Introducing new election-related advert knowledge units for researchers
We had previously announced that starting February 1, 2021, we would be making targeted information on more than 1.3 million social topics, election and political Facebook ads available to academic researchers for the first time. This data package contains ads that were run in the three months leading up to Election Day in the United States, August 3 to November 3, 2020, and are accessible via the Facebook Open Research and Transparency (FORT) platform.
As part of this launch, we will share access to two new data sets:
- Ad Targeting Record: Contains the targeting information selected by advertisers who ran social, election, and political ads between August 3, 2020 and November 3, 2020. We’re excluding ads with fewer than 100 impressions. This is one of several steps we have taken to protect user privacy.
- Ad Library Record: Contains the social, election, and political ads that are part of the Ad Library product. It’s included so researchers can analyze the ads and targeting information in the same environment. In other words, this data set is a copy of the corresponding ad library data made available on the FORT platform and is different from the ad library API product.
We are posting these datasets through the FORT platform to allow academic researchers to better study the impact of Facebook products on elections. The platform contains validated data protection and security measures and has been tested for penetration by internal and external experts.
How to access the records
To request access to these records, please fill out this form.
After you provide your details, Facebook will contact you with details on what to do next, including details of the Facebook Research Data Agreement (RDA) and the ID verification process. Once you and your university have signed the RDA and your ID has been verified, you will have access to the Facebook Open Research and Transparency platform and these two new records.
Learn more about the election related ad records
Ad targeting record
This includes the targeting options selected by advertisers when creating an ad. You can find more information about Facebook ads here.
Overview of the targeting options
More information about targeting options can be found here.
- Location: cities, municipalities and countries
- Demographics: age, gender, education, job title, and more
- Interests: Interests and hobbies of the people advertisers want to reach – these help make ads more relevant
- Behavior: consumer behavior, e.g. B. Previous purchases and device usage
- Connections: Audiences based on people connected to the advertiser’s Facebook page, app, or event
- Custom Audiences: Options that an advertiser can use to find their existing audiences among People on Facebook, such as: B. via customer lists, website or app traffic or engagement on Facebook. Learn more about the different types of custom audiences here. In the data set, we indicate whether the custom target group used is specifically based on a customer list.
- Lookalike Audiences: Help advertisers reach out to new people who may be interested in their business because they are similar to their existing customers. Learn more about how advertisers create like audiences.
Where applicable, we will indicate whether the targeting options were selected for inclusion or exclusion targeting.
Important Ad Targeting Record Considerations
Location data
This record contains the location targeting selected by an advertiser for the ad. Advertisers can enter location targeting in a number of ways, including: B. by selecting postal codes, countries, certain market areas (DMAs) or pindrops / addresses / locations with a certain radius.
We provided the location targeting chosen by advertisers. However, when an advertiser selects an address, city, or town, we instead record the type of selection, the city they fall in, and the radius the advertiser specified in the record. Larger geographic areas such as postcodes, cities or countries are included in the data set, but an address, town or city is not.
The following examples illustrate the transformations:
On the left, the advertiser’s targeting selection. on the right the transformation found in the data set. You will see addresses, pindrops (latitude and longitude), and locations are replaced with the edited text
,- Seattle + 5 miles -> Seattle (+5 miles)
- 1 North Almaden Blvd, San Jose + 5 miles ->
San Jose (+5 miles) - 95110, San Jose -> 95110, San Jose (so no change)
- 95110, San Jose + 5 miles -> 95110, San Jose (+5 miles)
- 37.335080; -121.895480 + 5 miles ->
San Jose (+5 miles) - Acme Park, San Jose (+ 1 mile) ->
San Jose (+1 mile)
Linking targeting data to ad library data
If researchers want to understand more information about an ad (its motivation, spending, etc.) and want to link that ad to their targeting information, they can perform a join between the ad_archive_id column of ad_archive_api (ad library data) and the archive_id column of the ad_library_targeting table (Ad Targeting data).
However, there may be inconsistencies between the ad_archive_api and ad_library_targeting tables:
If you join the two tables, you’ll see that the target table contains ads that don’t have a corresponding entry in the ad library table (or vice versa).
This is because advertisements can be classified as political / non-political long after they have been placed. In this case, the ad library will be updated. However, since the target data set is a one-time data release that took place on January 22nd, changes after this date will not be taken into account.
However, this doesn’t happen often. For example, when we did the one-time publication of targeting data on January 22, 2021, we found nine ads (out of approximately 1.3 million) that were in the targeting record but not in the ad library because they were later on found that they had been mistakenly flagged as political ads and removed from the library. Since the targeting record has already been generated at this point, these nine ads were included in this record.
We will be monitoring this situation for the next few weeks. If we find that there are a large number of ads with such a problem, we will consider whether to update this record with these new ads.
Ad library record
The ad library record contains the following fields:
- ad_archive_id: ID for the archived ad object
- ad_creation_time: The UTC date and time that someone created the ad. This is not the same time as when the ad went live. Contains the date and time separated by T. Example: 2019-01-24T19: 02: 04 + 0000, where +0000 is the UTC offset.
- ad_creative_body: The text that will appear in the ad. Usually 90 characters. See reference, ad motif.
- ad_creative_link_caption: If an ad contains a link, the text that appears in the link
- ad_creative_link_description: If an ad contains a link, a text description that will appear next to the link, e.g. B. a label or description
- ad_creative_link_title: If an ad contains a link, any title is given
- ad_delivery_start_time: Date and time at which an advertiser wants Facebook to start serving one of the ads. Is provided in UTC as in ad_creation_time
- ad_delivery_stop_time: The time at which an advertiser wants to stop showing their ad. If this is empty, Facebook will serve the ad until the advertiser stops it or spends their entire campaign budget. In UTC.
- ad_snapshot_url: String with URL link that shows the archived ad
- Currency: The currency in which the ad was paid for as an ISO currency code
- Impressions: A string that indicates the number of times the ad has created an impression. In areas of <1000, 1K-5K, 5K-10K, 10K-50K, 50K-100K, 100K-200K, 200K-500K, >1M.
- Demographic Distribution: The demographic distribution of the people that the ad will reach. Assuming age groups and gender:
- Age groups can be one of the following: 18-24, 25-34, 35-44, 45-54, 55-64, 65+
- The gender can be one of the following strings: “Male”, “Female”, “Unknown”
- Funding Entity: A string containing the name of the person, company, or organization that provided funding for the ad. Provided by the buyer of the ad.
- page_id: ID of the Facebook page on which the ad was shown
- Page Name: Name of the Facebook page on which the ad was shown
- region_distribution: Regional distribution of the people reached by the ad. Expressed as a percentage and when regions are at subland level.
- Expenses: A string representing the amount of money spent on running the ad, as specified in the currency. This is indicated in the following areas: <100, 100-499, 500-999, 1K-5K, 5K-10K, 10K-50K, 50K-100K, 100K-200K, 200K-500K, >1M.
- is_active: binary; describes whether an ad was active on January 22, 2021
- Countries reached: Facebook delivered the ads in these countries. Provided as an ISO country code.
- publisher_platforms: Search for ads based on whether they appear on a specific platform like Instagram or Facebook. You can provide a platform or a comma-separated list of platforms.
- Potential Reach: This is an estimate of the audience size who will be eligible to see this ad. It is based on targeting criteria, ad placements, and the number of ads served on Facebook apps and services in the last 30 days. This is not an estimate of how many people will actually see this ad, and the number may change over time. It is not designed to meet population or census estimates.
Via the Facebook Open Research and Transparency platform
The Facebook Open Research and Transparency (FORT) platform makes responsible research easier by providing flexible access to valuable data. The platform was created with validated data protection and security protection measures such as data access controls and tested for penetration by internal and external experts.
The FORT platform runs on an opinion-based version of JupyterHub, an open source tool that is widely used by the academic community. The FORT platform supports several standard programs, including SQL, Python and R, as well as a special bridge to certain Facebook Graph APIs.
Publishing guidelines
Researchers can publish research done with this data without Facebook’s permission. As with other research conducted under the Research Data Agreement, Facebook has the right to review (not approve or reject) research results prior to publication, and to remove any confidential or personally identifiable information.
Comments are closed.