Lie Machines: How to Save Democracy from Troll Armies, Deceitful Robots, Junk News Operations, and Political Operatives by Philip N. Howard

Author:Philip N. Howard [Howard, Philip N.]
Language: eng
Format: epub
ISBN: 9780300252415
Publisher: Yale University Press
Published: 2020-04-30T23:00:00+00:00

Mining Data for Political Inferences

Understanding how the modern political campaign works requires some knowledge of the data-mining industry, because this is the industry that supplies the information that campaign managers need to make strategic decisions about whom to target, where, when, with what message, and over which device and platform.15 Most citizens do not know how much information about them is held, bought, and sold by campaigns, advertising firms, political consultants, and social media platforms. Sometimes political campaigns have clear sponsors and declare their interests, but on other occasions such disclosures are hidden. Sometimes the public can see what kinds of advertisements are produced and distributed during a campaign season, but at other times the advertising is discrete and directed and not archived anywhere. Sometimes campaigns deliberately spread misinformation, lies, and rumors about rivals and opponents.

Raw data is an industry term for information that has not been processed or aggregated in some way. To make political inferences about someone, you usually need to take several kinds of data and put together an index that makes people more comparable to one another. For example, you could label someone a liberal if you acquire a list of their magazine purchases and find that they subscribe to a magazine where liberal commentators sometimes publish essays. But if you can find more raw data and learn that they are registered with a conservative party, that they have only ever voted for such a party, and that all their friends also vote for conservative politicians, you can make more precise inferences that this person should really be labeled a conservative. So the raw data about many of our behaviors, such as magazine purchases, party registration, voting history, and community ties, must be aggregated and weighed before reasonable political inferences can be made. And the person in this example would probably be scored as pretty conservative once all the relevant behaviors were evaluated.

Credit card data often helps kick-start such labeling exercises. Even though the transactions are between a buyer and a seller, credit card companies sell access to the transaction data to many third parties, including advertisers, public relations firms, and political consultants. Some third-party companies pay for access to the raw data, which they then must analyze themselves. Others prefer to pay for the finished analysis, so they will hire consultants like the now-defunct Cambridge Analytica to process the data and report on the important trends. The advantage of working with data brokers is that they can merge data sets from multiple sources and make even more powerful inferences.

Data mining has been an active industry for decades, so there are many kinds of aggregate categories available from data brokers.16 The difference that social media makes is that these services provide new kinds of data that allow more detailed insight into how particular people think and feel.

For example, a traditional political consultant might have compiled data on voting history and magazine purchases and be able to provide a list of who in a district is liberal and who is conservative.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.