Administrative Data Linking and Statistical Power Problems in Randomized Experiments

Tahamont, Sarah; Jelveh, Zubin; Chalfin, Aaron; Yan, Shi; Hansen, Benjamin

doi:10.3386/w25657

Cited by 3 publications

(2 citation statements)

References 116 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tahamont et al. (2020) show that minimizing the sum of false positive and false negative error rates minimizes the bias introduced by linking errors when linking administrative records to identify the presence of a binary outcome, as we do in this paper. Consequently, including the probabilistic matching tier is preferable to excluding it.…”

Section: Setting and Datamentioning

confidence: 63%

“…As might be expected, the validation procedure revealed that many of the false positives came from the probabilistic matching tier (the exact matching process had a 0.8 percent false positive error rate and 8.0 percent false negative error rate). Tahamont et al (2020) show that minimizing the sum of false positive and false negative error rates minimizes the bias introduced by linking errors when linking administrative records to identify the presence of a binary outcome, as we do in this paper. Consequently, including the probabilistic matching tier is preferable to excluding it.…”

Section: Matching Processmentioning

confidence: 65%

See 1 more Smart Citation

The Effects of Youth Employment on Crime: Evidence from New York City Lotteries

Kessler¹,

Tahamont²,

Gelber³

et al. 2022

J Policy Anal Manage

Self Cite

View full text Add to dashboard Cite

Recent policy discussions have proposed government‐guaranteed jobs, including for youth. One key potential benefit of youth employment is a reduction in criminal justice contact. Prior work on summer youth employment programs has documented little‐to‐no effect of the program on crime during the program but has found decreases in violent and other serious crimes among “at‐risk” youth in the year or two after the program. We add to this picture by studying randomized lotteries for access to the New York City Summer Youth Employment Program (SYEP), the largest such program in the United States. We link SYEP data to New York State criminal records data to investigate outcomes of 163,447 youth who participated in a SYEP lottery between 2005 and 2008. We find evidence that SYEP participation decreases arrests and convictions during the program summer, effects that are driven by the small fraction (3 percent) of SYEP youth who are at‐risk, as defined by having been arrested before the start of the program. We conclude that an important benefit of SYEPs is the contemporaneous effect during the program summer and that the effect is concentrated among individuals with prior contact with the criminal justice system.

show abstract

Section: Setting and Datamentioning

confidence: 63%

Section: Matching Processmentioning

confidence: 65%