Building the STEM pipeline: Findings of a 9-year longitudinal research project

•Quantitative ability and interest fit are relatively stable at adolescence.•They are reciprocally related from adolescence to early adulthood.•These factors predict the probability of obtaining a college degree in STEM.•Their effects and developments are similar for both men and women.

This research addresses the call for identifying the human capital for the U.S.'s future STEM workforce. Using a national dataset tracking 34,744 middle school students up to six years after their first enrollment in a four-year college, we investigated the developments of two individual difference predictors of STEM success based on the Person-Environment fit (P-E fit) model, quantitative ability and STEM interest fit, from adolescence to early adulthood. Results show that these individual difference factors are reciprocally related and thus mutually develop over time. They are relatively stable at adolescence and meaningfully predict the probability that the students obtained a college degree in STEM approximately nine years later. This finding has important theoretical implication as it helps resolve conflicting perspectives regarding the causal relationships between interests and abilities. Finally, we also found that the development and prediction of quantitative ability and interest fit are similar for both men and women, suggesting that they can be useful to identify future STEM participants at early age.