<p>
</p>
<p>Actually, I disagree. Income is not well understood, because it is inherently a choice variable and is therefore endogenous. To a great extent, people have some choice over whether they want to have a high-paying career or not. </p>
<p>One better way to measure ‘success’ is to still use income, but to control it for various job tracks. For example, compare only those people who entered certain professions against each other. For example, one can compare a guy who went to Stanford and became an engineer vs. a guy who went to a no-name school and become an engineer. Another way would be to survey college seniors to see what sorts of jobs they actually wanted to get vs. what jobs they actually ended up getting. </p>
<p>Granted, all of these techniques have limitations too. But less so than D&K. That doesn’t make D&K bad (for all studies have limitations), it just means that you can’t read more into the study than what it actually says.</p>