On the use of regression and verbal protocol analysis in modeling analysts' behavior in an unstructured task environment: A methodological note

We use two techniques - linear regression (LR) and verbal protocol analysis (VPA) - to model the decision behavior of four experienced financial analysts performing an unstructured valuation task. The research examines the structure and implied data use for each model type. We also assess the conclusion of prior research that the two techniques are complementary. This research also claims that these techniques are interchangeable in most decision contexts, in effect treating them as substitutes. Our results suggest that the techniques are complementary; however, each adds unique contributions to model development. (C) 1998 Elsevier Science Ltd. All rights reserved.