Academic Journal
The Impact of Sampling and Measurement Error in Y on the Estimation of Level 2 Standardized X-Y Relationships.
| Τίτλος: | The Impact of Sampling and Measurement Error in Y on the Estimation of Level 2 Standardized X-Y Relationships. |
|---|---|
| Συγγραφείς: | Konold, Timothy R.1 (AUTHOR) Konold@Virginia.edu, Sanders, Elizabeth A.2 (AUTHOR) |
| Πηγή: | Structural Equation Modeling. Jul/Aug2026, Vol. 33 Issue 4, p599-616. 18p. |
| Θεματικοί όροι: | *Sampling errors, *Measurement errors, *Mathematical variables, *Mathematical statistics, *Structural equation modeling, *Simulation methods & models, *Multilevel models, *Coefficients (Statistics) |
| Περίληψη: | This study focuses on Level 2 (L2) predictor-outcome bias in standardized coefficients as obtained through traditional multilevel models and multilevel structural equation models when sampling and measurement error is present in the outcome. While unreliability in the outcome variable does not bias unstandardized coefficients, the current study demonstrates how it can bias estimates in their more commonly reported standardized forms. Using simulations, we evaluate this issue under a variety of conditions in which sampling and measurement error are present in the outcome as a function of aggregation method. In doing so, we consider four methods for modeling L2 aggregate variables, including doubly manifest aggregation (MLM), doubly latent aggregation (MLSEM), and their two hybrids, in relation to the type of L2 variables researchers may encounter in practice and the extent to which a given methods can produce unbiased standardized coefficient values. [ABSTRACT FROM AUTHOR] |
| Βάση Δεδομένων: | Academic Search Index |
καταχωρήστε σχόλιο πρώτοι!