Precision and Differential Item Functioning on a Testlet-Based Test: The 1991 Law School Admissions Test as an Example

Abstract

Two components of the Law School Admissions Test (LSAT)--reading comprehension and analytic reasoning-are not constructed of individual items that can function autonomously. Instead they each consist of four clusters of items, in which each cluster refers to a common stem. In reading comprehension the common part is a single passage; in analytic reasoning it is a common situation. Such interdependent clusters of items have come to be called testlets. It has been found that when a test is constructed of testlets, traditional treatments of individual items as independent entities tends to yield overly optimistic estimates of reliability. It has also been found that even though individual items may pass muster in terms of their differential performance within various subgroups, this may not be true once the items within a testlet are treated as a coherent unit. Findings can be divided into three categories: (a) overall performance of individual subgroups on the test, (b) the reliability of each section, and (c) the differential performance of items/testlets. This article describes these findings and provides the details of the methods used in the investigation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 104 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.