In this study, a hierarchical linear model (HLM) was used within a cluster random assignment design to test the impact of Criterion sm on students' writing skill when Criterion was used as a supplementary instructional tool in ninth grade English classrooms over a 4 week period. The study was conducted in three states: Indiana, Pennsylvania, and New Jersey. The sample consisted of 1,464 ninth graders aggregated across 10 schools, 34 teachers, and 61 classes.Findings indicate that over a short time period, Criterion use improved the mechanical aspects of student essays as measured by the analytic scores given to their essays. Students in ninth gradeEnglish classes who used Criterion several times a week over a 1 month period received higher analytic scores on their essays written at the end of the study period than those who were in classes that did not. The effect was small, but statistically significant at the .05 level using a onetailed test. These findings suggest a role for automated feedback systems in supporting writing instruction.

This study examined math achievement of elementary school students when Math Out of the Box (MTB), an inquiry-based math program, was used to supplement curriculum. The sample consisted of 767 New Jersey students in the third, fourth, and fifth grades, with approximately one third using MTB. Math achievement was measured by an assessment developed by ETS and by New Jersey's standardized test of math proficiency (NJ ASK). On the ETS assessment, a small, statistically significant difference was found in each of the three grades between students who used MTB and those who did not. On NJ ASK, a small, statistically significant difference was found in the third grade only. While these findings are an important step in establishing the efficacy of MTB, selection bias may weaken the causal inferences drawn.

The purpose of this study was to investigate the ways in which universities use the GRE® General Test scores to award merit‐based fellowships to first‐year graduate students in doctoral programs. While GRE use in fellowship award decisions is a common practice, there is very little validity evidence to support its use in this context. This exploratory qualitative study uses an argument‐based validity approach as a conceptual framework to document GRE use systematically in award decisions. The findings from this study indicate that universities use merit‐based fellowships to recruit the best and the brightest graduate students. The GRE is used as part of the nominee selection process in departments and is also used in the final selection process by university‐wide committees. The GRE scores are primarily used to infer whether students have the baseline knowledge and skills needed to succeed in graduate school. These academic competencies were one of the characteristics the decision makers identified as required for successful doctoral students. The participants stated that the factors for success in graduate school are nebulous; however, the students with high GRE scores and other noteworthy supporting materials are expected to perform strongly in graduate school and, therefore, are worthy of the investment of a merit‐based fellowship.

This study investigates the benefits of no‐stakes educational testing during students' summer vacation as a strategy to mitigate summer learning loss. Fifty‐one students in Grades 3–8 from the Every Child Valued (ECV) and Lawrence Community Center (LCC) summer programs in Lawrenceville, NJ, took short, online assessments throughout the summer, covering knowledge and procedural skills in both mathematics and reading that they had previously learned. Students were randomly assigned to mathematics or reading conditions, whereby over the course of 5 weeks, students in the mathematics condition took online assessments in mathematics fluency two or three times per week and students in the reading condition took online assessments in reading components skills (word recognition and decoding, vocabulary, morphological awareness, and reading comprehension) two or three times per week. All students completed tests in both math and reading components at the beginning and end of the summer in order to compare baseline and final performance. Although students did not show significant summer learning loss in either reading components or mathematics fluency, students in the reading condition scored significantly higher, on average, on standardized tests of reading components administered at the end of the study than students in the mathematics condition. Given the small sample size and variations in grade level, academic subjects, computer programs, and testing conditions, it is premature to draw any firm conclusions from these findings.