Quasi Experimentation

Featuring engaging examples from diverse disciplines, this book explains how to use modern approaches to quasi-experimentation to derive credible estimates of treatment effects under the demanding constraints of field settings. Foremost expert Charles S. Reichardt provides an in-depth examination of the design and statistical analysis of pretest-posttest, nonequivalent groups, regression discontinuity, and interrupted time-series designs. He details their relative strengths and weaknesses and offers practical advice about their use. Reichardt compares quasi-experiments to randomized experiments and discusses when and why the former might be a better choice. Modern moethods for elaborating a research design to remove bias from estimates of treatment effects are described, as are tactics for dealing with missing data and noncompliance with treatment assignment. Throughout, mathematical equations are translated into words to enhance accessibility.

We shall examine the validity of 16 experimental designs against 12 common threats to valid inference. By experiment we refer to that portion of research in which variables are manipulated and their effects upon other variables observed. It is well to distinguish the particular role of this chapter. It is not a chapter on experimental design in the Fisher (1925, 1935) tradition, in which an experimenter having complete mastery can schedule treatments and measurements for optimal statistical efficiency, with complexity of design emerging only from that goal of efficiency. Insofar as the designs discussed in the present chapter become complex, it is because of the intransigency of the environment: because, that is, of the experimenter’s lack of complete control.

This book presents some quasi-experimental designs and design features that can be used in many social research settings. The designs serve to probe causal hypotheses about a wide variety of substantive issues in both basic and applied research. Each design is assessed in terms of four types of validity, with special stress on internal validity. Although general conclusions are drawn about the strengths and limitations of each design, emphasis is also placed on the fact that the relevant threats to valid inference are specific to each research setting. Consequently, a threat that is usually associated with a particular design need not invariably be associated with that design.

The role of group research designs to evaluate social work practice -- Pre-experimental group research designs -- Quasi-experimental group research designs -- Time-series research designs -- Evaluating and reporting quasi-experimental studies.

The contributors to Best Practices in Quantitative Methods envision quantitative methods in the 21st century, identify the best practices, and, where possible, demonstrate the superiority of their recommendations empirically. Editor Jason W. Osborne designed this book with the goal of providing readers with the most effective, evidence-based, modern quantitative methods and quantitative data analysis across the social and behavioral sciences. The text is divided into five main sections covering select best practices in Measurement, Research Design, Basics of Data Analysis, Quantitative Methods, and Advanced Quantitative Methods. Each chapter contains a current and expansive review of the literature, a case for best practices in terms of method, outcomes, inferences, etc., and broad-ranging examples along with any empirical evidence to show why certain techniques are better. Key Features: Describes important implicit knowledge to readers: The chapters in this volume explain the important details of seemingly mundane aspects of quantitative research, making them accessible to readers and demonstrating why it is important to pay attention to these details. Compares and contrasts analytic techniques: The book examines instances where there are multiple options for doing things, and make recommendations as to what is the "best" choice—or choices, as what is best often depends on the circumstances. Offers new procedures to update and explicate traditional techniques: The featured scholars present and explain new options for data analysis, discussing the advantages and disadvantages of the new procedures in depth, describing how to perform them, and demonstrating their use. Intended Audience: Representing the vanguard of research methods for the 21st century, this book is an invaluable resource for graduate students and researchers who want a comprehensive, authoritative resource for practical and sound advice from leading experts in quantitative methods.

Sections include: experiments and generalised causal inference; statistical conclusion validity and internal validity; construct validity and external validity; quasi-experimental designs that either lack a control group or lack pretest observations on the outcome; quasi-experimental designs that use both control groups and pretests; quasi-experiments: interrupted time-series designs; regresssion discontinuity designs; randomised experiments: rationale, designs, and conditions conducive to doing them; practical problems 1: ethics, participation recruitment and random assignment; practical problems 2: treatment implementation and attrition; generalised causal inference: a grounded theory; generalised causal inference: methods for single studies; generalised causal inference: methods for multiple studies; a critical assessment of our assumptions.

Using Propensity Scores in Quasi-Experimental Designs, by William M. Holmes, examines how propensity scores can be used to reduce bias with different kinds of quasi-experimental designs and to fix or improve broken experiments. Requiring minimal use of matrix and vector algebra, the book covers the causal assumptions of propensity score estimates and their many uses, linking these uses with analysis appropriate for different designs. Thorough coverage of bias assessment, propensity score estimation, and estimate improvement is provided, along with graphical and statistical methods for this process. Applications are included for analysis of variance and covariance, maximum likelihood and logistic regression, two-stage least squares, generalized linear regression, and general estimation equations. The examples use public data sets that have policy and programmatic relevance across a variety of disciplines.