Design effect

Abstract: The National Comorbidity Survey Replication (NCS-R) is a survey of the prevalence and correlates of mental disorders in the US that was carried out between February 2001 and April 2003. Interviews were administered face-to-face in the homes of respondents, who were selected from a nationally representative multi-stage clustered area probability sample of households. A total of 9,282 interviews were completed in the main survey and an additional 554 short non-response interviews were completed with initial non-respondents. This paper describes the main features of the NCS-R design and field procedures, including information on fieldwork organization and procedures, sample design, weighting and considerations in the use of design-based versus model-based estimation. Empirical information is presented on non-response bias, design effect, and the trade-off between bias and efficiency in minimizing total mean-squared error of estimates by trimming weights. Copyright © 2004 Whurr Publishers Ltd.

Abstract: A simple method for comparing independent groups of clustered binary data with group-specific covariates is proposed It is based on the concepts of design effect and effective sample size widely used in sample surveys, and assumes no specific models for the intracluster correlations It can be implemented using any standard computer program for the analysis of independent binary data after a small amount of preprocessing The method is applied to a variety of problems involving clustered binary data: testing homogeneity of proportions, estimating dose-response models and testing for trend in proportions, and performing the Mantel-Haenszel chi-squared test for independence in a series of 2 x 2 tables and estimating the common odds ratio and its variance Illustrative applications of the method are also presented

Abstract: Series Editor's Introduction Acknowledgments 1. Introduction 2. Sample Design and Survey Data Types of Sampling The Nature of Survey Data A Different View of Survey Data 3. Complexity of Analyzing Survey Data Adjusting for Differential Representation: The Weight Developing the Weight by Poststratification Adjusting the Weight in a Follow-Up Survey Assessing the Loss or Gain in Precision: The Design Effect The Use of Sample Weights for Survey Data Analysis 4. Strategies for Variance Estimation Replicated Sampling: A General Approach Balanced Repeated Replication Jackknife Repeated Replication The Bootstrap Method The Taylor Series Method (Linearization) 5. Preparing for Survey Data Analysis Data Requirements for Survey Analysis Importance of Preliminary Analysis Choices of Method for Variance Estimation Available Computing Resources Creating Replicate Weights Searching for Appropriate Models for Survey Data Analysis 6. Conducting Survey Data Analysis A Strategy for Conducting Preliminary Analysis Conducting Descriptive Analysis Conducting Linear Regression Analysis Conducting Contingency Table Analysis Conducting Logistic Regression Analysis Other Logistic Regression Models Design-Based and Model-Based Analyses 7. Concluding Remarks Notes References Index About the Authors

Abstract: Current methods for estimating the accuracy of diagnostic tests require independence of the test results in the sample. However, cases in which there are multiple test results from the same patient are quite common. In such cases, estimation and inference of the accuracy of diagnostic tests must account for intracluster correlation. In the present paper, the structural components method of DeLong, DeLong, and Clarke-Pearson (1988, Biometrics 44, 837-844) is extended to the estimation of the Receiver Operating Characteristics (ROC) curve area for clustered data, incorporating the concepts of design effect and effective sample size used by Rao and Scott (1992, Biometrics 48, 577-585) for clustered binary data. Results of a Monte Carlo simulation study indicate that the size of statistical tests that assume independence is inflated in the presence of intracluster correlation. The proposed method, on the other hand, appropriately handles a wide variety of intracluster correlations, e.g., correlations between true disease statuses and between test results. In addition, the method can be applied to both continuous and ordinal test results. A strategy for estimating sample size requirements for future studies using clustered data is discussed.