Dynamics of Gender Bias in Computing

TL;DR

This study introduces a new longitudinal dataset revealing three distinct periods of gender bias in computing from 1950 to 1980, challenging previous linear models and informing diversity efforts.

Contribution

It provides the first systematic, archival-based dataset on gender bias in early computing, revising historical understanding and emphasizing the non-permanent, contingent nature of bias.

Findings

Identifies three distinct periods of gender bias in computing.

Revises the timeline of when gender bias emerged, starting in the mid-1980s.

Challenges the linear 'pipeline' model of gender bias development.

Abstract

Gender bias in computing is a hard problem that has resisted decades of research. One obstacle has been the absence of systematic data that might indicate when gender bias emerged in computing and how it has changed. This article presents a new dataset (N=50,000) focusing on formative years of computing as a profession (1950-1980) when U.S. government workforce statistics are thin or non-existent. This longitudinal dataset, based on archival records from six computer user groups (SHARE, USE, and others) and ACM conference attendees and membership rosters, revises commonly held conjectures that gender bias in computing emerged during professionalization of computer science in the 1960s or 1970s and that there was a 'linear' one-time onset of gender bias to the present. Such a linear view also lent support to the "pipeline" model of computing's "losing" women at successive career stages. Instead, this dataset reveals three distinct periods of gender bias in computing and so invites temporally distinct explanations for these changing dynamics. It significantly revises both scholarly assessment and popular understanding about gender bias in computing. It also draws attention to diversity within computing. One consequence of this research for CS reform efforts today is data-driven recognition that legacies of gender bias beginning in the mid-1980s (not in earlier decades) is the problem. A second consequence is correcting the public image of computer science: this research shows that gender bias is a contingent aspect of professional computing, not an intrinsic or permanent one.