Selected Working Papers
Neighborhood Civic Capital and COVID Vaccine Uptake in Sweden: New Evidence from Population-Wide Registry Data
​
(Co-authored with Olof Mpumwire Östergren)
​
Civic capital – the prevalence of communal institutions and practices in a given area, linking people beyond biological kinship into sets of dense sets of local relationships – is often thought to improve resilience against public health cries by acting as a channel for the rapid dissemination of new protective behaviors. Yet gauging civic capital’s importance remains challenging. Most projects mapping civic capital use geographical tracts containing hundreds of thousands of individuals, making precise comparisons difficult. This paper, focused on the Swedish case, builds the most granular national map of civic capital to date (using areas with approximately 1,700 people each), to describe its association with vaccine uptake. We find that civic capital is positively associated with increased vaccine uptake, with its estimated effect across the distribution (+/- 2 S.D.) equivalent to two thirds of the gap between those with and without a college degree by the third vaccine. These findings have important implications for the study of neighborhood effects, the role of local norms in shaping health outcomes, and the sociology of crisis.​
​
​
Publication bias inflates effect size estimates for common educational interventions
​
(Co-authored with Bastian Betthäuser, Susan Swingler, Vanessa Wittemann, Anders Malthe Bach-Mortensen, Gaia Grassi, Niklas Ayris, Una Oljaca)
​
Robust evidence on the effectiveness of learning interventions that supplement regular in-school instruction is critical for supporting children in recovering learning setbacks that arise due to major disruptions of schooling, such as the recent COVID-19 pandemic. We provide a systematic review, quality appraisal, and meta-analysis of 194 estimates across 18 countries of the effects of four key types of supplementary learning intervention—afterschool programmes, extending instruction time, tutoring, and summer programmes—on children’s learning progress in maths and reading. We find evidence of substantial bias in the existing body of evidence. Correcting for publication bias substantially reduces the pooled effect size for extending instruction time, summer programmes, and tutoring, while the pooled effect size for afterschool programmes are rendered statistically insignificant. Examining heterogeneity within interventions, we find that interventions with an exclusive focus on maths or reading tend to be more effective than interventions that combine multiple subjects, interventions targeting students in secondary education tend to be more effective than those targeting primary students, and intervention intensity and duration are positively associated with intervention effects. Decision makers, practitioners, and researchers should consider the substantial extent of bias in the existing literature and the moderating role of intervention characteristics when designing, implementing, and evaluating supplementary learning interventions.