Understanding the crucial distinction between correlation and causation is an intellectual exercise whose importance cannot be emphasized enough and whose misunderstanding can lead to catastrophic policy recommendations. This has been a concern of many scientific discussions in fields like economics, sociology, psychology and medicine. Its arrival within the field of higher education research should be received with great satisfaction.
A recent paper by Kratz and Netz (2017) can be used to produce a neat illustration of
this point. These authors pursue in this work an examination of the benefits of
international student mobility (ISM) and, in particular, they set themselves to determine
the relative impact on future wages enjoyed by students enrolled in exchange programs
like ERASMUS. Any evaluation of these schemes is well justified given the vast amounts
of resources that many countries devote to their implementation. This may seem an
easy task given the large availability of data reporting the wages that students with and
without ISM-experience receive after graduation. Indeed, if there is a statistically and
economically significant wage gap in favor of the former, we can tap ourselves on the
shoulder and declare the task completed. And here it is where the above important
distinction fully kicks in. Before we run to congratulate policy-makers avid to hear the
good news, we need to ask ourselves the following counterfactual question: Would the
very same students that participated in these programs have accomplished similar
monetary gains anyway, had they not participated in ISM-programs? Answering these
question turns out to be an enormously difficult endeavor for which sophisticated
econometric models and not so easily available data is usually required. The challenge
derives from the possibility that those personal traits that are rewarded favorably in the
job market are precisely the typical ones of those students inclined to travel abroad to
study.
Kratz and Netz (2017) have at their disposal individual data comprising variables
capturing important aspects like social background and educational performance and
mobility experience prior to higher education for a large sample of German students.
Endowed with this information the authors first confirm that indeed “selection effects”
are present since those students with an academic family and previous mobility
experience are more likely to be found among the ranks of ISM-programs. In addition,
they use this information to produce additional variables so that their statistical models
can account for this potential source of biasedness.
All concerns in this direction are not removed as the authors themselves admit and their
efforts end up raising more questions than answers. Among their findings, they
document that international students display higher language skills (a competence
highly appreciated in the job market) and a higher inclination to pursue and complete a
doctorate (does this really pay off?). Furthermore, they establish that ISM experienced
workers tend to leverage a higher propensity to change jobs into higher salaries and
that they can be found with a higher likelihood as employees of international companies
which in turn pay higher salaries. However, they cannot safely attribute these patterns
to previous enrollment in ISM programs. As a result, their call to devote further research
to this issue is clearly understandable.
Text: Iñaki Rodriguez Longarela
Kratz, F., and Netz, N. (2018) Which mechanisms explain monetary returns to
international student mobility? Studies in Higher Education 43:2, 375-400.