Inequality in Abortion Service Access after Dobbs v. Jackson: a Search Engine Audit

In June 2022, The United States Supreme Court overturned a federal right to abortion in its ruling on Dobbs v. Jackson. As a result of new barriers, an increasing proportion of abortion care has moved to telehealth and online with search engines becoming critical tools for people seeking this vital reproductive health service. In this study, we performed an audit of the advertisements shown to Google Search users seeking information about abortion across the United States during the year following the Supreme Court decision. We focused on the presence of abortion clinics in the ads shown to the search engine’s users, as well as Crisis Pregnancy Centers (CPCs), which are organizations that target patients seeking abortion, and then often provide misinformation to discourage abortion procedures. We performed Google searches every day from 2022-06-27 to 2023-06-16 for 20 queries related to abortion access, while geolocating the request from each of the 435 voting districts of the United States. The resulting 1,387,550 search engine result pages were processed to extract advertisements. We then correlated the extracted results with various socio-economic indicators of those residing in each of the voting districts. Overall, the most frequent category of domains in our data is CPCs at 46.8%, followed by abortion clinics (29.9%), pro-choice advocacy or research (16.4%), and anti-choice advocacy or research (2.4%). We then focused on the results for queries around abortion cost, mainly “free abortion” and “abortion cost”, and compared them to the median household income in the voting district. We controlled for the percentage of Republican voters in the 2022 midterm elections and number of abortion clinics in the district. We found that advertisements pointing to CPCs were inversely related to income, such that districts in states with abortion limits had the highest predicted counts of CPC advertisements and these decreased by 11% with every $10,000 increase in district income. This decrease was at 20% in states with outright bans. Thus, we found evidence that the advertisements preferentially targeted places where delays in care could result in denial of care. The methodology of platform auditing used in this study is a powerful tool to measure the inequalities they help perpetuate, and we hope that platform auditing will be both prominent in the research on social policies, as well as supported by legislation ensuring access to data and algorithms of the digital platforms affecting our daily lives and wellbeing.