Over the past ten years, propensity score methods have made an important contribution to improving generalizations from studies that do not select samples randomly from a population of inference. However, these methods require assumptions and recent work has considered the role of bounding approaches that provide a range of treatment impact…

Recently Burkholder "et al." argued that class normalized gains over the entire population of courses is approximated by a Cauchy distribution, not by a normal distribution, and therefore should not be used to compare different classes because means and standard deviations cannot be calculated. They argued that multiple linear regression…

Understanding exam score distributions has implications for item response theory (IRT), grade curving, and downstream modeling tasks such as peer grading. Historically, grades have been assumed to be normally distributed, and to this day the normal is the ubiquitous choice for modeling exam scores. While this is a good assumption for tests…

When students with the capacity to succeed in a 4-year college do not take a college admission test, this represents a potential loss of opportunity for students and colleges alike. However, the costs of testing--both pecuniary and nonpecuniary--may exceed the benefits for students who lack the interest in or qualifications for college attendance.…

Test score distributions of schools or demographic groups are often summarized by frequencies of students scoring in a small number of ordered proficiency categories. We show that heteroskedastic ordered probit (HETOP) models can be used to estimate means and standard deviations of multiple groups' test score distributions from such data. Because…

This Brief compares three hypothetical Quality Rating and Improvement Systems (QRIS) that use different rating structures: block, points, and hybrid. Because the quality standards in the hypothetical QRIS are held relatively constant across structures, analyses can be conducted to determine how structure relates to key QRIS outcomes. Three…

Regression methods can locate student test scores in a conditional distribution, given past scores. This article contrasts and clarifies two approaches to describing these locations in terms of readily interpretable percentile ranks or "conditional status percentile ranks." The first is Betebenner's quantile regression approach that results in…

In Brazil, the mean of math test scores for students of the fourth grade declined by approximately 0.2 standard deviation in the late 1990s. However, the potential changes in the distribution of scores have never been addressed. It is unclear if the decline was caused by deterioration in student performance levels at the upper and/or lower tails…

Domestic attempts to use financial incentives for teachers to increase student achievement have been ineffective. In this paper, we demonstrate that exploiting the power of loss aversion--teachers are paid in advance and asked to give back the money if their students do not improve sufficiently--increases math test scores between 0.201 (0.076) and…

The continued accumulation of validity evidence for the intended uses of educational assessment scores is critical to ensure that inferences made using the scores are sound. To that end, the College Board has continued to collect college outcome data to evaluate the relationship between SAT® scores and college success. This report provides updated…

Recognizing the differences between three discrete distributions (Binomial, Hypergeometric and Negative Binomial) can be challenging for students. We present an activity designed to help students differentiate among these distributions. In addition, we present assessment results in the form of pre- and post-tests that were designed to assess the…

Equating under the external anchor design is frequently conducted using scaled scores on the anchor test. However, scaled scores often lead to the unique problem of creating zero frequencies in the score distribution because there may not always be a one-to-one correspondence between raw and scaled scores. For example, raw scores of 17 and 18 may…

This report summarizes five-year global GMAT testing trends and includes: (1) GMAT exams taken by residence; (2) GMAT exams taken by gender; (3) Mean age of GMAT examinees; (4) Mean GMAT Total score; and (5) GMAT score-sending breakdowns by program type (MBA, non-MBA master's, and doctoral/other), TY 2014. This data brief can be used to help build…

In this study, eight statistical strategies were evaluated for selecting the parameterizations of loglinear models for smoothing the bivariate test score distributions used in nonequivalent groups with anchor test (NEAT) equating. Four of the strategies were based on significance tests of chi-square statistics (Likelihood Ratio, Pearson,…

New York City's elite public specialized high schools have a long history of offering a rigorous college preparatory education to the City's most academically talented students. Though immensely popular and highly selective, their policy of admitting students on the basis of a single entrance exam has been heavily criticized. Many argue, for…