This paper examines the largely understudied effect of age at school entry on adult mental health. Using nationwide Finnish register data, we capture a broad spectrum of mental health conditions, ranging from severe disorders to less severe problems. To identify causal effects, we implement a regression discontinuity design around the January 1 school-entry cutoff. We find that being born at the beginning of the year improves mental health outcomes, with the effects concentrated among women in their late twenties.
We evaluate a 10-week intervention in Peru that randomly offered remote coaching to parents of preschool children during the 2021 summer break, complementing publicly available mathematics lessons delivered by broadcast media. Coaches provided guidance and encouragement through phone calls, SMS, and printed materials designed to increase mathematics-related parent-child interactions. We find positive effects on mathematics achievement of approximately 0.10 standard deviations, although statistical significance is sensitive to conservative adjustments for multiple hypothesis testing, particularly when mathematics and oral comprehension are jointly evaluated. We do not find clear evidence of improvements in oral comprehension. By contrast, we document important and precisely estimated increases in mathematics-related parental engagement, including more time spent on targeted activities, greater participation in mathematics games, and more mathematicsfocused tutoring. These findings suggest that light-touch remote coaching can shift parental inputs and may translate into gains in early numeracy.
This paper provides the first causal evidence on how the pandemic-induced remote learning disrupted schools' capacity to detect child maltreatment. Leveraging county-level variation in remote instruction during the 2020-21 school year, I find that counties with higher exposure to remote instruction experienced a 7.2% greater decline in child maltreatment reports involving school-aged children, but a 13.4% increase in maltreatment-related child fatalities within this group. The decrease in reports by education personnel primarily drove the overall decline. Effects on maltreatment-related fatalities persisted even after schools resumed in-person instruction. These results highlight an unintended cost of distance learning: remote instruction impaired the detection of child maltreatment, leading to fewer reports but more severe cases. Prompt policy interventions could safeguard children who remain undetected.
Economic downturns often come along with a reduction of apprenticeship training by firms, hereby threatening economies with a lack of skilled labour in the mid- and long-run. Against this background, the German government has implemented an apprenticeship subsidy for small and medium sized companies in the summer of 2020 during the COVID-19 pandemic. In brief, companies that maintained their number of new apprenticeships at the average level of the past three years could receive & euro;2000 for each apprenticeship starting in the autumn of 2020. We exploit eligibility for the subsidy due to firm size and compare eligible and non-eligible firms in an intent-to-treat difference-in-differences (DiD) design. The estimates show no impact on the number of new apprenticeships but a moderate effect on the probability to maintain the training level at the level of the past three years. This suggests that firms that were at the margin of maintaining the level without the subsidy have hired a small number of additional apprentices to receive the subsidy. This mirrors the incentive structure of the subsidy, as firms at the margin have received a large amount of money for a small number of additional apprenticeships. These results are robust to a wide range of sensitivity and robustness checks, including linear and non-linear DiD estimation, tests for diverging trends, placebo tests with placebo outcomes as well as checks for panel attrition. These results add to the scarce literature on the impact of training subsidies in general and their impact during economic crises in particular.
The emergence of artificial intelligence (AI) has heightened interest in personalizing computer assisted learning (CAL) programs to tailor their instruction to individual students. Despite the proliferation of AI-driven CAL programs, evidence for their effectiveness remains limited. We present findings from a field experiment in rural China examining whether an AI-driven personalized CAL program improves student achievement. We randomly assign 8,647 students from 315 primary school classes to one of three treatment arms: (i) AI-CAL, (ii) non-personalized CAL (active control), and (iii) non-CAL educational activities (pure control). Results indicate that AI-CAL does not significantly improve achievement, with estimates precise enough to rule out non-trivial positive effects. This finding holds across the difficulty of assessment items and across the baseline achievement of students. If anything, AI-CAL might be less effective than non-personalized CAL for students in the middle tercile of baseline achievement. These findings call for caution when scaling AI-driven educational technologies.
