Details

Autor(en) / Beteiligte

• Babalola, Stella
• Kincaid, D. Lawrence

Titel

New Methods for Estimating the Impact of Health Communication Programs

Ist Teil von

• Communication methods and measures, 2009-04, Vol.3 (1-2), p.61-83

Ort / Verlag

Taylor & Francis Group

Erscheinungsjahr

2009

Quelle

Taylor & Francis Journals Auto-Holdings Collection

Beschreibungen/Notizen

• This paper demonstrates four analytical methods for evaluating the impact of full-coverage communication programs with cross-sectional survey data: propensity score matching (PSM), sensitivity analysis, biprobit regression, and triprobit regression. PSM adjusts for measured covariates and assumes that unmeasured covariates can be ignored; sensitivity analysis helps to test the validity of the assumption that unmeasured covariates are ignorable. Biprobit tests and corrects for endogeneity between two outcomes, and triprobit does so for three outcomes. Estimates from these methods using the same data are different, although they all indicate that exposure significantly increases the probability of consistent condom use and having a high level of ideation for condom use. Compared to triprobit regression and PSM, biprobit regression appears to overstate the impact of exposure. The study underscores the importance of taking theoretical pathways of influence into account when evaluating the impact of communication programs on a population.