Working Paper: NBER ID: w19232
Authors: Jason Fletcher
Abstract: Recent evidence has suggested that popularity during high school is linked with wages during mid-life using the Wisconsin Longitudinal Study. The results were shown to be robust to a large set of individual-level heterogeneity included completed schooling, cognitive ability, and personality measures. This paper revisits this question by first replicating the results using an alternative dataset that is very similar in structure. Like the previous results, the Add Health baseline effects suggest that an additional high school friendship nomination is linked to a 2% increase in earnings around age 30. However, leveraging the unique sibling structure of the Add Health shows that sibling comparisons eliminate any associations between popularity and earnings. The findings suggest that families, rather than friends, may be the cause of the association.
Keywords: High School Popularity; Adult Earnings; Social Skills; Family Level Heterogeneity
JEL Codes: J01; J1; J3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Family Background (J12) | High School Popularity (I24) |
| Family Background (J12) | Adult Earnings (J31) |
| Completed Schooling (I23) | Adult Earnings (J31) |
| Personality Measures (D91) | Adult Earnings (J31) |
| High School Popularity (I24) | Adult Earnings (J31) |
| High School Popularity (I24) | Adult Earnings (J31) |