13CHRONICLES OF RESEARCHERS AT SOPHIA #34Associate ProfessorDepartment of EconomicsEffects of study abroad scholarships on language abilitySince Adam Smith’s time, the field of economics has concerned itself primarily with building theory, relegating empirical analysis to a secondary status. But the tides changed in the 1990s when empirical analysis began to assert its influence. Higuchi points out that a major reason for this is the establishment of analytical methods along with the development of IT.“In economics, as well as in other social science fields, a desire has developed for immediate results closely connected to real-ity. In particular, emphasis has been placed on illuminating causal relationships between the two related factors of cause and effect, information that is invaluable for Evidence Based Policy Making (EBPM).”An example methodology comes from the field of medicine’s “randomized controlled trial (RCT)” used in clinical trials of new drugs. Patients selected as subjects are randomly divided into two groups. Their individual differences are averaged, making it pos-sible to regard the groups as having equal attributes. The drug under review is administered to only one group but the progress of the two groups is observed carefully for clear differences in the improvement of symptoms.“RCT has been used in the social sciences for around 20 years, and widely used in particular in my own field of Development Economics. For instance, to verify the effects of a new initiative expected to combat poverty, two groups are created randomly with one group participating in the initiative, the other being sim-ply observed. Since this methodology poses the significant ethical question of a government intentionally creating social inequali-ties, it is particularly difficult to implement in developed coun-tries,” Higuchi explains.As an alternative, a quasi-experimental test called “regression dis-continuity design (RDD)” can be used. To confirm the effects of an intervention applied to people whose income falls below a cer-tain standard amount, for example, one can compare the data of people who qualify for the intervention but are slightly above the threshold with people who are slightly below it. “The income ‘borderline’ subjects in the above example can be as-sumed to be essentially homogeneous despite not being random-ized. This is because they have similar living conditions and live-lihoods. The RDD results are therefore considered as reliable as those from RCT.”“It isn’t exactly within my specialization on developing countries, but I’ve recently conducted an RDD analysis in the context of Japan,” Higuchi says, referring to his research examining the im-pact studying abroad on an MEXT scholarship program has on enhancing foreign language ability. Simply comparing the lan-guage ability of students with study abroad experience versus those without it does not demonstrate the effects of study abroad. The reason has to do with the issue of self-selection – students who choose to study abroad being different from those who do not.“Study abroad is more likely for students who are motivated to learn a language and improve their ability to actively communi-cate with people from other countries. They believe they should study abroad. The difference in language proficiency between them and those who did not study abroad relates to a difference YUKIHIGUCHI
元のページ ../index.html#13