Dimensionality assesment with factor analysis methods

Mariska Barendse

Dimensionality assesment with factor analysis methods

Mariska Barendse

Psychometrics and Statistics

Onderzoeksoutput

1621 Downloads (Pure)

Samenvatting

Onderzoekers in de sociale en gedragswetenschappen verzamelen via vragenlijsten vaak informatie over niet observeerbare variabelen die bijvoorbeeld attitudes, vaardigheden en opvattingen uitdrukken. De dimensionaliteit van een set itemscores is het aantal latente variabelen. Het bepalen van dit aantal is van belang bij schaalconstructie en –evaluatie. In dit proefschrift wordt factor analytische benaderingen ter bepaling van de dimensionaliteit onderzocht. In de factoranalyse wordt groot aantal geobserveerde items door een kleiner aantal latente variabelen beschreven. De klassieke factor analytische methoden zijn geschikt voor het beschrijven van lineaire relaties tussen multivariaat normaal verdeelde itemscores. In de praktijk zijn itemscores vaak categorisch en is er sprake van niet-lineaire relaties. Bovendien is er soms een hiërarchische structuur in de data, zoals wanneer meerdere kinderen in meerdere klassen zijn onderzocht. In dit proefschrift wordt onderzocht in hoeverre meer geavanceerde factor analytische methoden dan de klassieke geschikt zijn om de dimensionaliteit te bepalen bij complexe datastructuren, en welke procedure hierbij dan het meest geschikt is. De gehanteerde factor analytische methoden bleken in de bestudeerde situaties goed te kunnen omgaan met categorische items, niet-lineaire relaties en hiërarchisch structuren.

Originele taal-2	English
Kwalificatie	Doctor of Philosophy
Toekennende instantie	Rijksuniversiteit Groningen
Begeleider(s)/adviseur	Timmerman, Marieke, Supervisor Oort, F.J., Supervisor, Externe Persoon Meijer, Rob, Supervisor
Datum van toekenning	16-feb.-2015
Plaats van publicatie	[S.l.]
Uitgever	[S.n.]
Gedrukte ISBN's	978-90-367-7581-6
Elektronische ISBN's	978-90-367-7580-9
Status	Published - 2015

Toegang tot document

Title and contents Final publisher's version, 628 KB
Chapter 1Final publisher's version, 550 KB
Chapter 2Final publisher's version, 1,15 MB
Chapter 3Final publisher's version, 700 KB
Chapter 4Final publisher's version, 754 KB
Chapter 5Final publisher's version, 773 KB
Chapter 6 Publisher versionFinal publisher's version, 477 KB
Chapter 7Final publisher's version, 578 KB
BibliographyFinal publisher's version, 642 KB
Summary in DutchFinal publisher's version, 568 KB
Curriculum VitaeFinal publisher's version, 517 KB
List of PublicationsFinal publisher's version, 521 KB
AcknowledgementsFinal publisher's version, 545 KB
PropositionsFinal publisher's version, 151 KB

Handle.net

Permanente koppeling

Document onder embargo

Chapter 6
Final author's version, 837 KB
Kopie aanvragen

Document onder embargo

Complete dissertation
Final publisher's version, 2,37 MB
Kopie aanvragen

Citeer dit

@phdthesis{16a8c55b1c6d4465a5e0267480751fe4,

title = "Dimensionality assesment with factor analysis methods",

abstract = "Researchers in the social and behavioral sciences often collect scores on a set of items, to measure unobservable (i.e. latent) variables. The latter may express, for example, beliefs, traits, attitudes or skills. The dimensionality of a set of items is the number of latent variables. This number is important information in scale construction and scale evaluation. In this dissertation approaches to assess the dimensionality of a set of items with factor analysis methods are investigated. Within the factor analysis framework the relation between a large amount of observed items is described by a smaller amount of latent variables (i.e. factors). The classical factor analysis involves linear relations between variables with multivariate normally distributed continuous item scores. I examined to what extent more elaborated factor analysis methods are suited, and which procedures have to be applied, to determine the dimensionality in situations that are often encountered in practice. Those cases go beyond the classical case. Examples are discrete item scores (e.g., Likert scales ranging from “definitely does not apply” to “definitely does apply{"}), nonlinear relations between the items, and data with a hierarchical structure (e.g., data pertaining to different pupils from different schools). Under the conditions studied here, the factor analysis models and associated procedures seemed to be able to handle discrete responses, nonlinear relations, and multilevel structures.",

author = "Mariska Barendse",

year = "2015",

language = "English",

isbn = "978-90-367-7581-6",

publisher = "[S.n.]",

school = "University of Groningen",

}

TY - BOOK

T1 - Dimensionality assesment with factor analysis methods

AU - Barendse, Mariska

PY - 2015

Y1 - 2015

N2 - Researchers in the social and behavioral sciences often collect scores on a set of items, to measure unobservable (i.e. latent) variables. The latter may express, for example, beliefs, traits, attitudes or skills. The dimensionality of a set of items is the number of latent variables. This number is important information in scale construction and scale evaluation. In this dissertation approaches to assess the dimensionality of a set of items with factor analysis methods are investigated. Within the factor analysis framework the relation between a large amount of observed items is described by a smaller amount of latent variables (i.e. factors). The classical factor analysis involves linear relations between variables with multivariate normally distributed continuous item scores. I examined to what extent more elaborated factor analysis methods are suited, and which procedures have to be applied, to determine the dimensionality in situations that are often encountered in practice. Those cases go beyond the classical case. Examples are discrete item scores (e.g., Likert scales ranging from “definitely does not apply” to “definitely does apply"), nonlinear relations between the items, and data with a hierarchical structure (e.g., data pertaining to different pupils from different schools). Under the conditions studied here, the factor analysis models and associated procedures seemed to be able to handle discrete responses, nonlinear relations, and multilevel structures.

AB - Researchers in the social and behavioral sciences often collect scores on a set of items, to measure unobservable (i.e. latent) variables. The latter may express, for example, beliefs, traits, attitudes or skills. The dimensionality of a set of items is the number of latent variables. This number is important information in scale construction and scale evaluation. In this dissertation approaches to assess the dimensionality of a set of items with factor analysis methods are investigated. Within the factor analysis framework the relation between a large amount of observed items is described by a smaller amount of latent variables (i.e. factors). The classical factor analysis involves linear relations between variables with multivariate normally distributed continuous item scores. I examined to what extent more elaborated factor analysis methods are suited, and which procedures have to be applied, to determine the dimensionality in situations that are often encountered in practice. Those cases go beyond the classical case. Examples are discrete item scores (e.g., Likert scales ranging from “definitely does not apply” to “definitely does apply"), nonlinear relations between the items, and data with a hierarchical structure (e.g., data pertaining to different pupils from different schools). Under the conditions studied here, the factor analysis models and associated procedures seemed to be able to handle discrete responses, nonlinear relations, and multilevel structures.

M3 - Thesis fully internal (DIV)

SN - 978-90-367-7581-6

PB - [S.n.]

CY - [S.l.]

ER -