The number of respondents who included salary is almost half of the overall sample size. So it could just be that developers who use tabs make more and are less

The number of respondents who included salary is almost half of the overall sample size. So it could just be that developers who use tabs make more and are less likely to hand out their salary.