Online advertising offers an effective means of reaching target audiences so it is unsurprising that it is now integral to any modern, political campaign. However, the lack of transparency presents significant risks and challenges and could potentially undermine the integrity of the electoral process. The European Commission’s Code of Practice on Disinformation is a welcome step towards securing transparency in online political advertising. As the effectiveness of the Code can only be ascertained though careful monitoring, the work of the European Regulators Group for Audiovisual Media Services (ERGA) and the Broadcasting Authority of Ireland (BAI) is invaluable. FuJo was delighted to partner with the BAI to implement the monitoring process for Ireland during the European elections.
In April 2019 European Regulators Group for Audiovisual Media Services (ERGA) established a research project covering 13 member states to monitor the political advertising archives established by Facebook, Google and Twitter prior to the European elections in May 2019. This work was undertaken to support the European Commission in monitoring the implementation of the commitments made by Google, Facebook and Twitter under the self-regulatory Code of Practice on Disinformation. The platforms advised ERGA to rely on the online libraries or archives of adverts that were made publicly available ahead of the elections. This report summarises the data and key findings in relation to the monitoring of these archives conducted in Ireland by the Institute for Future Media and Journalism at the request of the Broadcasting Authority of Ireland.
The data consisted of 1,554 political advertisements from four archives – Facebook’s Ad Library (280); Facebook’s Ad Library Report (1,091), Twitter’s Ad Transparency Centre (88); and Google’s Ad Transparency Report (95) – during the period April 18th 2019 to May 24th 2019.
As outlined in the report, ERGA identified a set of nine questions to assess the extent to which each platform disclosed relevant information about political advertising. Only six of these questions could be answered by using the political advertising archives provided by Facebook, Google and Twitter. To answer those six questions, adverts were monitored for the presence of the following information: (i) whether the advert was paid for; (ii) who paid for it; (iii) if it carried a disclaimer stating that it was a political or issue-based advert; (iv) information on micro-targeting options; and (v) spending information.
We found that:
1. The level of transparency was mixed. In the majority of cases studied both Facebook and Twitter provided a disclaimer on the immediate image of the advert that advised it was related to social issues or politics. Google however, although hosting the adverts on a webpage entitled ‘political adverts in the European Union’ did not provide any individual disclaimers on the immediate images of the adverts. In relation to the disclosure of payer information, both Facebook and Google made this clear in the majority of cases whilst Twitter required the researcher to click through various links to subsections of the ad details pages in order to obtain this information.
2. Overall, we found that microtargeting information was available in all cases but that it was limited to geography, gender and age. We were unable to assess whether other targeting options were available to the advertisers (such as special interests or political persuasion).
3. We found that all platforms labelled individual adverts as sponsored and the sponsor name was present in the plurality of adverts as was the disclaimer with the exception of Google. Although each platform adopted a different approach to classifying and labelling relevant adverts.
4. Across all three platforms the majority of adverts clearly contained the sponsor’s name, however Facebook demonstrated some inconsistency here with a portion of the adverts analysed omitting sponsor information.
5. We found that all adverts carried some spending information, but that much information was presented in almost meaningless brackets and/or in aggregated form.
6. Overall, we found that only Facebook labelled any adverts as issue-based. However, we found examples of issue-based adverts on Facebook that were not clearly labelled. Neither Twitter nor Google made efforts to identify issue-based adverts. The issues addressed in this report are elaborated on in more detail in the accompanying essays.
You can read the full report here: ElectCheck 2019 A Report on Political Advertising Online During The 2019 European Elections