Research Framework - Statistical Analysis

Interpretation of Estimated Parameters

The estimated parameters and their standard errors must be investigated and interpreted.  Deciding how best to interpret estimation results is often an open research question. 

Ability Distribution

The nature of the ability distribution (for example, the shape and location of the scale-score distribution or that of a continuous unidimensional ability parameter) can sometimes have important repercussions that should be carefully investigated (for example, a distribution that may have been expected to be symmetrical may turn out to be highly skewed).

Item Parameters

How to interpret estimated item parameters and what to do about them is an important area of research, especially for new models or new applications. Results of this sort of research may inform the test design process.

---

Research

Measurement Error, Human Error, and Decisions Based on a Test, by Stuart Kahl (2003)