The Effect of Supplementary Tutoring on Students' Mathematics Achievement: A Comparative Study of Japan and the United States

Open Access
Mori, Izumi
Graduate Program:
Educational Theory and Policy
Doctor of Philosophy
Document Type:
Date of Defense:
July 12, 2012
Committee Members:
  • Suet Ling Pong, Dissertation Advisor/Co-Advisor
  • Suet Ling Pong, Committee Chair/Co-Chair
  • David P Baker, Committee Member
  • David R Johnson, Committee Member
  • Kathryn Bancroft Hynes, Committee Member
  • Soo Yong Byun, Committee Member
  • Supplementary Tutoring
  • Shadow Education
  • Private Tutoring
  • Mathematics Achievement
  • Propensity Score Matching
Supplementary tutoring, also known as shadow education, private tutoring, or out-of-school tutoring, refers to a range of organized tutoring practices in academic subjects that occur outside regular school hours. This study used the 2006 Programme for International Student Assessment (PISA) and compared between the United States and Japan, two countries with different patterns of dominant use of supplementary tutoring. The study addressed the following three questions: (1) What factors affect students’ participation in supplementary tutoring in the United States and Japan? (2) What are the effects of supplementary tutoring on students’ mathematics achievement in the two countries? (3) Do the effects differ by student subgroups in each country? This study distinguished between two types of supplementary tutoring: out-of-school tutoring (taught by non-school teachers) and school-tutoring (taught by schoolteachers). The study used propensity score matching as an analytic strategy, which created counterfactual groups that were as similar as possible to facilitate comparison between the treated and controlled subjects. Nearest-neighbor method, stratification method, and kernel method were used along with the conventional OLS method. Regarding the background of participation, supplementary tutoring in Japan was largely represented by out-of-school tutoring as a private service, used by middle-class students for obtaining academic excellence. In contrast, supplementary tutoring in the United States was typically represented by in-school tutoring as a social service, used by low-achieving students in low-SES schools for ensuring minimum proficiency. The study obtained no statistically significant estimates of the effects of either type of tutoring in two countries. These results suggested that while the students’ opportunities to receive tutoring varied, the academic consequences of tutoring did not vary among students. Methodological issues in using propensity score methods were identified in the study, and their implications for meeting causal assumptions were discussed.