Multisite field experiments using the (generalized) randomized block design that assign treatments to individuals within sites are common in education and the social sciences. Under this design, there are two possible estimands of interest and they differ based on whether sites or blocks have fixed or random effects. When the average treatment…

Structural equation modeling (SEM) is a statistical method widely used in educational research to investigate relationships between variables. SEM models are typically constructed based on theoretical foundations and assessed through fit indices. However, a well-fitting SEM model alone is not sufficient to verify the causal inferences underlying…

BASIE is a framework for interpreting impact estimates from evaluations. It is an alternative to null hypothesis significance testing. This guide walks researchers through the key steps of applying BASIE, including selecting prior evidence, reporting impact estimates, interpreting impact estimates, and conducting sensitivity analyses. The guide…

This comment discusses some issues about the use and reporting of experimental studies in education, illustrated by a recently published study that claimed (i) that an educational innovation was effective despite outcomes not reaching statistical significance, and (ii) that this refuted the findings of an earlier study. The two key issues raised…

Conventionally, regression discontinuity analysis contrasts a univariate regression's limits as its independent variable, "R," approaches a cut point, "c," from either side. Alternative methods target the average treatment effect in a small region around "c," at the cost of an assumption that treatment assignment,…

Nonnormality of univariate data has been extensively examined previously (Blanca et al., 2013; Micceri, 1989). However, less is known of the potential nonnormality of multivariate data although multivariate analysis is commonly used in psychological and educational research. Using univariate and multivariate skewness and kurtosis as measures of…

Multisite trials, which are being used with increasing frequency in education and evaluation research, provide an exciting opportunity for learning about how the effects of interventions or programs are distributed across sites. In particular, these studies can produce rigorous estimates of a cross-site mean effect of program assignment…

Bayesian statistical methods have become more feasible to implement with advances in computing but are not commonly used in educational research. In contrast to frequentist approaches that take hypotheses (and the associated parameters) as fixed, Bayesian methods take data as fixed and hypotheses as random. This difference means that Bayesian…

Recently, Stephen Gorard has outlined strong objections to the use of significance testing in social research. He has argued, first, that as the samples used in social research are almost always non-random it is not possible to use inferential statistical techniques and, second, that even if a truly random sample were achieved, the logic behind…

Selection bias is a persistent, and often hidden, problem in educational research. It is the primary obstacle standing in between increasingly available large education datasets and the ability to make valid causal inferences to inform policymaking, research, and practice (Stuart, 2010). This article provides an accessible discussion on the…

When fitting hierarchical regression models, maximum likelihood (ML) estimation has computational (and, for some users, philosophical) advantages compared to full Bayesian inference, but when the number of groups is small, estimates of the covariance matrix (S) of group-level varying coefficients are often degenerate. One can do better, even from…

When fitting hierarchical regression models, maximum likelihood (ML) estimation has computational (and, for some users, philosophical) advantages compared to full Bayesian inference, but when the number of groups is small, estimates of the covariance matrix [sigma] of group-level varying coefficients are often degenerate. One can do better, even…

As applications of multilevel modelling in educational research increase, researchers realize that multilevel data collected in many educational settings are often not purely nested. The most common multilevel non-nested data structure is one that involves student mobility in longitudinal studies. This article provides a methodological review of…

In randomized control trials (RCTs) in the education field, the complier average causal effect (CACE) parameter is often of policy interest, because it pertains to intervention effects for students who receive a meaningful dose of treatment services. This article uses a causal inference and instrumental variables framework to examine the…

Experimental data are analysed statistically to allow researchers to draw conclusions from a limited set of measurements. The hard fact is that researchers can never be certain that measurements from a sample will exactly reflect the properties of the entire group of possible candidates available to be studied (although using a sample is often the…