Investigation of Cross-Cultural Measurement Invariance through Multi-Group Confirmatory Factor Analysis and Multi-Group Latent Class Analysis

Fatma Coşkun; Hamide Deniz Gülleroğlu

Makale detayları

Investigation of Cross-Cultural Measurement Invariance through Multi-Group Confirmatory Factor Analysis and Multi-Group Latent Class Analysis

Yazarlar:

Fatma Coşkun | Hamide Deniz Gülleroğlu

Yayıncı:

Kastamonu University

Ana Başlıklar:

Alan Eğitimleri (Diğer)

Dergi:

Kastamonu Eğitim Dergisi Vol.33 Issue 4, pp.816-828

Belge Tipi:

Makale

Yayın Tarihi:

11/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001360035

pISSN:

2147-9844

Anahtar kelimeler:

Özet:

 Purpose: This study aims to examine cross-cultural measurement invariance using multi-group confirmatory factor analysis MG-CFA and multi-group latent class analysis MG-LCA. DesignMethodologyApproach: For this purpose, data obtained from the Students Like Learning Mathematics Scale in the TIMSS 2019 study were used. The sample of the research was determined using the maximum variation sampling method. Measurement invariance analyses were conducted on 15 comparison groups formed by Singapore, Hong Kong, Japan, Norway, Turkey, and South Africa. Findings: The MG-CFA results showed that strict measurement invariance was achieved only between Singapore-Hong Kong and Hong Kong-Norway. Between South Africa and Turkey, measurement invariance was observed at the level of structural invariance, while in other groups, it was achieved at the level of metric invariance. According to the MG-LCA results, measurement invariance was established at the partially homogeneous model level for some groups and at the heterogeneous model level for others. Highlights: In MG-LCA, it was determined that for the cross-cultural comparisons where measurement invariance was achieved at the partially homogeneous model level, measurement invariance was largely achieved at the metric invariance level according to MG-CFA. The MG-CFA and MG-LCA approaches address measurement invariance from different perspectives. MG-CFA is a sample- and data-oriented approach, whereas MG-LCA is a person-centered approach. Using these two methods together provides more comprehensive insights into why measurement invariance could not be achieved.

 Purpose: This study aims to examine cross-cultural measurement invariance using multi-group confirmatory factor analysis MG-CFA and multi-group latent class analysis MG-LCA. DesignMethodologyApproach: For this purpose, data obtained from the Students Like Learning Mathematics Scale in the TIMSS 2019 study were used. The sample of the research was determined using the maximum variation sampling method. Measurement invariance analyses were conducted on 15 comparison groups formed by Singapore, Hong Kong, Japan, Norway, Turkey, and South Africa. Findings: The MG-CFA results showed that strict measurement invariance was achieved only between Singapore-Hong Kong and Hong Kong-Norway. Between South Africa and Turkey, measurement invariance was observed at the level of structural invariance, while in other groups, it was achieved at the level of metric invariance. According to the MG-LCA results, measurement invariance was established at the partially homogeneous model level for some groups and at the heterogeneous model level for others. Highlights: In MG-LCA, it was determined that for the cross-cultural comparisons where measurement invariance was achieved at the partially homogeneous model level, measurement invariance was largely achieved at the metric invariance level according to MG-CFA. The MG-CFA and MG-LCA approaches address measurement invariance from different perspectives. MG-CFA is a sample- and data-oriented approach, whereas MG-LCA is a person-centered approach. Using these two methods together provides more comprehensive insights into why measurement invariance could not be achieved.

 Purpose: This study aims to examine cross-cultural measurement invariance using multi-group confirmatory factor analysis MG-CFA and multi-group latent class analysis MG-LCA. DesignMethodologyApproach: For this purpose, data obtained from the Students Like Learning Mathematics Scale in the TIMSS 2019 study were used. The sample of the research was determined using the maximum variation sampling method. Measurement invariance analyses were conducted on 15 comparison groups formed by Singapore, Hong Kong, Japan, Norway, Turkey, and South Africa. Findings: The MG-CFA results showed that strict measurement invariance was achieved only between Singapore-Hong Kong and Hong Kong-Norway. Between South Africa and Turkey, measurement invariance was observed at the level of structural invariance, while in other groups, it was achieved at the level of metric invariance. According to the MG-LCA results, measurement invariance was established at the partially homogeneous model level for some groups and at the heterogeneous model level for others. Highlights: In MG-LCA, it was determined that for the cross-cultural comparisons where measurement invariance was achieved at the partially homogeneous model level, measurement invariance was largely achieved at the metric invariance level according to MG-CFA. The MG-CFA and MG-LCA approaches address measurement invariance from different perspectives. MG-CFA is a sample- and data-oriented approach, whereas MG-LCA is a person-centered approach. Using these two methods together provides more comprehensive insights into why measurement invariance could not be achieved.

 Çalışmanın amacı: Bu çalışmada, kültürlerarası ölçme değişmezliğinin çok gruplu doğrulayıcı faktör analizi ÇG-DFA ve çok gruplu örtük sınıf analizi ÇG-ÖSA ile incelenmesi amaçlanmıştır. Materyal ve Yöntem: Bu amaç doğrultusunda, TIMSS 2019 uygulamasında Matematik Öğrenmeyi Sevme Ölçeği ile elde edilen veriler kullanılmıştır. Araştırmanın örneklemi, maksimum çeşitlilik yöntemi ile belirlenmiştir. Ölçme değişmezliği analizleri Singapur, Hong Kong, Japonya, Norveç, Türkiye ve Güney Afrika Cumhuriyetinin oluşturduğu 15 karşılaştırma grubu üzerinde gerçekleştirilmiştir. Bulgular: ÇG-DFA sonuçları, katı ölçme değişmezliğinin yalnızca Singapur-Hong Kong ve Hong Kong-Norveç arasında sağlandığını göstermiştir. Güney Afrika Cumhuriyeti ve Türkiye arasında ölçme değişmezliği yapısal değişmezlik düzeyinde kalırken; diğer gruplarda arasında ölçme değişmezliğinin metrik değişmezlik düzeyinde sağlandığı görülmüştür. ÇG-ÖSA sonuçlarına göre, bazı gruplar arasında kısmi homojen model, diğerleri arasında ise heterojen model düzeyinde ölçme değişmezliği sağlanmıştır. Önemli Vurgular: ÇG-ÖSA da ölçme değişmezliğinin kısmi homojen model düzeyinde sağlandığı kültürlerarasında ÇG-DFAya göre ölçme değişmezliğinin büyük oranda metrik değişmezlik düzeyinde sağlandığı belirlenmiştir. ÇG-DFA ve ÇG-ÖSA yaklaşımları ölçme değişmezliğini farklı boyutlardan ele almaktadır. ÇG-DFA, örneklem ve veri odaklı bir yaklaşımken; ÇG-ÖSA birey merkezli bir yaklaşımdır. Bu iki yöntemin birlikte kullanılması, ölçme değişmezliğinin neden sağlanamadığı hakkında daha kapsamlı bilgiler sağlamaktadır.

Makale detayları

Investigation of Cross-Cultural Measurement Invariance through Multi-Group Confirmatory Factor Analysis and Multi-Group Latent Class Analysis

Yazarlar:

Fatma Coşkun | Hamide Deniz Gülleroğlu

Yayıncı:

Kastamonu University

Ana Başlıklar:

Alan Eğitimleri (Diğer)

Dergi:

Kastamonu Eğitim Dergisi Vol.33 Issue 4, pp.816-828

Belge Tipi:

Makale

Yayın Tarihi:

11/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001360035

pISSN:

2147-9844

Measurement invariance

Confirmatory factor analysis

Latent class analysis

TIMSS 2019

Student surveysÖlçme değişmezliği

Doğrulayıcı faktör analizi

Örtük sınıf analizi

TIMSS 2019

Öğrenci anketleri

Sayı

Makale detayları

Yazarlar: Fatma Coşkun | Hamide Deniz Gülleroğlu

Yayıncı: Kastamonu University

Ana Başlıklar: Alan Eğitimleri (Diğer)

Dergi: Kastamonu Eğitim Dergisi Vol.33 Issue 4, pp.816-828

Belge Tipi: Makale

Yayın Tarihi: 11/2025

Belge Dili: İngilizce

Document ID (Org No.): 2001360035

pISSN: 2147-9844

Sayı

Yazarlar:

Fatma Coşkun | Hamide Deniz Gülleroğlu

Yayıncı:

Kastamonu University

Ana Başlıklar:

Alan Eğitimleri (Diğer)

Dergi:

Kastamonu Eğitim Dergisi Vol.33 Issue 4, pp.816-828

Belge Tipi:

Makale

Yayın Tarihi:

11/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001360035

pISSN:

2147-9844