Evaluating the Missing Data Assumptions of the Chain and Poststratification Equating Methods

Sandip Sinharay,Paul W. Holland,Alina A. von Davier

doi:10.1007/978-0-387-98138-3_17

Evaluating the Missing Data Assumptions of the Chain and Poststratification Equating Methods

Sandip Sinharay, Paul W. Holland + Show 1 more

https://doi.org/10.1007/978-0-387-98138-3_17

Copy DOI

Publication Date: Jan 1, 2009

Citations: 1

Affiliation: Educational Testing Service

#Nonequivalent Groups With Anchor Test #Nonequivalent Groups With Anchor Test Design + Show 8 more

Abstract
Full-Text PDF
Similar Papers

Abstract

The nonequivalent groups with anchor test (NEAT) design, also known as the common-item, nonequivalent groups design (Kolen & Brennan, 2004), is used for equating several operational tests. Two types of observed-score equating methods often used with the NEAT design are chain equating (CE) and poststratification equating (PSE). Here, we consider their nonlinear versions, that is, the frequency-estimation equipercentile method for PSE and the chained equipercentile methods for CE (see Kolen & Brennan, 2004).

Full Text