NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED553903
Record Type: Non-Journal
Publication Date: 2013
Pages: 132
Abstractor: As Provided
Reference Count: N/A
ISBN: 978-1-3031-2216-3
ISSN: N/A
Identifying Conditions That Support Causal Inference in Observational Studies in Education: Empirical Evidence from within Study Comparisons
Hallberg, Kelly
ProQuest LLC, Ph.D. Dissertation, Northwestern University
This dissertation is a collection of three papers that employ empirical within study comparisons (WSCs) to identify conditions that support causal inference in observational studies. WSC studies empirically estimate the extent to which a given observational study reproduces the result of a randomized clinical trial (RCT) when both share the same treatment group. Chapter 1 highlights the role of one or more waves of true pretest data for reducing bias in observational studies in education when random assignment is not possible. Drawing on data from two WSCs and one strong quasi-experimental study, I find that the bias reduction associated with conditioning on pretest measures is related to how correlated the pretest is with both selection and outcomes. If the pretest/selection correlation is weak, less bias reduction should be expected. However, across all three datasets bias never increased as a result of matching on the pretest. Chapter 2 examines the performance of two primary approaches used to examine the effect of a school level intervention when both student and school level data are available: intact school matching and matching at the student-level regardless of what school a student attends. Drawing on data from two WSCs, I found that in all cases when balance was achieved on a rich set of student and school covariates the quasi-experimental effect estimate corresponded closely with the RCT benchmark. This was true whether intact school matching or student level matching was employed. However, ignoring school level attributes entirely led to biased effect estimates in one WSC. Chapter 3 provides insights to applied education researchers who face a tradeoff between matching on observables and finding a geographically proximal match when implementing intact school matching. We examine the performance of matching on observable characteristics alone, matching only within district, and employing a hybrid approach the preferences local matches as long as they are not too dissimilar on observable characteristics. Drawing on data from an empirical application in education, we find that all three approaches reduce initial bias, but the hybrid approach leads to closest correspondence with the RCT benchmark. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]
ProQuest LLC. 789 East Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106. Tel: 800-521-0600; Web site: http://www.proquest.com/en-US/products/dissertations/individuals.shtml
Publication Type: Dissertations/Theses - Doctoral Dissertations
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A