Konnikov, Alla; Denier, Nicole; Hu, Yang; Hughes, Karen; Deutsch, Rebecca; Ding, Lei; Jabir Alshehabi, Al-Ani (; ... Tarafdar, Monideepa; + view all Konnikov, Alla; Denier, Nicole; Hu, Yang; Hughes, Karen; Deutsch, Rebecca; Ding, Lei; Jabir Alshehabi, Al-Ani (; Rets, Irina; Tarafdar, Monideepa; - view fewer (2022) BIAS Word inventory for work and employment diversity, (in)equality and inclusivity (Version 2.0). [Digital scholarly resource]. https://doi.org/10.31235/osf.io/t9v3a

Preview

Text BIAS Word Inventory v2.pdf - Published Version Download (277kB) | Preview

Abstract

This version provides a number of updates to enhance the word list for use. Based on additional reading of related academic literature, experience using the list with word embedding models, and practical questions that emerged in our own analyses, we undertook to review the previous word list and add words, review which words should be counted as roots or exact phrase matches, and clarify the classification of words that may have had ambiguous meaning. To do so, our expert team reviewed the existing list, proposed missing words to be added, flagged words that were ambiguous or may require exact matching, and compared this list of expanded or flagged words to a random sample of job postings. The team then compared notes, assessed the list in relation to a database of job postings for common usage, and updated this document in December 2023. // ABSTRACT: The language used in job advertisements contains explicit and implicit cues, which signal employers’ preferences for candidates of certain ascribed characteristics, such as gender and ethnicity/race. To capture such biases in language use, existing word inventories have focused predominantly on gender and are based on general perceptions of the ‘masculine’ or ‘feminine’ orientations of specific words and socio-psychological understandings of ‘agentic’ and ‘communal’ traits. Nevertheless, these approaches are limited to gender and they do not consider the specific contexts in which the language is used. To address these limitations, we have developed a comprehensive word inventory for work and employment diversity, (in)equality, and inclusivity that builds on a number of conceptual and methodological innovations. The BIAS Word Inventory was developed as part of our work in an international, interdisciplinary project – BIAS: Responsible AI for Labour Market Equality – in Canada and the United Kingdom (UK). Conceptually, we rely on a sociological approach that is attuned to various documented causes and correlates of inequalities related to gender, sexuality, ethnicity/race, immigration and family statuses in the labour market context. Methodologically, we rely on ‘expert’ coding of actual job advertisements in Canada and the UK, as well as iterative cycles of inter-rater verification. These expert codings can be paired with natural language processing (NLP) techniques and word embeddings from large language models (LLM) to further expand the list. Our inventory is particularly suited for studying labour market inequalities, as it reflects the language used to describe job postings, and the inventory takes account of cues at various dimensions, including explicit and implicit cues associated with gender, ethnicity, citizenship and immigration statuses, role specifications, equality, equity and inclusivity policies and pledges, work-family policies, and workplace context.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item