Monday, July 14, 2014
Assaf Romm writes from Israel about the newly instituted stable matching clearinghouse for psychologists, and more briefly about the new match for Israeli doctors.
First, it was good seeing you in Israel, and it seems like you were very strategic with choosing the dates for the summer school... We are fine here, even if slightly alarmed by all the alarms. Hopefully we'll reach a cease-fire soon enough.
More market design related, let me tell you about the Israeli psychologists match, which Avinatan Hassidim and I have helped redesigning over the last year. Last Thursday (July 10) was "match day" for the Israeli psychologists match. 970 registered students were matched to 52 MA programs in 12 institutions (universities and "michlalot"). Out of the 970 students that participated, 540 were matched to a program. Out of those, 314 (58.15%) got their first choice, and 518 (95.93%) got their fifth choice or better. From what we heard from programs, they were also very pleased with the process and the results, and would like to continue using the automated matching system for years to come. Furthermore, some of them specifically mentioned that using the new system allowed them stop being strategic, and receiving students that they wouldn't have received otherwise, because they were too much of a long shot (and so, not a good strategic choice under the old system).
The Israeli psychology match comes to solve a problem that the psychology departments had when competing over students for clinical psychology studies (in Israel, MA in psychology represent the first step in becoming a psychotherapist). While seats in clinical programs are in very high demand, there are two things that make the process slightly difficult. The first is the restrictions on accepting students (the state enforces quotas), and the second is the competition over top students.
This created in the past a problem of unraveling that was partially solved by a gentlemen agreement among all the psychology departments. The agreement consisted of three scheduled rounds of acceptance and wait listing, such that no department may accept a candidate in any other fashion. This system required the departments to be very strategic in offering acceptances, and often ended with a blitz of calling people from waiting lists before they were being called by others.
In May 2013 we approached the heads of the psychology departments, got them to convene and initiated a pilot run for an automated system of admissions for all the MA degrees in psychology in Israel (not including some less competitive programs, and including all the institutions that previously were part of the gentlemen agreement). After a year of designing, implementing, announcing, and finally actually managing the preferences submission process, we finally successfully reached the deadline and the pilot was a great success.
The main obstacles we encountered during the process were:
* Convincing programs that the new matching system will not be biased in favor of large/small programs.
* Convincing applicants that the system is strategyproof, and that it is better for them compared to the old system.
* Dealing with extensive courting behavior (i.e. scholarships being offered under the table, etc.).
* Supporting complicated preferences by the programs (minority quotas, several tracks within the same program).
* Couples (since there aren't many couples, we deferred dealing with couples to next year, and we will implement the algorithm suggested by Ashlagi, Braverman and Hassidim, forthcoming).
* Dealing with politics that have to do with programs approaching candidates and informally "admitting" them prior to the match (in theory, this should have almost zero influence, but in practice it creates a race of arms between the departments).
Up until now we only received a handful of emails from students who participated in the match, and it seems that in retrospect they understood the advantages of a fast and automated system, even if it wasn't clear to them in advance why they now have to submit their preferences without knowing where they were accepted. We plan on sending surveys soon to get a better feeling of the students' experience throughout this process, and possibly we will send a similar survey to the departments that participated.
This successful redesign is the second in Israel to happen this year. Earlier this year we (Slava Bronfman and Avinatan Hassidim being the main protagonists, with Arnon Afek, the current director-general of the ministry of health, and myself as side-kicks) have finished redesigning the Israeli medical internship match. If you are interested, I will write a different email explaining the details of that other match, and specifically the theoretical problems we have encountered and how we solved them."