Measuring Cognitive Activities in Software Engineering

TL;DR

This paper introduces a multidisciplinary approach to studying cognitive activities in software engineering through video analysis, coding, and modeling, aiming to improve understanding and identify best practices.

Contribution

It presents a formal hierarchical coding scheme and combines psychological and statistical analyses to create a generic cognitive modeling approach for software engineering.

Findings

Initial data from technical review meetings analyzed

Hierarchical coding scheme enables comparison of observations

Merging psychological and statistical methods advances cognitive modeling

Abstract

This paper presents an approach to the study of cognitive activities in collaborative software development. This approach has been developed by a multidisciplinary team made up of software engineers and cognitive psychologists. The basis of this approach is to improve our understanding of software development by observing professionals at work. The goal is to derive lines of conduct or good practices based on observations and analyses of the processes that are naturally used by software engineers. The strategy involved is derived from a standard approach in cognitive science. It is based on the videotaping of the activities of software engineers, transcription of the videos, coding of the transcription, defining categories from the coded episodes and defining cognitive behaviors or dialogs from the categories. This project presents two original contributions that make this approach generic in software engineering. The first contribution is the introduction of a formal hierarchical coding scheme, which will enable comparison of various types of observations. The second is the merging of psychological and statistical analysis approaches to build a cognitive model. The details of this new approach are illustrated with the initial data obtained from the analysis of technical review meetings.