No abstract

We examine a largely unexplored source of downward bias in peer effect estimation, namely, exclusion bias. We derive formulas for the magnitude of the bias in tests of random peer assignment, and for the combined reflection and exclusion bias in peer effect estimation. We show how to consistently test random peer assignment and how to estimate and conduct consistent inference on peer effects without instruments. The method corrects for the presence of reflection and exclusion bias but imposes restrictions on correlated effects. It allows the joint estimation of endogenous and exogenous peer effects in situations where instruments are not available and cannot be constructed from the network matrix. We estimate endogenous and exogenous peer effects in two datasets where instrumental approaches fail because peer assignment is to mutually exclusive groups of identical size. We find significant evidence of positive peer effects in one, negative peer effects in the other. In both cases, ignoring exclusion bias would have led to incorrect inference. We also demonstrate how the same approach applies to autoregressive models.

No abstract

Poor early childhood development in low-and middle-income countries is a major public health problem. Efficacy trials have shown the potential of early childhood development interventions but scaling up is costly and challenging. Guidance on effective interventions' delivery is needed. In an open-label cluster-randomized control trial, we compared the effectiveness of weekly home visits and weekly mother-child group sessions. Both included nutritional education, whose effectiveness was tested separately.METHODS: In Odisha, India, 192 villages were randomly assigned to control, nutritional education, nutritional education and home visiting, or nutritional education and group sessions. Mothers with children aged 7 to 16 months were enrolled (n = 1449). Trained local women ran the two-year interventions, which comprised demonstrations and interactions and targeted improved play and nutrition. Primary outcomes, measured at baseline, midline (12 months), and endline (24 months), were child cognition, language, motor development, growth and morbidity.RESULTS: Home visiting and group sessions had similar positive average (intention-to-treat) impacts on cognition (home visiting: 0.324 SD, 95% confidence interval [CI]: 0.152 to 0.496, P = .001; group sessions: 0.281 SD, 95% CI: 0.100 to 0.463, P = .007) and language (home visiting: 0.239 SD, 95% CI: 0.072 to 0.407, P = .009; group sessions: 0.302 SD, 95% CI: 0.136 to 0.468, P = .001). Most benefits occurred in the first year. Nutrition-education had no benefit. There were no consistent effects on any other primary outcomes.CONCLUSIONS: Group sessions cost $38 per child per year and were as effective on average as home visiting, which cost $135, implying an increase by a factor of 3.5 in the returns to investment with group sessions, offering a more scalable model. Impacts materialize in the first year, having important design implications. WHAT'S KNOWN ON THIS SUBJECT: In low-and middle-income countries millions of young children have poor development. Efficacy trials show that stimulation and nutritional interventions benefit children' s development but evidence on cost-effective methods of going to scale is urgently needed.WHAT THIS STUDY ADDS: Mother-child group sessions were as effective on average as home visits in improving child cognition and language in Odisha, India. Groups required 28% of the cost of home visits, substantially improving scalability of child psychosocial interventions.

This paper reports on a randomized survey experiment among 1840 households, designed to compare pen-and-paper interviewing (PAPI) to computer-assisted personal interviewing (CAPI). We find that PAPI data contain a large number of errors, which can be avoided in CAPI. We show that error counts are not randomly distributed across the sample, but are correlated with household characteristics, potentially introducing sample bias in analysis if dubious observations need to be dropped. We demonstrate a tendency for the mean and spread of total measured consumption to be higher on paper compared to CAPI, translating into significantly lower measured poverty, higher measured inequality and higher income elasticity estimates. Investigating further the nature of PAPI's measurement error for consumption, we fail to reject the hypothesis that it is classical: it attenuates the coefficient on consumption when used as explanatory variable and we find no evidence of bias when consumption is used as dependent variable. Finally, CAPI and PAPI are compared in terms of interview length, costs and respondents' perceptions. 1 We gratefully acknowledge financial support from the World Bank's multi-year research agenda in survey methodology (LSMS Phase IV). We appreciate permission from the Millennium Challenge Corporation (MCC) to build on their existing survey in Pemba. We thank Kathleen Beegle, David McKenzie, Kinnon Scott and participants at the Conference on Survey Design and Measurement in Washington DC for feedback on the experiment's design and an earlier draft of this paper. The paper was substantially improved after incorporating suggestions made by the editors and two anonymous referees. Leonard Kyaruzi, Deogratias Mitti and Mujobu Moyo lead the field teams, while Alessandro Romeo and Thaddaeus Rweyemamu took care of data entry of the paper questionnaires.