RESEARCH

ABSTRACT

The extensive literature on reuse in software engineering has focused on technical and organizational factors, largely ignoring cognitive characteristics of individual developers. Despite anecdotal evidence that cognitive heuristics play a role in successful artifact reuse, few empirical studies have explored this relationship. This paper proposes how a cognitive heuristic, called anchoring, and the resulting adjustment bias can be adapted and extended to predict issues that might arise when developers reuse code and/or designs. The research proposes that anchoring and adjustment can be manifested in three ways: propagation of errors in reuse artifacts, failure to include requested functionality absent from reuse artifacts, and inclusion of unrequested functionality present in reuse artifacts. Results from two empirical studies are presented. The first study examines reuse of object classes in a programming task, using a combination of practicing programmers and students. The second study uses a database design task with student participants. Results from both studies indicate that anchoring occurs. Specifically, there is strong evidence that developers tend to use the extraneous functionality in the artifacts they are reusing and some evidence of anchoring to errors and omissions in reused artifacts. Implications of these findings for both practice and future research are explored.