Engineer/Scientist Careers: Patents, Online Profiles, and Misclassification Bias
61 Pages Posted: 2 Dec 2014
Date Written: November 27, 2014
Through an inventor survey, we find substantial error in tracking mobility of engineers and scientists by patents. These errors cause misclassification of mobility -- false positives (wrongly recording change of employer), and false negatives (failing to record change of employer). Errors are higher among inventors with shorter patent careers.
Econometric theory shows that misclassification of the dependent variable causes systematic bias in regression estimates (not merely attenuation). We introduce LinkedIn as a more accurate source of career data. Taking LinkedIn as a benchmark, the rate of false positives in patent measures of mobility is 12 percent, while the rate of false negatives is 83 percent.
Measuring mobility by LinkedIn, we investigate the effect of aspects of human capital previously found to affect mobility. One previous finding is robust: that mobility is higher among Silicon Valley inventors than those elsewhere. Other findings are sensitive to sample or misclassification. We interpret our results as the outcome of targeted retention of human capital.
Keywords: Engineer/scientist careers; patents; LinkedIn; misclassification
JEL Classification: O32
Suggested Citation: Suggested Citation