| Autor: |
Kolbe, L. |
| Přispěvatelé: |
Oort, Frans, Jorgensen, Terrence, Jak, Suzanne, Methods and Statistics (RICDE, FMG) |
| Jazyk: |
angličtina |
| Rok vydání: |
2022 |
| Popis: |
The measurement of psychological constructs like mathematical ability or social anxiety is at the heart of social and behavioral sciences. Psychological constructs are often referred to as latent constructs because they are not directly observable. As a result, these constructs are measured with observable variables such as responses to questionnaire or test items that are assumed to represent the latent construct. Given the dominant role of such tests and questionnaires in social and behavioral sciences and applications, it is important that the items on these measurement instruments function the same way across individuals or groups. This requirement is also referred to as measurement invariance. If measurement invariance does not hold, differences in observed scores across individuals or groups may arise from differences on background variables (such as age or gender) instead of differences on the latent construct. It is therefore important to assess measurement invariance before comparing individuals or groups on latent constructs. This dissertation addresses novel approaches to assess measurement invariance in the framework of structural equation modeling. More specifically, the focus of this dissertation is on new ways to assess measurement invariance using single-group methods like restricted factor analysis and moderated nonlinear factor analysis |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
