Abstract
Learning progressions are hypothetical models of student learning in a domain over extended periods of time. In many cases these progressions describe multiple 'big ideas' or constructs. Relationships between these constructs, i.e., how development along one might affect the other, are difficult to ascertain. Such relationships can be described from the perspectives of either item characteristics or student abilities. Existing methods of analyses focus predominantly on the 'item-side' of the equation and much less research addresses construct relationships from the 'student-side'. In this study, we supplemented a Multidimensional Item Response Modeling approach with a Latent Class Analysis to more fully explore both within and between-construct relationships. We analyzed student written responses (n=317) to 31 ordered-multiple-choice items targeted at five constructs in a genetics learning progression. We present our finding with the goal of comparing and contrasting the types of inferences that can be made with both measurement approaches.
| Original language | American English |
|---|---|
| Pages (from-to) | 607-614 |
| Number of pages | 8 |
| Journal | Proceedings of International Conference of the Learning Sciences, ICLS |
| Volume | 1 |
| Issue number | January |
| State | Published - 2014 |
| Event | 11th International Conference of the Learning Sciences: Learning and Becoming in Practice, ICLS 2014 - Boulder, United States Duration: Jun 23 2014 → Jun 27 2014 |
ASJC Scopus subject areas
- Computer Science (miscellaneous)
- Education