Toward a Causal Interpretation of the Common Factor Model

Bork, Riet van; Wijsen, Lisa D.; Rhemtulla, Mijke

doi:10.1515/disp-2017-0019

Cited by 31 publications

(25 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…item 8 "I prefer to just let things happen rather than to understand why they turned out that way"). These three domains as originally described are to be interpreted as passive and interchangeable consequences of the construct itself, and therefore are not active contributors to alexithymia (Van Bork, Wijsen, & Rhemtulla, 2017). RUNNING HEAD: NETWORK ANALYSIS OF ALEXITHYMIA In the last decade, network analysis has affirmed itself as new way of analyzing data in psychiatry and psychology, which allows to conceive constructs or mental disorders as a complex system of mutually influencing elements (Borsboom & Cramer, 2013).…”

Section: Introductionmentioning

confidence: 99%

“…The Toronto Alexithymia Scale (TAS) is the most commonly used measure of alexithymia in empirical research (Bagby, Parker, & Taylor, 1994) and is composed of 20 items designed to assess three domains: difficulty identifying feelings (e.g., item 1 “I am often confused about what emotion I am feeling”), difficulty describing feelings (e.g., item 4 “It is difficult for me to find the right words for my feelings”), and externally oriented thinking (e.g., item 8 “I prefer to just let things happen rather than to understand why they turned out that way”). These three domains as originally described are to be interpreted as passive and interchangeable consequences of the construct itself and therefore are not active contributors to alexithymia (Van Bork, Wijsen, & Rhemtulla, 2017).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Network Approach to Items and Domains From the Toronto Alexithymia Scale

Briganti

Linkowski

2019

Psychol Rep

View full text Add to dashboard Cite

The aim of this paper is to explore network structures of the Toronto Alexithymia Scale (TAS) in a large sample of 1925 French-speaking Belgian university students and compare results with previous studies from different samples and tools to identify potential targets for clinical intervention. We estimated network models for the twenty items of the TAS and for its three domains difficulty identifying feelings, difficulty describing feelings and externally-oriented thinking. We explored item connectivity through node predictability (shared variance with other network components). We performed an Exploratory Graph Analysis (EGA) to explore the dimensionality of our dataset and compare results with the original three-factor model; because a different model was proposed, we estimated an additional network structure on the new structure. Items from the TAS connect both within and between domains. The three-domain network identifies difficulty describing feelings as the most connected domain. The EGA reported that three items from externally-oriented thinking form a new domain, distraction. In the new four-domain network, difficulty describing feelings remains the most interconnected domain; however, two negative connections are found.Our findings support the relative importance of identifying and describing feelings as a meaningful target for intervention.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Network Approach to Items and Domains From the Toronto Alexithymia Scale

Briganti

Linkowski

2019

Psychol Rep

View full text Add to dashboard Cite

show abstract

“…A realist philosophy of psychometrics makes sense in some ways (Borsboom, Mellenbergh, & Van Heerden, 2004;Hood, 2013;Maul, 2013;Van Bork, Wijsen, & Rhemtulla, 2017), and the common factor model specifically invites a realist interpretation. Chapters 4 and 5 both contribute to a realist or causal understanding of psychometric models.…”

Section: A Pragmatist Philosophy Of Psychometricsmentioning

confidence: 99%

Characterizations of Psychometrics

Wijsen¹

2021

Preprint

View full text Add to dashboard Cite

Psychometrics is a mostly technical and model-based discipline that is concerned with (quantitative) psychological measurement. This dissertation aims to uncover the intricacies of this discipline by drawing up three characterizations of psychometrics: a historical characterization, an ethnographic characterization, and a philosophical characterization. The historical characterization involves an academic genealogy of psychometrics, in which the history of psychometrics is visualized by tracing back the doctoral advisors of prominent psychometricians. In the ethnographic characterization, psychometricians are interviewed on the history and future of their discipline, as well as the dealings of contemporary psychometrics. The philosophical characterization offers a realist reading of psychometric models and an analysis of the use of values in psychometric research. Though each characterization certainly has its own distinct methodological features, together the characterizations paint a multidimensional and nuanced portrait of psychometrics. Psychometrics turns out to be a diverse discipline that incorporates a number of tensions: 1) psychometrics seems to be a subfield of psychology but also a discipline that can be characterized as a subdiscipline of statistics; 2) psychometricians use models that imply a realist philosophy (useful for formulating hypotheses), but at the same time engage in mostly building useful applications; and 3) though psychometrics has played a considerable role in historical and modern-day society, the discipline also seems quite isolated from real social involvement. All in all, I hope that this dissertation unlocks some of the complexities of psychometrics, and specifically, makes psychometrics a more accessible research area for fellow historians, philosophers, and sociologists of science.

show abstract

“…In that case, to argue for a common factor model requires a case to be made that only the shared variance and not the unique variance of items is of interest. We can think of no rationale for such a case that does not hinge on a common cause (van Bork, Wijsen, & Rhemtulla, 2018) If psychology can achieve both these goals-if methodologists can make available a more complete range of measurement models and if researchers can be encouraged to justify their measurement models-it will have made much progress toward improving the inferences resulting from structural equation models.…”

Section: Conclusion and Next Stepsmentioning

confidence: 99%

WorseThanMeasurementError

Rhemtulla¹,

Bork²,

Borsboom³

2018

Preprint

Self Cite

View full text Add to dashboard Cite

Previous research and methodological advice has focused on the importance of accounting for measurement error in psychological data. That perspective assumes that psychological variables conform to a common factor model, such that they consist of construct variance plus error. In this paper, we explore what happens when a set of items that are not generated from a common factor construct model are nonetheless modeled as reflecting a common factor. Through a series of hypothetical examples and an empirical re-analysis, we show that (1) common factor models tend to produce extremely biased and highly variable structural parameter estimates when the population model is not a common factor model; (2) model fit is a poor indicator of the degree of bias; and (3) composite models are sometimes more reliable than common factor models under alternative measurement structures, though they also lead to unacceptably bad solutions in some cases.

show abstract

Toward a Causal Interpretation of the Common Factor Model

Cited by 31 publications

References 23 publications

Network Approach to Items and Domains From the Toronto Alexithymia Scale

Network Approach to Items and Domains From the Toronto Alexithymia Scale

Characterizations of Psychometrics

WorseThanMeasurementError

Contact Info

Product

Resources

About