Bayesian methods seem a natural choice for combining sources of evidence in meta-analyses. However, in practice, their sensitivity to the choice of prior distribution is much less attractive, particularly for parameters describing heterogeneity. A recent non-Bayesian approach to fixed-effects meta-analysis provides novel ways to think about…

Post-instrument covariates are often included as controls in instrumental variable (IV) analyses to address a violation of the exclusion restriction. However, we show that such analyses are subject to biases unless strong assumptions hold. Using linear constant-effects models, we present asymptotic bias formulas for three estimators (with and…

There exists a variety of interval estimators for the overall treatment effect in a random-effects meta-analysis. A recent literature review summarizing existing methods suggested that in most situations, the Hartung-Knapp/Sidik-Jonkman (HKSJ) method was preferable. However, a quantitative comparison of those methods in a common simulation study…

While the use of inferential statistics is a nearly universal practice in the social sciences, there are instances where its application is unnecessary and potentially misleading. This is true for a portion of research using administrative data in educational research in the United States. Surveying all research articles using administrative data…

The ability of a scientific discipline to build cumulative knowledge depends on its predominant method of data analysis. A steady accumulation of knowledge requires approaches which allow researchers to consider results from comparable prior research. Bayesian statistics is especially relevant for establishing a cumulative scientific discipline,…

Researchers may have multiple motivations for combining disparate pieces of evidence in a meta-analysis, such as generalizing experimental results or increasing the power to detect an effect that a single study is not able to detect. However, while in meta-analysis, the main question may be simple, the structure of evidence available to answer it…

This working paper contains the overheads used in a seminar designed to introduce some basic concepts of statistics to nonstatisticians. The seminar has been presented on several occasions. The first part of the seminar, and the first set of overheads, deals with the essentials of statistics, including: (1) population, sample, and inference; (2)…

Presents an opinion on the appropriate use of significance tests, especially in the context of regression analysis, the most commonly encountered statistical technique in education and related disciplines. Briefly discusses the appropriate use of power analysis. Contains 47 references. (Author/SV)

A solid understanding of "inferential statistics" is of major importance for designing and interpreting empirical results in any scientific discipline. However, students are prone to many misconceptions regarding this topic. This article structurally summarizes and describes these misconceptions by presenting a systematic review of publications…

Presents an overview of a wide range of measurement models currently available to the analyst who needs to make accurate and valid inferences about respondents and stimuli from data. Reviews models with and without predictor variables or observed and latent predictors, as well as parametric and nonparametric models, and models for order-restricted…

Hierarchical models are extremely promising tools for data analysis, but it is important not to lessen hard thinking about data and iterative model checking when fitting hierarchical models. More and better software, methods to assure proper calibration, and materials in support of hierarchical model use are all needed. (SLD)

Examines potential problems in the use of inferential statistics for single population proportions, differences between two population proportions, and quotients of two population proportions. Discusses hypothesis testing versus interval estimation. Emphasizes the importance of selecting the appropriate formula for the standard error and…

Contrasts aptitude by treatment interaction (ATI) and hierarchical linear modeling (HLM) methods for making cross-level inferences between individual-level and group-level factors in school effectiveness research. Recommends HLM when intraclass correlations are high. ATI is suitable when intraclass correlations are low, but partitioning the…

Three versions of a probabilistic model adapted from the theory of information retrieval--a binary version, a version using the full value of the data, and a version using principal components--were tested and applied to data available from application forms to predict graduate school performance of library school students. (EM)