Meta-analysis as a method for integrating results of studies about effects of practice and coaching on test scores


Division of Educational Measurement, Department of Education, University of Umeå, S-901 87, Umeå, Sweden.


The effects of practice and coaching on test scores have been investigated in numerous studies. The studies, as well as their focus, have had considerable variations and, as a consequence, the results obtained manifest extreme variations, which in several cases depend on the test. Focusing on one test, the SAT, and meta-analysis as a tool for a synthesis of the effects obtained, the question is raised whether the meta-analyses carried out regarding the effects of practice and coaching on the SAT have resulted in the same conclusions. The summarised results of these analyses indicate that meta-analysis cannot fully answer the question whether, and under what circumstances, the SAT is susceptible to practice and coaching. Different meta-analyses of nearly the same studies have yielded different conclusions. The reasons emphasised are the test itself, the definition of coaching and practice and various types of methodological flaws regarding both the studies incorporated as well as the meta-analyses themselves. From the obtained results, suggestions are made aiming at a more proper application and use of meta-analyses as a tool for research synthesis.