Using a Two-Staged Propensity Score Matching Strategy and Multilevel Modeling to Estimate Treatment Effects in a Multisite Observational Study PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Using a Two-Staged Propensity Score Matching Strategy and Multilevel Modeling to Estimate Treatment Effects in a Multisite Observational Study PDF full book. Access full book title Using a Two-Staged Propensity Score Matching Strategy and Multilevel Modeling to Estimate Treatment Effects in a Multisite Observational Study by Jordan H. Rickles. Download full books in PDF and EPUB format.
Author: Jordan H. Rickles
Publisher:
ISBN:
Category :
Languages : en
Pages : 13
Get Book Here
Book Description
The study is designed to demonstrate and test the utility of the proposed two-stage matching method compared to other analytic methods traditionally employed for multisite observational studies. More specifically, the study addresses the following research questions: (1) How do different specifications of the matching method influence covariate balance? (2) How do different specifications in the matching method influence inferences about treatment effect and effect heterogeneity? The different matching method specifications include differences in the propensity score model and whether a between-site match, within-site match, or two-stage matching process is used. The simulation results indicate that the two-stage matching method balances the desire for within-site covariate balance and the desire to retain as many treatment units in the analysis as possible. Relative to more straightforward matching methods, however, the two-stage matching method does not result in greater covariate balance nor less biased effect estimation. As a result, more straightforward methods that address the nested data structure--such as within-site matching or pooled matching with a random-intercept-and-slope propensity score model--might be preferable to the more complex two-stage matching method. These conclusions are based on a finite set of data generating conditions, with a small set of important confounders at both the unit and site level and a reasonable within-site sample size for matching. Future research should examine the performance of various propensity score model and matching methods under more extreme data conditions. (Contains 2 tables and 5 figures.).
Author: Jordan H. Rickles
Publisher:
ISBN:
Category :
Languages : en
Pages : 13
Get Book Here
Book Description
The study is designed to demonstrate and test the utility of the proposed two-stage matching method compared to other analytic methods traditionally employed for multisite observational studies. More specifically, the study addresses the following research questions: (1) How do different specifications of the matching method influence covariate balance? (2) How do different specifications in the matching method influence inferences about treatment effect and effect heterogeneity? The different matching method specifications include differences in the propensity score model and whether a between-site match, within-site match, or two-stage matching process is used. The simulation results indicate that the two-stage matching method balances the desire for within-site covariate balance and the desire to retain as many treatment units in the analysis as possible. Relative to more straightforward matching methods, however, the two-stage matching method does not result in greater covariate balance nor less biased effect estimation. As a result, more straightforward methods that address the nested data structure--such as within-site matching or pooled matching with a random-intercept-and-slope propensity score model--might be preferable to the more complex two-stage matching method. These conclusions are based on a finite set of data generating conditions, with a small set of important confounders at both the unit and site level and a reasonable within-site sample size for matching. Future research should examine the performance of various propensity score model and matching methods under more extreme data conditions. (Contains 2 tables and 5 figures.).
Author: Jordan Harry Rickles
Publisher:
ISBN:
Category :
Languages : en
Pages : 208
Get Book Here
Book Description
In this study I present, demonstrate, and test a method that extends the Stuart and Rubin (2008) multiple control group matching strategy to a multisite setting. Three primary phases define the proposed method: (1) a design phase, in which one uses a two-stage matching strategy to construct treatment and control groups that are well balanced along both unit- and site-level key pretreatment covariates; (2) an adjustment phase, in which the observed outcomes for non-local control group matches are adjusted to account for differences in the local and non-local matched control units; and (3) an analysis phase, in which one estimates average causal effects for the treated units and investigates heterogeneity in causal effects through multilevel modeling. The main novelty of the proposed method occurs in the design phase, where propensity score matching is executed in two stages. In the first stage, treatment units are matched to control units within the same site. In the second stage, treatment units without an acceptable within-site match are matched to control units in another site (between-site match). The two-stage matching method provides researchers with an alternative to strict within-site matching or matching that ignores the nested data structure (pooled matching). I employ an empirical illustration and a set of simulation studies to test the utility and feasibility of the proposed two-stage matching method. The results document the two-stage matching method's conceptual appeal, but indicate that effect estimation under the two-stage matching method does not, in general, outperform more traditional matching-based or regression-based methods. Alternative specifications within the proposed method can improve performance of two-stage matching. In addition to extending the work of Stuart and Rubin, this study complements the small set of studies that have examined propensity score matching in multisite settings and provides guidance for researchers looking to estimate treatment effects from a multisite observational study. The dissertation concludes with directions for future research and considerations for researchers conducting multisite observational studies.
Author: Ann A. O'Connell
Publisher: IAP
ISBN: 164802873X
Category : Education
Languages : en
Pages : 645
Get Book Here
Book Description
Multilevel Modeling Methods with Introductory and Advanced Applications provides a cogent and comprehensive introduction to the area of multilevel modeling for methodological and applied researchers as well as advanced graduate students. The book is designed to be able to serve as a textbook for a one or two semester course in multilevel modeling. The topics of the seventeen chapters range from basic to advanced, yet each chapter is designed to be able to stand alone as an instructional unit on its respective topic, with an emphasis on application and interpretation. In addition to covering foundational topics on the use of multilevel models for organizational and longitudinal research, the book includes chapters on more advanced extensions and applications, such as cross-classified random effects models, non-linear growth models, mixed effects location scale models, logistic, ordinal, and Poisson models, and multilevel mediation. In addition, the volume includes chapters addressing some of the most important design and analytic issues including missing data, power analyses, causal inference, model fit, and measurement issues. Finally, the volume includes chapters addressing special topics such as using large-scale complex sample datasets, and reporting the results of multilevel designs. Each chapter contains a section called Try This!, which poses a structured data problem for the reader. We have linked our book to a website (http://modeling.uconn.edu) containing data for the Try This! section, creating an opportunity for readers to learn by doing. The inclusion of the Try This! problems, data, and sample code eases the burden for instructors, who must continually search for class examples and homework problems. In addition, each chapter provides recommendations for additional methodological and applied readings.
Author: Scott F. Kosten
Publisher:
ISBN:
Category : Statistics
Languages : en
Pages : 236
Get Book Here
Book Description
Estimating the treatment effect between a treatment group and a control group in an observational study is a challenging problem in statistics. Without random assignment of subjects, there are likely to be differences between the treatment group and control group on a set of baseline covariates. If one of these baseline covariates is correlated to the response variable, then the difference in sample means between the groups is likely to be a biased estimate of the true treatment effect. Propensity score matching has become an increasingly popular strategy for reducing bias in estimates of the treatment effect. This reduction in bias is accomplished by identifying a subset of the original control group, which is similar to the treatment group in terms of the measured baseline covariates. Our research focused on the development of a new procedure that combines propensity score matching and a rank-based analysis of the general linear model. Our procedure was compared to several others in a Monte Carlo simulation study. Overall, our procedure produced highly efficient and robust confidence intervals for a treatment effect in an observational study. In addition to the Monte Carlo simulation study, our procedure and several other propensity score matching techniques were used to analyze two real world datasets for the presence of a treatment effect.
Author: Chih-Lin Li (Statistician)
Publisher:
ISBN:
Category :
Languages : en
Pages : 110
Get Book Here
Book Description
Chapter 3 focuses on longitudinal observational studies, where the same subjects are followed-up at multiple occasions. In longitudinal studies, the selection of treatment changes over time, and typically depends on the previous outcomes and covariates, which in turn may be correlated with later outcomes. We propose a propensity score matching approach for this setup. Our approach allows estimation of treatment effects at each time point while controlling for time-varying confounders. We show through simulation studies that the matching method is more robust to model misspecification than two other model-based methods. The matching method is also demonstrated using data from a study of multimodal treatment for children with attention deficit hyperactivity disorder.
Author: Erinn M. Hade
Publisher:
ISBN:
Category :
Languages : en
Pages : 96
Get Book Here
Book Description
Abstract: In medical and public health research, many studies are observational. In these studies, the treatment is not randomly assigned to participants. Therefore, the dif- ferences in outcomes between treatment groups could be due to imbalances of char- acteristics that are related to the outcome of interest prior to the treatment. Herein we investigate how we can use propensity scores, the conditional probability of re- ceiving treatment given the observed information, to make valid causal inference in observational studies. Theoretical results for the bias are given for linear response models that use the propensity score as a linear covariate. The bias depends on the relationship between the propensity score, the treatment indicator and the functional form of the covariate. Various methods for estimation of the treatment effect are explored. We show that the bias is influenced by the overlap in the distributions and functional forms of the covariates. When the distributions of the covariates have substantial overlap between treated and control groups, matching does well in terms of bias. In the second half of our work, we continue to investigate propensity score meth- ods for causal inference in observational studies, however our focus turns to studies with multiple groups. These methods are motivated by an example from the Pre- maturity Prevention Clinic at The Ohio State University. Our innovation relies on matching triplets of patients, which includes one patient from each of our groups of interest (those treated on-time, those with delayed treatment, and those who never were treated with 17P). Within each of these triplets, we attempt to balance pre- treatment characteristics by two matching techniques. We investigate two match- ing algorithms via simulation. Theese simulation studies found that a sub-optimal ii matching approach will in most circumstances provide better overall matches, than a nearest-neighbor approach. We implement our sub-optimal triplet matching for our motivating data and provide some conclusions about these data and future work with these methods.
Author: Yu-Sung Su
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
The fundamental problem of causal inference is that an individual cannot be simultaneously observed in both the treatment and control states (Holland 1986). Propensity score methods that compare the treatment and control groups by discarding the unmatched units are now widely used to deal with this problem. Propensity score matching works well when using individual level data (persons, countries, counties, etc.); however, when using data that have a multilevel structure, such as time-series-cross-sectional (TSCS) data we need to combine propensity score matching procedures with multilevel modeling in order to take into account the unique structure of the data. In this paper we conduct Monte Carlo simulations with 36 different scenarios to test the performance of the two combined methods. The result shows that combining propensity score methods with multilevel modeling yields less biased and more efficient estimates. Two empirical case studies that reexamine the relationship between democratization and development and democracy and militarized interstate disputes also show the advantage of combining these two methods.
Author: Walter Leite
Publisher: SAGE Publications
ISBN: 1483313395
Category : Social Science
Languages : en
Pages : 225
Get Book Here
Book Description
Practical Propensity Score Methods Using R by Walter Leite is a practical book that uses a step-by-step analysis of realistic examples to help students understand the theory and code for implementing propensity score analysis with the R statistical language. With a comparison of both well-established and cutting-edge propensity score methods, the text highlights where solid guidelines exist to support best practices and where there is scarcity of research. Readers will find that this scaffolded approach to R and the book’s free online resources help them apply the text’s concepts to the analysis of their own data.
Author: Wei Pan
Publisher: Guilford Publications
ISBN: 1462519490
Category : Psychology
Languages : en
Pages : 417
Get Book Here
Book Description
This book is designed to help researchers better design and analyze observational data from quasi-experimental studies and improve the validity of research on causal claims. It provides clear guidance on the use of different propensity score analysis (PSA) methods, from the fundamentals to complex, cutting-edge techniques. Experts in the field introduce underlying concepts and current issues and review relevant software programs for PSA. The book addresses the steps in propensity score estimation, including the use of generalized boosted models, how to identify which matching methods work best with specific types of data, and the evaluation of balance results on key background covariates after matching. Also covered are applications of PSA with complex data, working with missing data, controlling for unobserved confounding, and the extension of PSA to prognostic score analysis for causal inference. User-friendly features include statistical program codes and application examples. Data and software code for the examples are available at the companion website (www.guilford.com/pan-materials).
Author: Sungur Gurel
Publisher:
ISBN:
Category :
Languages : en
Pages : 61
Get Book Here
Book Description
We investigated the performance of four different propensity score (PS) methods to reduce selection bias in estimates of the average treatment effect (ATE) in observational studies: inverse probability of treatment weighting (IPTW), truncated inverse probability of treatment weighting (TIPTW), optimal full propensity score matching (OFPSM), and propensity score stratification (PSS). We compared these methods in combination with three methods of standard error estimation: weighted least squares regression (WLS), Taylor series linearization (TSL), and jackknife (JK). We conducted a Monte Carlo Simulation study manipulating the number of subjects and the ratio of treated to total sample size. The results indicated that IPTW and OFPSM methods removed almost all of the bias while TIPTW and PSS removed about 90% of the bias. Some of TSL and JK standard errors were acceptable, some marginally overestimated, and some moderately overestimated. For the lower ratio of treated on sample sizes, all of the WLS standard errors were strongly underestimated, as designs get balanced, the underestimation gets less serious. Especially for the OFPSM, all of the TSL and JK standard errors were overestimated and WLS standard errors under estimated under all simulated conditions.
