In addition, we control for individual characteristics and hospital fixed effects.
The effect becomes insignificant after controlling for individual characteristics in model 3.
Finally, the results in column (5) control for individual characteristics, primary diagnosis, and hospital fixed effects.
Our results confirm the existence of an exporter wage premium after controlling for individual characteristics of workers.
Second, it would be interesting to include direct measures of the outside wage rather than controlling for individual characteristics.
Finally, we analyze the robustness of the estimated school contribution to the achievement gap to controlling for individual characteristics.
The estimated effect becomes smaller but stays positive when we control for individual characteristics and hospital characteristics.
In the last specification we analyze the contribution of all observable school characteristics after controlling for individual characteristics.
Without controlling for individual characteristics, estimated returns to ICT skills are at 27%, significant at the 11% level [column (3)].
Therefore, heuristics suggest that one should look for a set of characteristics, and not for individual characteristics.
The estimates in column (5) control for individual characteristics, the full diagnosis vector, and hospital fixed effects, which is our preferred set of controls.
