NLT Special Feature: Observation, Selection, and Signaling
EdLab is constantly seeking to find the keys to and analyzing the impact of disruptive forces in education. Indeed, many of the recent technological advances will serve to make education significantly cheaper. Yet there are some economic frictions that prevent disruptive and productive ideas from gaining further traction. The friction that I see as the biggest to overcome is the observability problem and what it has brought on.
Education is one of the few areas where output is very difficult to observe. It is hard to determine the value of an education at various levels. As it is, a student never enters a classroom with a blank slate. Then try teasing out the effects family, geographical location, and income have on education and you start to have a mess on your hands. Then when running value-added analyses that try to take into account for all of these factors, the difficulty relates to the dependent variable. Whether for legitimate or political reasons, many deem test scores too inaccurate and point to the incentives created for teachers to cheat the system. Keep in mind all of this is done with the goal to better observe output in education.
Move up to higher education and this observation problem starts to look worse. The inability to observe leads to gross selection effects. The whole process of who goes to what college is one big self-fulfilling prophecy. Why do students want to go to an elite school? It's considered an elite school. And why is it considered an elite school? Students want to go there. In all of this, the student only peripherally thinks about the quality of education he/she is receiving. Behind this is a hidden asymmetry of information. Colleges have all of the information regarding the quality of the student, SAT, transcript, extracurriculars, rec letters, and all. There's a name for what colleges do here...it's called cherry-picking. Is he smart because of the Columbia education he received or because he got in to Columbia? Not an impossible question to answer, and the same goes for value-added analysis of teachers, but it's still damn hard to answer and then act on.
What education means in the labor market is probably the most telling. The sad truth is that employers don't really care what you learned in college for most non-STEM professions. Employers for these professions assume that human capital is not appreciating in college. They simply assume that whatever you could do coming out of college, you could have done almost as well coming out of high school.
Now reverse engineer the process and see what the smartest of students will do in each case. They will not be focused on learning useful material in college; they can instead slack off provided their gpa is acceptable (thank you grade inflation). In high school, students are focused on the nicest college application possible. Only learning which furthers your chance of getting into a more selective school is considered valuable. This bleak picture is not too different from what the traditional educational institutions offer: a lot of expensive signaling and not a whole lot of learning.
Bringing all of this back to EdLab, I think it's important to realize that disruption of education is going to require more than just reducing the cost of learning through the posting of online educational materials for free. The cost of signaling needs to come down more than commensurately for education to be cheaper and more accessible. All of this means that even if we make educational output and human capital marginally more observable, we will be well on our way to creating a better education system for all.