When researchers estimate labor market discrimination, they control for many observable factors that are thought to impact wages (e.g., education, occupation, experience, location). Even after accounting for these factors, they find that men are typically paid more than women, concluding that any remaining differences may indicate discrimination. Of course, it is not possible to control for all differences that may affect earnings. Consider whether this methodology overestimates or underestimates the total effect of discrimination on wages.
a. Women may choose certain jobs to avoid the impact of discrimination in higher-paying professions. If this is the case, the methodology the total effect of discrimination on wages.
b. Jobs that are dominated by women may pay less than similarly skilled jobs that are dominated by men. If this is the case, the methodology the total effect of discrimination on wages. underestimates overestimates

a. In the case when the women might select the specific job for avoiding the effect of discrimination in a profession that pay high salary so the methodology would underestimate the total impact with respect to the discrimination of the wages

b. In the case when the job i.e dominated by the women that pay lower having same skilled job dominated by men. So here the methodology also be underestimated