Skip to main content

Some districts harder to track than others, but not many competitive races are affected

(MIDDLETOWN, CT) July 24, 2019 – The Wesleyan Media Project obtains ad data from Kantar Media/CMAG, which represents the most comprehensive and systematic collection on the content and targeting of political advertisements.

There is, however, one feature of Kantar’s data that we want to highlight for our data users: each media market in the country is either a “discovery” market or a “non-discovery” market. Discovery markets have technology that allows them to compare new sound wave patterns from ads to the existing database to determine whether an ad is one the system has heard before. If the ad is new, then it adds it to the database.

Markets that do not have this capability (i.e., non-discovery or tracking markets) can only record instances of ads that are already in the system. Kantar has distributed discovery markets such that there is at least one discovery market for each state (and frequently multiple markets based typically on population size), but we sometimes find that, for example, congressional races that are in small markets may not always have advertising in Kantar’s dataset because the ads ran only in non-discovery markets. This should be less of a problem for statewide campaigns and product advertising, but it can be an issue for congressional and local races.

The Wesleyan Media Project has obtained and is providing here information on the discovery status of all 210 media markets for each election cycle since 2010. Table 1 shows a breakdown of the number of markets that are discovery and non-discovery from 2010 to 2018. A market-level breakdown is available in Table 2. While we have data for each election cycle, we do not have the exact dates that the status of markets switched (i.e., from non-discovery to discovery). 

Table 1: Discovery vs. Non-Discovery Market Counts 2010 – 2018

Election
Cycle
Discovery
Markets
Non-Discovery
Markets
201012387
201212585
201412585
201612684
201813476

Table 2: Media Market Discovery Status 2010 – 2018

DMAMedia MarketDominant
State*
20102012201420162018Status
743Anchorage AKAK11111Always discovery
745Fairbanks AKAK00000Always non-discovery
747Juneau AKAK00000Always non-discovery
630Birmingham ALAL11111Always discovery
606Dothan ALAL00000Always non-discovery
691Huntsville ALAL11111Always discovery
686Mobile ALAL11111Always discovery
698Montgomery ALAL11111Always discovery
670Fort Smith ARAR11111Always discovery
734Jonesboro ARAR00000Always non-discovery
*Primary state covered by the media market. Many media markets cover multiple states.
Click here to download the file as a .csv.
Showing 1 to 10 of 210 entries

The Impact on Ad Tracking

As mentioned above, this discovery market issue is not likely to have much of an impact on analyses of statewide advertising efforts (presidential, U. S. Senate, gubernatorial, etc.) because any particular creative is likely to air in multiple markets, some of which are discovery. The presence of non-discovery media markets, then, is primarily a concern for U.S. House races, as well as for other local races such as mayor and for marketing efforts where the ad message is likely to differ by media market.

Table 3 lists congressional districts where greater than 50% of constituents lived in a non-discovery media market in 2016. We’ll illustrate the issue by looking specifically at Mississippi’s 4th district. This district lies almost entirely within the Hattiesburg and Biloxi media markets, which are both non-discovery markets. This means that political ads that air in these markets will only be recorded by Kantar Media/CMAG if they have already been “discovered” by another media market. So if the ads for the U.S. House race in this district only aired in the Hattiesburg and Biloxi media markets, Kantar would not have ANY record of those ads.

The table also lists the competitiveness of the congressional race in the district, with higher scores indicating the most competitive races. Most races were not very competitive and thus probably saw minimal television advertising, but a few races, such as those in Michigan’s 1st congressional district, were competitive and thus likely saw substantial advertising.

Table 3: Flagged Congressional Districts with Competitive Scores 2012-2016

Congressional
District
2012 2014 2016
TEXAS-19111
MICHIGAN-01424
TEXAS-13111
MISSISSIPPI-04111
OREGON-04111
NEBRASKA-03111
SOUTH CAROLINA-07111
KANSAS-02111
TEXAS-11111
LOUISIANA-05111
CALIFORNIA-01111
MONTANA-00222
NEW YORK-22114
MINNESOTA-01111
OREGON-02111
WASHINGTON-04111
WEST VIRGINIA-01111
WISCONSIN-03111
WISCONSIN-07211
IDAHO-02111
NEBRASKA-01111
WYOMING-00111
MAINE-02134
ILLINOIS-16111
WEST VIRGINIA-03241
NEW YORK-23121
NORTH CAROLINA-07411
IOWA-04311
MISSISSIPPI-01111
Note: Congressional districts where greater than 50% of constituents lived
in a non-discovery media market in 2016. Competitive Scores come from
CQ/Roll Call’s October pre-election ratings.
1 = safe seats,
2 = Democratic or Republican favored seats,
3 = Democratic or Republican leaning seats,
4 = Too-close-to-call races.

The exact boundaries of media markets are determined by Nielsen, and WMP does not have access to them through our contract with Kantar Media/CMAG. Therefore, we have approximated the media market boundaries with counties as building blocks using resources in the public domain, not the proprietary Nielsen database. Calculations here related to percent population in a particular congressional district and media market are approximations.


About This Report

The Wesleyan Media Project (WMP) provides real-time tracking and analysis of political advertising in an effort to increase transparency in elections. Housed in Wesleyan’s Quantitative Analysis Center – part of the Allbritton Center for the Study of Public Life – the Wesleyan Media Project is the successor to the Wisconsin Advertising Project, which disbanded in 2009. It is directed by Erika Franklin Fowler, associate professor of government at Wesleyan University, Michael M. Franz, professor of government at Bowdoin College and Travis N. Ridout, professor of political science at Washington State University. WMP staff include Laura Baum (Project Manager), Dolly Haddad (Project Coordinator) and Matthew Motta (Research Associate).

Periodic releases of data will be posted on the project’s website and dispersed via Twitter @wesmediaproject. To be added to our email update list, click here.

Please use our contact form to submit any questions or comments about this post.

Close Menu