many researchers find that there is an asymmetric impact of executive compensation on firm performance, and provide various possible explanations for this. Essentially, they use different data segmentation methods to identify the factors that could influence the incentive effect in stock-based compensation. However, in general, these segmentations are arbitrary and exogenous, and normally they would change the original distribution of the full sample, which might invalidate the statistical tests. The QR method is one way to alleviate these problems, and thus we apply it to test the relationship between stock-based compensation and firm performance.