UID:
almafu_9960819773902883
Format:
1 online resource (89 pages) :
,
digital, PDF file(s).
ISBN:
1-108-85178-9
,
1-108-85239-4
,
1-108-86251-9
Series Statement:
Cambridge elements. Elements in quantitative and computational methods for the social sciences
Content:
The goal of this Element is to provide a detailed introduction to adaptive inventories, an approach to making surveys adjust to respondents' answers dynamically. This method can help survey researchers measure important latent traits or attitudes accurately while minimizing the number of questions respondents must answer. The Element provides both a theoretical overview of the method and a suite of tools and tricks for integrating it into the normal survey process. It also provides practical advice and direction on how to calibrate, evaluate, and field adaptive batteries using example batteries that measure variety of latent traits of interest to survey researchers across the social sciences.
Note:
Title from publisher's bibliographic system (viewed on 18 Jul 2022).
,
Cover -- Title Page -- Copyright Page -- Adaptive Inventories -- Contents -- 1 Introducing Adaptive Inventories -- 1.1 Motivating Example: Measuring Political Knowledge -- 1.2 AIs and Computerized Adaptive Testing -- 1.3 A Brief Introduction to CAT -- 1.4 Example: An Adaptive Measure of Political Knowledge -- 1.4.1 Setting Up an Adaptive Inventory -- 1.5 Example Applications -- 1.5.1 Adaptive Right-Wing Authoritarianism -- 1.5.2 Adaptive Need for Cognition -- 1.6 Overview of the Element -- 2 Introduction to CAT for Binary Outcomes -- 2.1 Item Response Theory for Dichotomous Survey Responses -- 2.1.1 Calculating Probabilities: The Item Response Function -- 2.1.2 Likelihood -- 2.1.3 Prior and Posterior -- 2.2 Estimating the Respondents' Location: EAP Estimation -- 2.3 Item Selection: Minimum Expected Posterior Variance -- 2.4 Check Stopping Rules -- 2.5 Creating Final Estimates -- 2.6 Application to NPI -- 3 Exploring Your Options: Alternative CAT Algorithms -- 3.1 Measuring Knowledge -- 3.2 Likelihood and Prior Distributions -- 3.3 Estimating θ -- 3.3.1 MLE and MAP -- 3.3.2 Weighted Maximum Likelihood -- 3.3.3 Fisher Information and Measures of Uncertainty -- 3.4 Item Selection -- 3.4.1 A General Framework for Understanding Item Selection -- 3.4.2 Fisher Information and Observed Information -- 3.4.3 Kullback-Leibler Information -- 3.4.4 Explore on Your Own -- 3.5 Stopping Rules -- 3.5.1 Stopping Thresholds -- 3.5.2 Stopping Overrides -- 3.6 Binary Responses with Guessing -- 3.7 Technical Appendix -- 3.7.1 Weighted Maximum Likelihood Estimation -- 4 CAT for Polytomous Outcomes -- 4.1 Graded Response Model (GRM) -- 4.1.1 Calculating Probabilities -- 4.1.2 Likelihood -- 4.1.3 Estimating the Respondents' Locations -- 4.1.4 Item Selection and Stopping Rules -- 4.2 Generalized Partial Credit Model (GPCM) -- 5 Evaluating Adaptive Inventories.
,
5.1 Item Selection Speed -- 5.2 Performance Based on Final Estimates -- 5.2.1 Simulations from Real World Data -- 5.2.2 Performance Based on Simulated Data -- 5.2.3 Metrics for Model Evaluation and Test Information -- 5.3 Example: Assessing Batteries for Neuroticism -- 5.3.1 Thinking about Diagnostics -- 5.3.2 Assessment in Practice -- 5.4 Summary -- 6 Tools and Tricks for Applied Researchers -- 6.1 Building a Battery -- 6.1.1 How Big? -- 6.1.2 Calibration Samples and Choosing a Prior -- 6.2 Getting to the Field -- 6.3 Interfacing with Webservice and Qualtrics -- 7 Implications and Future Directions -- References -- Acknowledgments.
Additional Edition:
ISBN 1-108-79726-1
Language:
English
URL:
https://doi.org/10.1017/9781108862516
