Margins of Error: A Study of Reliability in Survey Measurement
John Wiley & Sons, 2007. gada 9. jūl. - 416 lappuses
Enhance the quality of survey results by recognizing and reducing measurement errors.
Margins of Error: A Study of Reliability in Survey Measurement demonstrates how and hwy identifying the presence and extent of measurement errors in survey data is essential for improving the overall collection and analysis of the data. The author outlines the consequences of ignoring survey measurement errors and also discusses ways to detect and estimate the impact of these errors. This book also provides recommendations of improving the quality of survey data.
Logically organized and clearly written, this book:
In conjunction with research data gathered on nearly 500 survey measures and the application of an empirical approach grounded in classical measurement theory, this book discusses the sources of measurement error and provides the tools necessary for improving survey data collection methods.
Margins of Error enables statisticians and researchers in the fields of public opinion and survey research to design studies that can detect, estimate, and reduce measurement errors that may have previously gone undetected. This book also serves as a supplemental textbook for both undergraduate and graduate survey methodology courses.
1.5. rezultāts no 39.
Even in the simplest regression models, measurement unreliability in predictor variables generally biases regression coefficients downward, making it more difficult to reject the null hypothesis; and unreliability in both dependent and ...
... statistical matters, especially tests of statistical significance, correlation estimates, regression coefficients, and other information based on sample estimates of response variance, it hardly seems realistic to ignore them.
... but we strenuously object to the interpretation of such internal consistency coefficients as estimates of reliability of survey measurement. In developing this argument we first provide a formal statement of classical true-score ...
... The (G X K) matrix of regression coefficients relating observed measures to true scores in population S (I) The (K X K) matrix of covariances among latent true scores in population S O2 The (G X G) matrix of covariances among errors ...
Esat sasniedzis šīs grāmatas aplūkošanas reižu limitu.