Analysing global professional gender gaps using LinkedIn advertising data

Although women’s participation in tertiary education and the labour force has expanded over the past decades, women continue to be underrepresented in technical and managerial occupations. We analyse if gender inequalities also manifest themselves in online populations of professionals by leveraging...

Full description

Bibliographic Details
Main Authors: Kashyap, R, Verkroost, FCJ
Format: Journal article
Language:English
Published: Springer 2021
_version_ 1797108707207675904
author Kashyap, R
Verkroost, FCJ
author_facet Kashyap, R
Verkroost, FCJ
author_sort Kashyap, R
collection OXFORD
description Although women’s participation in tertiary education and the labour force has expanded over the past decades, women continue to be underrepresented in technical and managerial occupations. We analyse if gender inequalities also manifest themselves in online populations of professionals by leveraging audience estimates from LinkedIn’s advertisement platform to explore gender gaps among LinkedIn users across countries, ages, industries and seniorities. We further validate LinkedIn gender gaps against ground truth professional gender gap indicators derived from the International Labour Organization’s (ILO) Statistical Database, and examine the feasibility and biases of predicting global professional gender gap indicators using gender gaps computed from LinkedIn’s online population. We find that women are significantly underrepresented relative to men on LinkedIn in countries in Africa, the Middle East and South Asia, among older individuals, in Science, Technology, Engineering and Mathematics (STEM) fields and higher-level managerial positions. Furthermore, a simple, aggregate indicator of the female-to-male ratio of LinkedIn users, which we term the LinkedIn Gender Gap Index (GGI), shows strong positive correlations with ILO ground truth professional gender gaps. A parsimonious regression model using the LinkedIn GGI to predict ILO professional gender gaps enables us to expand country coverage of different ILO indicators, albeit with better performance for general professional gender gaps than managerial gender gaps. Nevertheless, predictions generated using the LinkedIn population show some distinctive biases. Notably, we find that in countries where there is greater gender inequality in internet access, LinkedIn data predict greater gender equality than the ground truth, indicating an overrepresentation of high status women online in these settings. Our work contributes to a growing literature seeking to harness the ‘data revolution’ for global sustainable development by evaluating the potential of a novel data source for filling gender data gaps and monitoring key indicators linked to women’s economic empowerment.
first_indexed 2024-03-07T07:32:27Z
format Journal article
id oxford-uuid:c31c1a78-03ae-4ed7-bab9-2998ad93b7ee
institution University of Oxford
language English
last_indexed 2024-03-07T07:32:27Z
publishDate 2021
publisher Springer
record_format dspace
spelling oxford-uuid:c31c1a78-03ae-4ed7-bab9-2998ad93b7ee2023-01-31T17:16:56ZAnalysing global professional gender gaps using LinkedIn advertising dataJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:c31c1a78-03ae-4ed7-bab9-2998ad93b7eeEnglishSymplectic ElementsSpringer2021Kashyap, RVerkroost, FCJAlthough women’s participation in tertiary education and the labour force has expanded over the past decades, women continue to be underrepresented in technical and managerial occupations. We analyse if gender inequalities also manifest themselves in online populations of professionals by leveraging audience estimates from LinkedIn’s advertisement platform to explore gender gaps among LinkedIn users across countries, ages, industries and seniorities. We further validate LinkedIn gender gaps against ground truth professional gender gap indicators derived from the International Labour Organization’s (ILO) Statistical Database, and examine the feasibility and biases of predicting global professional gender gap indicators using gender gaps computed from LinkedIn’s online population. We find that women are significantly underrepresented relative to men on LinkedIn in countries in Africa, the Middle East and South Asia, among older individuals, in Science, Technology, Engineering and Mathematics (STEM) fields and higher-level managerial positions. Furthermore, a simple, aggregate indicator of the female-to-male ratio of LinkedIn users, which we term the LinkedIn Gender Gap Index (GGI), shows strong positive correlations with ILO ground truth professional gender gaps. A parsimonious regression model using the LinkedIn GGI to predict ILO professional gender gaps enables us to expand country coverage of different ILO indicators, albeit with better performance for general professional gender gaps than managerial gender gaps. Nevertheless, predictions generated using the LinkedIn population show some distinctive biases. Notably, we find that in countries where there is greater gender inequality in internet access, LinkedIn data predict greater gender equality than the ground truth, indicating an overrepresentation of high status women online in these settings. Our work contributes to a growing literature seeking to harness the ‘data revolution’ for global sustainable development by evaluating the potential of a novel data source for filling gender data gaps and monitoring key indicators linked to women’s economic empowerment.
spellingShingle Kashyap, R
Verkroost, FCJ
Analysing global professional gender gaps using LinkedIn advertising data
title Analysing global professional gender gaps using LinkedIn advertising data
title_full Analysing global professional gender gaps using LinkedIn advertising data
title_fullStr Analysing global professional gender gaps using LinkedIn advertising data
title_full_unstemmed Analysing global professional gender gaps using LinkedIn advertising data
title_short Analysing global professional gender gaps using LinkedIn advertising data
title_sort analysing global professional gender gaps using linkedin advertising data
work_keys_str_mv AT kashyapr analysingglobalprofessionalgendergapsusinglinkedinadvertisingdata
AT verkroostfcj analysingglobalprofessionalgendergapsusinglinkedinadvertisingdata