by Van Zyl, A.S.A.J.; Unpublished doctoral dissertation, University of Pretoria, South Africa. on 1991

This study was conducted in South Africa and results were published as part of a doctoral dissertation in 1991. Eighty students in fifth through tenth grades were a part of the treatment group and participated in weekly after-school chess clubs at ten different schools. The control group was composed of eighty students who did not participate in chess clubs and were matched to treatment group peers based on total IQ score, mathematics performance, and first language performance. Comparing the marks that students in the treatment and control groups received in mathematics, the study found that the effect of the chess intervention on mathematics marks was 0.640 standard deviation units. The results were statistically significant. The study was eligible for inclusion in the systematic literature review and categorized as Tier I. While students were not randomly assigned to treatment and control groups, students were matched based on student characteristics and tests of baseline performance indicated that the groups were comparable.

by Sallon, S.; Unpublished dissertation, Manchester Metropolitan University. on 2013

This dissertation, completed in 2013, evaluated the effects of an in-school chess intervention on second grade students in England. Treatment group students (n=201) participated in a total of 30 hours of chess instruction. Schools in which chess was taught to students at this age were matched, based on student characteristics, with schools that did not offer chess to their students or taught chess in a later grade. Students in both groups completed an author-designed math quiz composed of 19 math and reasoning items. The quiz specifically sought to measure student ability in the areas of numeracy spatial awareness, logical deduction, and problem solving. The results of independent t-tests comparing the mean performance of students in treatment schools to that of students in control schools revealed significant differences on each of the four areas of mathematical ability measured by the quiz. The overall effect size across these areas was 0.515 standard deviation units. The results were statistically significant. This study was eligible for inclusion in the systematic literature review and categorized as a Tier I study. The randomization of schools into the treatment and control groups provides confidence that the higher levels of mathematics achievement observed among treatment group students is attributable to the chess intervention.

by Hermelin, R.; Unpublished master’s thesis, University of KwaZulu-Natal, Durban, South Africa. on 2004

In this 2004 master’s thesis, Hermelin evaluated the effects of an after-school chess club on South African students’ end-of-course math grades. The sample was comprised of 76 grade 5-7 students, with an equal number of students in the treatment and control groups. Treatment group students, who self-selected to participate in the chess club, were matched with non-chess club classmates based on grade level and similar prior year end-of-course math grades. The author uses a test of statistical significance to determine if observed mean differences are statistically significant, and found an effect size of 0.840 standard deviation units. The findings were statistically significant. The study was eligible for inclusion in the systematic literature review and categorized as a Tier II study. While the study does use a quasi-experimental matching design, it does not show group equivalence on pretest results. Because the analysis does not control for student-level characteristics beyond matching on grade-level and past math achievement, exclusion of other variables may bias results.

by Garcia, N.V.; Unpublished doctoral dissertation, University of Houston. on 2008

Conducted during the 2006-2007 school year, this dissertation evaluated the impact of a weekly after-school chess club intervention on Hispanic students’ performance on the reading and math sections of the Texas Assessment of Knowledge and Skills (TAKS). The author randomly selected 27 treatment students from a pool of chess club students across five different elementary schools in a predominantly Hispanic district in Texas; 27 control group students were selected from the same schools. Treatment students self-selected to participate in the chess club, and control students did not play chess—it is unknown if control students participated in any other after-school activity. Using ANCOVA, the study found an effect size of 1.455 standard deviation units in math and 1.436 standard deviations in reading. The results were statistically significant. The study was eligible for inclusion in the systematic literature review and categorized as a Tier III study. The sampling technique does not qualify as random assignment, because students self-selected to participant in the intervention. If there are systematic differences in student characteristics between the two groups, results may be biased. The study focuses on Hispanic students which limits its generalizability to other student populations.

by Fried, S., & Ginsburg, N.; Unpublished manuscript. (n.d.). on

This random assignment study measured the impact of an in-school chess intervention on spatial relations analysis, measured by the Wechsler Intelligence Scale for Children-Revised (WISC-R), and behavior, measured by the Survey of School Attitudes. The assessments were administered at the end of the study period. The sample was comprised of 30 students exhibiting persistent behavior problems. Treatment students received 18 lessons over the course of one academic year, one control group received counseling, and another control group did not receive any intervention. The authors used ANOVA and tests of statistical significance to determine whether observed mean differences were statistically significant. For the students in the treatment group compared with students in the no-contact control group, an effect size of 0.070 standard deviation units was found on the cognitive assessment and an effect size of 0.103 was found on the behavior survey. However, the results were statistically insignificant suggesting no difference between treatment and control students. The study was eligible for inclusion in the systematic literature review and categorized as a Tier I study. The authors note there was random student attrition from the study but do not present statistics on the dropout rate in the treatment or control group. Because the original sample size was small, attrition further reduces the study’s ability to detect small effect sizes, which may be evidenced by the null finding of significantly different mean scores between the treatment group and control groups. The study’s focus on students with behavior problems limits it generalizability to other student populations.

by Yap, K.O.; Portland, Oregon: Northwest Regional Educational Laboratory on 2006

This 2006 report evaluated the performance of 233 students participating in an after-school chess instruction program, Chess for Success (CFS), on the mathematics and reading portions of the Oregon state test, the Coopersmith Inventory, and a student behavior scale. Treatment group student performance was compared to the performance of 88 students that did not participate in chess instruction. Based on t-test results, effect sizes were 0.276 standard deviation units in mathematics, 0.152 standard deviation units in reading, and -0.018 standard deviation units on the behavioral measures. The math results were statistically significant, while the reading and behavior results were not. The study was eligible for inclusion in the systematic literature review and categorized as Tier III. Neither students nor schools were randomly assigned to treatment or control groups, and students in the two groups were not matched on student characteristics. While treatment group baseline performance on the Coopersmith Inventory and state reading assessment were not significantly different from the control group performance on those assessments, there were statistically significant differences in the performance of treatment and control group students on the mathematics assessment at baseline. Therefore, there is reason for caution in interpreting the effects reported by the study as being the result of participation in chess instruction.

by Trinchero, R. (n.d.); Kasparov Chess Foundation Europe

Conducted in Italy, this study was designed to test the impact on an in-school chess intervention on the mathematics performance of 3rd through 7th grade students. The treatment group was composed of 412 students and the control group was composed of 156 students. Students who participated in the in-school chess intervention were also invited to complete additional chess practice through computer-assisted training online. Each group was further subdivided and a portion of students in each group took a pre-test composed of OECD-PISA mathematics items. All students completed the same assessment following the conclusion of the chess intervention. For students who took both pre- and post-intervention assessments, individual student gains in mathematics were calculated. Then, the gains of the experimental groups were compared to the gains of the control group using univariate ANOVA. The effect size in mathematics was 0.421 standard deviation units. The results were statistically significant. The study was eligible for inclusion in the systematic literature review and categorized as Tier III. Assignment to treatment and control groups was not randomized, and no matching of students in the treatment and control groups based on student characteristics was conducted.

by Thompson, M.; Issues in Educational Research, 13. on 2003

This study, published in 2003, examined the effect of participation in chess on academic success in science at a boys’ school in Australia. Of the 508 students in grades 6-12 at the school, 64 participated in the school’s competitive chess program. The students’ results on the Australian Schools Science Competition were used to evaluate the effect of the chess program. Rasch scaling was used to include students from all grade levels on the same scale. Students who participated in the chess program achieved higher results on the science test than students who did not; however, after controlling for grade level and IQ. The effect size was 0.128 standard deviation units. However, the results were statistically insignificant suggesting no difference between treatment and control students. This study was eligible for inclusion in the systematic literature review and categorized as a Tier III study. However, the lack of pre-test results for students and the self-selection of students into the chess program may have generated bias in the results.

by Sigirtmac, A. D.; Early Child Development and Care, 182(6). on 2012

Published in 2012, this study evaluated the effect of chess training on development of cognitive skills for six-year-old children in Turkey. Fifty students who had previously received chess instruction and fifty students who had not were given a concept test that evaluated their understanding of spatial concepts. The children were tested individually by the researcher. The Mann-Whitney U test indicated statistically significant differences between children who had and had not received chess instruction for all tested concepts, resulting in an effect size of 1.600 standard deviation units. The results were statistically significant. This study was eligible for inclusion in the systematic literature review and categorized as a Tier III study. While the study compares the cognitive abilities of children who have and have not received chess instruction, the lack of a pre-test means that the differences between the groups may have been caused by something other than chess, biasing the results.

by Scholz, M., Niesch, H., Steffen, O., Ernst, B., Loeffler, M., Witruk, E., & Schwarz, H.; International Journal of Special Education, 23(3), 138-148. on 2008

This study was published in 2008 and examined the effect of chess as a mathematics lesson for students with learning disabilities in four German schools. Some classes were randomly assigned to have one hour of chess lessons per week in replacement of one hour of math lessons for one academic year. Students in the treatment and control groups were given a math assessment designed by the authors at the beginning and end of the year to measure their concentration and calculation abilities. The chess lessons were designed specifically for children. Thirty-one students from the treatment group and 22 students from the control group completed the pre- and post-test and were considered in the analysis. The authors found that calculation abilities for counting and simple addition tasks improved significantly more in students receiving the chess lessons. The effect size in mathematics was 0.204 standard deviation units. However, the results were statistically insignificant suggesting no difference between treatment and control students. This study was eligible for inclusion in the systematic literature review and categorized as a Tier I study. The results may not be generalizable to students without learning disabilities, or to students with learning disabilities in larger class sizes (classes in this study had a class size of 8-13 students).