ERIC Number: ED218309
Record Type: Non-Journal
Publication Date: 1981-Oct
Reference Count: N/A
An Investigation of Models for the Estimation of Test Score Reliability Using Longitudinal Data.
Goldstein, Harvey; Ecob, Russell
Using data from a National Child Development Study (NCDS) in Great Britain, the applications of instrumental variable methods and structural equation models to estimating instrumental variables are presented. A subset of the longitudinal educational and home background data on children born in England, Wales and Scotland in a March week of 1958 is used. The models for quantitative variables are discussed in terms of reliability in measurement error variance, relationships and models for true score measures on two or more occasions, and identification problems in measurement error. A correction for unreliability or measurement error variance of the independent variable, instrumental variable estimation, and estimation in structural equation models are discussed. The methods used with the NCDS data are summarized. The appendices include the research project submission, a description of the NCSD, and a comparison of the classical test score model and the latent variable model. Further appendices discuss the distributions of variables, transformations, instrumental variables theory and test score reliability, LISREL applications, estimate inconsistency, missing data and a social class variable. (CM)
Publication Type: Reports - Research
Education Level: N/A
Sponsor: National Inst. of Education (ED), Washington, DC.
Authoring Institution: London Univ. (England). Inst. of Education.
Identifiers - Location: United Kingdom (Great Britain)