In order to be considered “proven,” programs had to have at least one study that met the ESSA Strong, Moderate, or Promising standards of evidence, and had a significant and substantial positive outcome on achievement outcomes (an effect size of at least +0.10). Specific standards were as follows:
a. Studies had to compare students who received tutoring to a similar control group of students who did not receive tutoring. Assignment to the tutoring or control conditions could be random or matched in advance, but in practice, all but one accepted study used randomized assignment. Quasi-experimental studies that conduct the matching/weighting prior to posttest collection and during the early stage of intervention implementation, and define the experimental group as all students who received any treatment, must demonstrate comparable baseline characteristics including pretest equivalence and similar demographics.
b. Students had to be pre- and post-tested on valid achievement measures. Pretest measures for the final (analytic) sample had to be similar in tutoring and control groups (within 0.25 standard deviation units). Differential attrition had to be no more than 15%.
c. Studies had to be at least 12 weeks in duration.
d. There had to be at least 30 students in each group.
e. Outcome measures could not be ones created by developers or researchers. Tutors could not test their own students on research measures.