Related Experiment Videos
Using LinLog and FACETS to model item components in the LLTM.
Tracy L Kline1, Karen M Schmidt, Ryan P Bowles
1University of Virginia, USA. tkline@rti.org
Journal of Applied Measurement
|December 31, 2005
Summary
This study compared LinLog and FACETS software for analyzing spatial memory data using the linear logistic test model (LLTM). Both programs yielded similar parameter estimates, confirming FACETS
Area of Science:
- Psychometrics
- Cognitive Psychology
Background:
- The linear logistic test model (LLTM) is a powerful tool for analyzing item complexity in educational and psychological assessments.
- Investigating within-item complexity factors requires robust measurement programs capable of handling multifaceted data structures.
- Existing Rasch measurement programs may differ in their applicability to LLTM analyses, necessitating comparative studies.
Purpose of the Study:
- To evaluate the performance of LinLog and FACETS, two Rasch measurement programs, in estimating parameters for the linear logistic test model (LLTM).
- To determine if FACETS, typically used for the many-facet Rasch model (MFRM), can accurately estimate LLTM parameters when items have multiple dimensions.
- To compare parameter estimations from LinLog and FACETS using both original and simulated data for a spatial memory measure.
Main Methods:
- The study utilized data from a spatial memory measure, including an original dataset (114 persons) and simulated datasets (500 and 1000 persons).
- Analyses were conducted using LinLog, an LLTM-specific program employing conditional maximum likelihood, and FACETS, adapted for a multifaceted approach to item dimensions.
- Parameter estimations for persons and items were compared between the two programs under different data conditions.
Main Results:
- LinLog and FACETS analyses produced strikingly similar parameter estimates for both the original and simulated datasets.
- The results indicate that FACETS can be effectively used to estimate LLTM parameters, even when treating item dimensions as separate facets.
- The performance of both programs was robust across different sample sizes, supporting their reliability in LLTM analyses.
Conclusions:
- FACETS provides accurate parameter estimates for the linear logistic test model (LLTM), comparable to the specialized LinLog program.
- The multifaceted approach within FACETS is suitable for analyzing within-item complexity factors in measures like spatial memory.
- This finding expands the utility of FACETS for researchers employing LLTM in various psychological and educational measurement contexts.