International trade has rapidly increased in the past decades, affecting production and labor demand across various economic sectors. The impact of trade on employment and welfare relies heavily on data about worker reallocation, which often contains coding errors. This study demonstrates that such errors bias the estimated effects of trade and structural parameters in standard models. An econometric framework is developed to estimate misclassification probabilities, correct mobility matrices, and structural parameters. The findings reveal that the true effects of trade shocks differ significantly from those estimated using uncorrected data, highlighting the importance of addressing coding errors in economic analyses.