We introduce the best-friend methodology for using surveys to measure the population prevalence of sensitive behaviors. We demonstrate the effectiveness of this tool by comparing self-reports to best-friend reports of sexual behavior and abortion history among young women in Malawi (n = 1493). Best-friend reports reveal higher and more believable estimates of abortion and multiple sexual partners. In contexts in which best friends commonly discuss such behaviors, best-friend reports are an inexpensive and easily implemented tool.

Survey researchers struggle to measure sensitive and stigmatizing behaviors that are critical for understanding health risks, patterns of disease transmission, and contraceptive need.1–4 Novel methodological approaches to reduce underreporting, including audio computer-assisted self-interviewing and confidential voting techniques, have had limited success. These methods can elicit higher (and arguably more accurate5) reports of certain sexual behaviors and abortion, but they produce lower reports of other behaviors and can increase reporting errors.6–11

To better estimate the population prevalence of sexual risk behavior and abortion in a community in southern Malawi, we administered a survey asking a series of questions about the behaviors of the respondent's best friend. This method leverages 2 sociological insights. The first is that friendship networks are homogenous; friends tend to be similar with respect to sociodemographic characteristics and behaviors.12 The second is that social desirability bias should have less influence on reports of the sensitive or stigmatizing behaviors of one's friends than on self-reports.13 Here we present the initial results obtained by using this easily implemented approach to survey research on sensitive topics.

Malawi is in the midst of a generalized and mature AIDS epidemic. Young women are particularly likely to become infected, with HIV incidence peaking in the mid-20s.14 Premarital sex, although common, is nonnormative, creating a social desirability bias that encourages individuals, and women in particular, to underreport extramarital partnerships and sexual activity before marriage.15,16

Abortion is illegal in Malawi except to save the life of the pregnant woman.17,18 Nonetheless, clandestine abortion—largely unsafe—is a common response to unwanted or mistimed pregnancies throughout the country. There are no nationwide estimates of abortion in Malawi, but data from major hospitals suggest that abortion complications are the most common reason for hospital admission to gynecological wards.19,20

Tsogolo la Thanzi (TLT) is a new and ongoing longitudinal study of young (15–25 years) Malawians living within a 7-kilometer radius of the southern town of Balaka. The baseline data were collected from a random sample of women and men from June 2009 through August 2009. Only the data collected from the sample of women (N = 1493) are discussed here.

Interviewers collected the data in face-to-face interviews that lasted approximately 80 minutes each. The survey contained questions about the respondent's sexual behavior and the history and outcome of each pregnancy she had ever experienced. Because of the sensitive nature of such questions, TLT complemented this direct approach with questions about the respondent's best friend's sexual behavior and abortion history. The interviewer introduced the best-friend section by asking the respondent to “think about your best female friend in Balaka. Picture her. Do you have this person in mind?” Interviewers then asked the respondent a series of questions about the best friend, starting with less sensitive questions and moving on to more sensitive questions. Because the relationship between best friends is typically reciprocal, this approach approximates a 1-to-1 association.

In the aggregate, TLT respondents’ answers to nonsensitive questions about their own marriage and childbearing closely mirrored the answers they gave to the same questions about their best friends (Figure 1). For example, 42.1% of respondents reported being currently married, compared with 42.8% reporting that their best friends were married. The same was true for having children (49.2% for both) and current pregnancy (9.6% for respondents and 8.8% for their best friends). For more sensitive questions, however, significant differences emerged between self-reports and best-friend reports. The percentage of respondents who reported having had sex in the past 12 months (61.3%) was a full 11 percentage points lower than the percentage who reported that their best friends had had sex in the past 12 months. Underreporting was systematic by marital status; the difference between self-reports and best-friend reports was negligible for married women and large for their unmarried counterparts, reflecting prevailing norms discouraging premarital sexual activity.

Additionally, although very few respondents acknowledged having had more than 1 sexual partner in the past year (1.7%), multiple partners were reported to be relatively common among best friends (20.3%). In the retrospective pregnancy history, only 0.4% of respondents admitted to having had an abortion in their lifetime. By contrast, 7.6% of respondents reported that they were confident their best friend had ever had an abortion, and an additional 0.5% suspected as much.

Best-friend reports do not solve all problems with self-reports on sensitive issues. They are insufficient for examining behaviors that are unlikely to be discussed between friends or for establishing causal mechanisms. However, best-friend reports can easily provide better estimates of the prevalence of sensitive and underreported behaviors in populations. Soliciting a respondent's reports on the characteristics and behavior of his or her best friend (who is likely similar to the respondent) is an effective and inexpensive way to collect information on sensitive issues. At the population level, the best-friend approach could be validated with large-scale surveillance studies that compare patterns of sexual behavior as established by best-friend reports to observed trends in sexually transmitted infections. At the individual level, the best-friend approach could be validated with intensive network data attaching best-friend reports to biomarker data from respondents and their best friends. Once best-friend reports have been validated thoroughly, they may have extensive applications, including gauging the accuracy of self-reported behavior and enhancing epidemiological models of the spread of sexually transmitted infections by improving their inputs.


This research was supported by funding from the National Institute of Child Health & Human Development (grant R01-HD058366).

We thank Abdallah Chilungo, Sydney Lungu, Hazel Namadingo, and the Tsogolo la Thanzi (TLT) interviewers for their assistance in collecting TLT data. We also thank Michelle Frisco for helpful comments on an earlier version of the article.

Human Participant Protection

This study protocol was approved by the Pennsylvania State University Office for Research Protections and the Malawi National Health Sciences Research Committee.


1. Plummer ML, Ross DA, Wight D, et al. “A bit more truthful”: the validity of adolescent sexual behaviour data collected in rural northern Tanzania using five methods. Sex Transm Infect. 2004;80:ii49ii56. Crossref, MedlineGoogle Scholar
2. Fenton KA, Johnson AM, McManus S, Erens B. Measuring sexual behaviour: methodological challenges in survey research. Sex Transm Infect. 2001;77(2):8492. Crossref, MedlineGoogle Scholar
3. Jones EF, Forrest JD. Underreporting of abortion in surveys of United States women: 1976 to 1988. Demography. 1992;29(1):113136. Crossref, MedlineGoogle Scholar
4. Jagannathan R. Relying on surveys to understand abortion behavior: some cautionary evidence. Am J Public Health. 2001;91(11):18251831. LinkGoogle Scholar
5. Hewett PC, Mensch BS, Ribeiro MC, et al. Using sexually transmitted infection biomarkers to validate reporting of sexual behavior within a randomized, experimental evaluation of interviewing methods. Am J Epidemiol. 2008;168(2):202211. Crossref, MedlineGoogle Scholar
6. Jaya, Hindin MJ, Ahmed S. Differences in young people's reports of sexual behaviors according to interview methodology: a randomized trial in India. Am J Public Health. 2008;98(1):169174. LinkGoogle Scholar
7. Mensch BS, Hewett PC, Gregory R, Helleringer S. Sexual behavior and STI/HIV status among adolescents in rural Malawi: an evaluation of the effect of interview mode on reporting. Stud Fam Plann. 2008;39(4):321334. Crossref, MedlineGoogle Scholar
8. Gregson S, Mushati P, White PJ, Mlilo M, Mundandi C, Nyamukapa C. Informal confidential voting interview methods and temporal changes in reported sexual risk behaviour for HIV transmission in sub-Saharan Africa. Sex Transm Infect. 2004;80:ii36ii42. Crossref, MedlineGoogle Scholar
9. Hewett PC, Mensch BS, Erulkar AS. Consistency in the reporting of sexual behaviour by adolescent girls in Kenya: a comparison of interviewing methods. Sex Transm Infect. 2004;80:ii43ii48. Crossref, MedlineGoogle Scholar
10. Mensch BS, Hewett PC, Erulkar AS. The reporting of sensitive behavior by adolescents: a methodological experiment in Kenya. Demography. 2003;40(2):247268. Crossref, MedlineGoogle Scholar
11. Fu H, Darroch JE, Henshaw SK, Kolb E. Measuring the extent of abortion underreporting in the 1995 National Survey of Family Growth. Fam Plann Perspect.1998;30(3):128133, 138. Crossref, MedlineGoogle Scholar
12. McPherson M, Smith-Lovin L, Cook JM. Birds of a feather: homophily in social networks. Annu Rev Sociol. 2001;27:415444. CrossrefGoogle Scholar
13. Fisher RJ. Social desirability bias and the validity of indirect questioning. J Consum Res.1993;20(2):303315. CrossrefGoogle Scholar
14. White RG, Vynnycky E, Glynn JR, et al. HIV epidemic trend and antiretroviral treatment need in Karonga District, Malawi. Epidemiol Infect. 2007;135(6):922932. Crossref, MedlineGoogle Scholar
15. Nnko S, Boerma JT, Urassa M, Mwaluko G, Zaba B. Secretive females or swaggering males? An assessment of the quality of sexual partnership reporting in rural Tanzania. Soc Sci Med. 2004;59(2):299310. Crossref, MedlineGoogle Scholar
16. Poulin M. Reporting on first sexual experience: the importance of interviewer-respondent interaction. Demogr Res.2010;22(11):237288. Crossref, MedlineGoogle Scholar
17. United Nations Population Division. Abortion policies: a global review. Country profile: Malawi. Available at: http://www.un.org/esa/population/publications/abortion/doc/malawi.doc. Accessed March 5, 2010. Google Scholar
18. Government of Malawi. Penal code, section 150. Available at: http://www.hsph.harvard.edu/population/abortion/Malawi.abo.html. Accessed March 5, 2010. Google Scholar
19. Munthali AC, Chimbiri A, Zulu E. Adolescent Sexual and Reproductive Health in Malawi: A Synthesis of Research Evidence. New York, NY, and Washington, DC: Alan Guttmacher Institute; 2004. Google Scholar
20. Mtimavyale LAR. Sharpening the Malawian youth on issues related to abortion. Paper presented at: Demographic Society Seminar; 1997; Lilongwe, Malawi. Google Scholar


No related items


Downloaded 219 times



Sara Yeatman, PhD, and Jenny Trinitapoli, PhDSara Yeatman is with the Department of Health and Behavioral Sciences, University of Colorado Denver. Jenny Trinitapoli is with the Department of Sociology and Population Research Institute, Pennsylvania State University, University Park. “Best-Friend Reports: A Tool for Measuring the Prevalence of Sensitive Behaviors”, American Journal of Public Health 101, no. 9 (September 1, 2011): pp. 1666-1667.


PMID: 21778489