socially, undesirable health behaviours, significantly distorts information on sensitive behaviours, gained from self-reports and prevents accurate estimation of the prevalence of those, behaviours. We contribute to a growing body of literature that seeks to assess the performance of the list experiment method to improve estimation of these sensitive health behaviours.

METHOD

We use a double-list experiment design in which respondents serve as the treatment group for one list and as the control group for the other list to estimate the prevalence of two sensitive health behaviours in different settings: condom use among 500 female sex workers in urban Senegal and physical intimate partner violence among 1700 partnered women in rural Burkina Faso. First, to assess whether the list experiment improves the accuracy of estimations of the prevalence of sensitive behaviours, we compare the prevalence rates estimated from self-reports with those elicited through the list experiment. Second, we test whether the prevalence rates of the sensitive behaviours obtained using the double-list design are similar, and we estimate the reduction in the standard errors obtained with this design. Finally, we compare the results obtained through another indirect elicitation method, the polling vote method.

RESULTS

We show that the list experiment method reduces misreporting by 17 percentage points for condom use and 16-20 percentage points for intimate partner violence. Exploiting the double-list experiment design, we also demonstrate that the prevalence estimates obtained through the use of the two lists are identical in the full sample and across sub-groups and that the double-list design reduces the standard errors by approximately 40% compared to the standard errors in the simple list design. Finally, we show that the list experiment method leads to a higher estimation of the prevalence of sensitive behaviours than the polling vote method.

CONCLUSION

The study suggests that list experiments are an effective method to improve estimation of the prevalence of sensitive health behaviours.

Evidence of high bed net usage from a list randomization experiments in rural Gambia

BACKGROUND

Recording behaviours that have the potential to impact health can be doubly challenging if the behaviour takes place in private spaces that cannot be observed directly, and where respondents answer what they think the recorder may want to hear. Sleeping under a long-lasting insecticidal net (LLIN) is an important intervention for malaria prevention, yet it is difficult to gauge the extent to which coverage (how many nets are in the community) differs from usage (how many people actually sleep under a net). List randomization, a novel method which partially obscures respondents' answers to sensitive questions, was employed to estimate LLIN usage in The Gambia.

METHODS

802 heads-of-household from 15 villages were recruited into a randomized controlled trial assessing the effect of a housing intervention on malaria. These houses were randomly assigned to a housing intervention versus control, with stratification by village so as to ensure balance between arms. From these, 125 households (63 intervention, 52 control) were randomly selected for participation in the list randomization experiment, along with 68 households from the same villages but which were not part of the housing improvement study, resulting in a total of 196 households for the list randomization experiment. Approximately half (n = 97) of the 196 study participants were randomly assigned to the control group and received a four-question list about non-sensitive behaviours; the intervention group (n = 99) received the same list, with the addition of one question on a sensitive behaviour: whether or not they had used a bed net the previous night. Participants were read the list of questions and then said how many of the statements were true. Bed net usage was estimated by calculating the difference in means between the number of affirmative responses between the two groups.

RESULTS

The mean number of affirmative responses in the control group was 2.60 of four statements (95% confidence interval, 95% CI 2.50-2.70), compared with 3.68 (95% CI 3.59-3.78) in the intervention group. Such difference (1.08; 95% CI 94.9-100%) suggests near universal bed net usage.

CONCLUSIONS

Bed net usage by household heads in these rural villages was found to be high. Though not entirely unexpected given other studies' estimates of high bed net usage in the area, the list randomization method should be further validated in an area with lower coverage.