REVIEW ARTICLE


The Logic of Latent Variable Analysis as Validity Evidence in Psychological Measurement



Purya Baghaei*, Mona Tabatabaee Yazdi
English Department, Mashhad Branch, Islamic Azad University, Mashhad, Iran


Article Metrics

CrossRef Citations:
18
Total Statistics:

Full-Text HTML Views: 5519
Abstract HTML Views: 2630
PDF Downloads: 943
ePub Downloads: 666
Total Views/Downloads: 9758
Unique Statistics:

Full-Text HTML Views: 2652
Abstract HTML Views: 1378
PDF Downloads: 681
ePub Downloads: 457
Total Views/Downloads: 5168



Creative Commons License
© Baghaei and Tabatabaee Yazdi; Licensee Bentham Open

open-access license: This is an open access article licensed under the terms of the Creative Commons Attribution-Non-Commercial 4.0 International Public License (CC BY-NC 4.0) (https://creativecommons.org/licenses/by-nc/4.0/legalcode), which permits unrestricted, non-commercial use, distribution and reproduction in any medium, provided the work is properly cited.

* Address correspondence to this author at the Islamic Azad University, Ostad Yusofi St., 91871 -Mashhad, Iran; Tel: +98 51 36635064; Fax: +98 51 36634763; E-mails: pbaghaei@mshdiau.ac.ir, puryabaghaei@gmail.com


Abstract

Background:

Validity is the most important characteristic of tests and social science researchers have a general consensus of opinion that the trustworthiness of any substantive research depends on the validity of the instruments employed to gather the data.

Objective:

It is a common practice among psychologists and educationalists to provide validity evidence for their instruments by fitting a latent trait model such as exploratory and confirmatory factor analysis or the Rasch model. However, there has been little discussion on the rationale behind model fitting and its use as validity evidence. The purpose of this paper is to answer the question: why the fit of data to a latent trait model counts as validity evidence for a test?

Method:

To answer this question latent trait theory and validity concept as delineated by Borsboom and his colleagues in a number of publications between 2003 to 2013 is reviewed.

Results:

Validating psychological tests employing latent trait models rests on the assumption of conditional independence. If this assumption holds it means that there is a ‘common cause’ underlying the co-variation among the test items, which hopefully is our intended construct.

Conclusion:

Providing validity evidence by fitting latent trait models is logistically easy and straightforward. However, it is of paramount importance that researchers appreciate what they do and imply about their measures when they demonstrate that their data fit a model. This helps them to avoid unforeseen pitfalls and draw logical conclusions.

Keywords: Validity, Latent trait models, Local independence, Classical test theory.