The Exemplary Work of Promising Researchers is an invited session hosted by the Division D GSC each year at the Annual Meeting. During the Exemplary Works sessions, the graduate student winners of the In-Progress Research Gala from the previous year have the opportunity to present their finalized research study and receive feedback from experts in the field of educational research.
Coming Soon! AERA 2016 Exemplary Work of Promising Researchers
AERA 2015 Exemplary Work of Promising Researchers
The AERA 2015 Exemplary Work session included presentations from four distinct graduate student scholars. In addition, Dr. Hariharan Swaminathan from the University of Connecticut served as the session's discussant. Thank you very much to Dr. Hariharan and all of the graduate students for sharing your research and expertise at the 2015 Expemplary Works session.
Exploring the Progression of Math Content Standards by Aligning Assessments to the Common Core
Angelica Rankin, Adam Reeger, Catherine Welch, and Stephen Dunbar
University of Iowa
Math content and standards have evolved along with education, and the Common Core State Standards for Mathematics (CCSSM) are the next phase in that evolution. Integrating the CCSSM into current educational procedures has caused much debate. Some stakeholders are concerned the CCSSM will cause more harm than good, while others welcome the change. This study aims to shine some light on the extent of the changes the CCSSM might bring by examining the progression of math content standards over the last 50 years through the alignment of old math assessments to the CCSSM. Initial findings suggest the CCSSM may bring some new perspectives to math education, but may not differ as drastically from current standards as perhaps intended.
Anchor Selection Embedded Mediated MIMIC Method for Understanding DIF Mechanism
Can Shao, Ying Cheng & Quinn Lathrop
University of Notre Dame
Differential item functioning (DIF) has attracted much attention from researchers because of its close linkage to test fairness. This study proposes a procedure to first detect DIF items using the Bootstrap, Visualize, and Build (BVB) method, then use mediated MIMIC model to further understand the underlying mechanism of DIF. Through our simulation study, it is found that in the first step, the procedure gives good detection in terms of anchor set and DIF item; in the second step, the Mediated MIMIC model can successfully detect the potential mediation effect with Type-I error well controlled.
An Investigation of Small Sample Equating Methods
MinJeong Shin
University of Massachusetts
Equating with small sample size can occur when a test’s population is limited, such as credentialing exams in some medical fields. The accuracy and stability of the equating becomes a concern when the sample size is not large enough. Equating may not be necessary with such a small sample size; however, if reference and new forms differ in difficulty and ability distributions of two groups of examinees who took either the reference form or the new form are different, test scores of the two forms will not be comparable unless they are equated. In other words, no equating will lead to invalid interpretations of test scores as well as inaccurate pass/fail decisions. Thus, it is crucial to consider small sample equating methods. This study investigates the applicability of the transformation method converting classical item statistics to IRT item parameters when classical statistics are available in the bank but without IRT parameters, especially in the context of IRT observed score equating with small samples. This study also employs one of the small sample equating methods from traditional equating—the chained circle-arc method. Depending on the conditions such as sample size, test length, difference in form difficulty, and ability distributions shifts, applying equating methods can produce more accurate equating results than no equating. This study will provide guidelines to apply equating methods when the sample size is small.
Observational Procedures and Markov Chain Models for Estimating the Prevalence and Incidence of a Behavior
James E. Pustejovsky
The University of Texas at Austin
Data based on direct observation of behavior are used extensively in certain areas of educational and psychological research. A number of different procedures are used to record data during direct observation, including continuous recording, momentary time sampling, and interval recording (partial interval recording and whole interval recording). Among these, interval recording has long been recognized as problematic because the mean of such data measures neither the prevalence nor the incidence of a behavior. In this project, I do two things. First, I propose a model for interval recording data based on a latent alternating Poisson process for the behavior being observed. The model can be used to recover estimates of the behavior's prevalence and incidence, using maximum likelihood methods. Second, I propose some new recording procedures that involve combining momentary time sampling and interval recording. By comparing the asymptotic efficiency of the new methods versus existing methods, I show that for certain behaviors (i.e., certain parts of the prevalence-incidence parameter space), the new methods yield more precise estimates of prevalence and/or incidence. In further work, I am planning to field-test the new procedures I've proposed. I'm also planning to study the robustness of the parameter estimates to violations of the alternating Poisson process modeling assumptions.