A Quantitative Examination of Gender Bias and Language Paradigm in the Evaluation of Source Code

• Main Author: Rangel, Pamela J.
• Language: EN
• Published: Northcentral University 2018
• Subjects: Computer engineering|Engineering|Computer science
• Online Access: http://pqdtopen.proquest.com/#viewpdf?dispub=10784617

<p> Women represent 16% of the software development workforce in the United States. Simultaneously, organizations are unable to meet the increasing demands of software development and maintenance due to a lack of personnel. Gender bias influences female departure from STEM, but for software engineering the problem is possibly exacerbated by disregarded preferences for programming paradigms which impact self-efficacy. The purpose of this study examining gender bias in STEM was to determine if software source code developed by female and male software engineers is evaluated differently, to determine if code developed under structured and object oriented paradigms is evaluated differently, and to determine if there is an interaction effect between developer gender and programing paradigm. This study was framed by Social Cognitive Career Theory, which encompasses perspectives of career choice, occupational roles, and gender bias, and their impact on self-efficacy. Respondents were asked to score source code written by a fictive male or female developer, and written in an object oriented or structured programming language, for readability, quality and organization. The independent variables, gender of code author and software language paradigm, provided the basis for 2 x 2 between subjects experimental analyses examining main effects related to readability, quality, and organization, and any interaction effect. The target population for this study was software developers or students in the United States, sought via an online survey. Participants were randomly assigned one of four code examples. No interaction effect was demonstrated, but there were significant main effects related to gender of code author and quality, and related to code paradigm and readability. The overall trend suggested the fictive female author was scored higher than the fictive male author. These unexpected results support the need for further understanding of the complexities of gender related to software engineering, and should not provide a foundation for complacency in regard to improving female participation in software engineering. This extension of Social Cognitive Career Theory to software engineering may serve to influence stakeholders in understanding why female persistence and participation in software engineering remains minimal.</p><p>