In spite of the recent trends in Educational Research highlighting the importance of non-cognitive traits, such as well-being, as priority learning outcomes, longitudinal measurement models that would be the most suited to measure such states and traits have not been fully explored. The current study illustrates some of the potential advantages of longitudinal assessment designs using data collected to study the relationship between sleep quality and well-being of university of education students through a prospective research study. The data were collected using a four-week assessment design and required students to respond to "weekly" self-ratings scales on-line. The relationships between weekly measurements were evaluated using the Multi Trait Multi Method, while the predictive power of sleep quality or other indicators on well-being were estimated using the Latent Growth Curve Modeling. The findings support that longitudinal measurement models can be useful assessment tools when making inferences about state versus trait nature of the variables from the affective domain and whether the intra- or the inter-relationships among them change over time. Moreover, the results suggest that using longitudinal, instead of single-take, measurements may greatly enhance the validity arguments when it the competencies of interest are prone show growth or change over time.
Key words: Longitudinal assessment designs, multi-trait multi-method, latent growth curve, emotional states, sleep quality.
|
