What is it about?
When we look at change of a construct over time, we can fit a growth curve/multilevel model using observed scores of items. This assumes the measured behaviors reflect the construct in the same way over time, but behaviors may also change its meaning over time. For example, 'crying easily' is an indicator of depression for adults but not so much for children. To capture change in both observed behaviors and the underlying construct in a flexible manner, such as across continuous time (e.g., age) or over background variables (e.g., sex, age x sex), we developed a longitudinal moderated factor model. In it we also suggested an automatic way to figure out how to specify the pattern of effects on the measurement model (whether an item changed its meaning or not).
Featured Image
Photo by Jeremy Bishop on Unsplash
Why is it important?
This model overcomes limitations of traditional growth curve models; it facilitates measurement evaluation over multiple background variables and growth modeling with many time points. The model can inform us on how manifest behaviors and their underlying construct each change over time and across groups in a flexible way based on researcher's hypothesis, without confounding these processes.
Perspectives
Read the Original
This page is a summary of: Modeling construct change over time amidst potential changes in construct measurement: A longitudinal moderated factor analysis approach., Psychological Methods, August 2024, American Psychological Association (APA),
DOI: 10.1037/met0000685.
You can read the full text:
Contributors
The following have contributed to this page