Reliability of the Marlowe-Crowne social desirability scale in Ethiopia, Kenya, Mozambique, and Uganda
1 International Emergency and Public Health Fellowship Program, Department of Emergency Medicine, The Johns Hopkins School of Medicine, Baltimore, USA
2 Health Systems Program, Department of International Health, The Johns Hopkins Bloomberg School of Public Health, Baltimore, USA
3 Department of Population & Family Health and Department of Biostatistics, The Johns Hopkins Bloomberg School of Public Health, Baltimore, USA
BMC Medical Research Methodology 2011, 11:162 doi:10.1186/1471-2288-11-162Published: 2 December 2011
Studies of HIV often use self-reported surveys to measure sexual knowledge, attitudes, and practices. However, the self-reported data are vulnerable to social desirability (SD), a propensity of individuals to report favorable responses. The Marlowe-Crowne Social Desirability Scale (MC-SDS) was developed as a measure of the effect of social desirability, but it has not been adapted for or used in Africa. This study aimed to apply the MC-SDS nested in an HIV behavioral intervention program and to measure its reliability in four African countries.
The MC-SDS was adapted based on consultations with local stakeholders and pilot tested in Ethiopia, Kenya, Mozambique, and Uganda. Trained interviewers administered the modified 28-item MC-SDS survey to 455 men and women (ages 15-24 years). The scores for the social desirability scales were calculated for all participants. An analysis of the internal consistency of responses was conducted using the Cronbach's α coefficient. Acceptable internal consistency was defined as an α coefficient of ≥ 0.70.
Mean social desirability scores ranged from a low of 15.7 in Kenya to a high of 20.6 in Mozambique. The mean score was 17.5 for Uganda and 20.6 for Mozambique. The Cronbach's α coefficients were 0.63 in Kenya, 0.66 in Mozambique, 0.70 in Uganda, and 0.80 in Ethiopia.
The MC-SDS can be effectively adapted and implemented in sub-Saharan Africa. The reliability of responses in these settings suggest that the MC-SDS could be a useful tool for capturing potential SD in surveys of HIV related risk behaviors.