Essays on Consumer Preferences Online

Abstract

This dissertation is composed of three separate empirical analyses. Each is a separate analysis and article.

Theory suggests information about quality is especially important for experience goods. However, in an environment with multiple sources of information it is unclear which sources of information will be valued and if those sources will be valued differently conditional on product characteristics. Chapter I examines the impact of heterogeneous sources of information about quality on the demand for experience goods. In particular, I study the effects of user and expert reviews on the demand for video games online and how the valuation of information may vary based on if the game is produced by an 'independent' or 'major' developer. Using a unique dataset from the online video game platform Steam, I find consumers do value information about quality, they value information about quality from both user and expert reviewers, and they value user and expert reviews for products from independent developers and for products from major developers.

As individuals consume an increasing amount of news on social media it is important to understand consumer preferences, a key component of demand, for individuals on this platform. Chapter II examines the importance of ideological similarity between a news source and an individual and the importance of the reliability of a news source in individual's consumption decisions in the news market on social media. The findings suggest, conditional on who an individual chooses to follow, liberal individuals have an increasing preference for news as it becomes more conservative and no preference for news as it becomes more liberal. Liberals are also found to have a preference for less reliable news which may be a preference for sensationalism. Alternatively, conservatives have no preference for more liberal or more conservative news and have no preference for more reliable or less reliable news.

Chapter III provides greater insight into Twitter accounts, account networks, and activity on the platform. I find, on average, active US Twitter accounts have an audience of 6,000 accounts and receive information from 1,500 accounts. High-profile accounts have much larger audiences than lower-profile accounts. High-profile accounts constitute a very low percentage of the total number of accounts. Both high and lower-profile accounts have similar levels of activity on Twitter. Activity related to news constitutes a small percentage of total activity on Twitter. Finally, the majority of activity on Twitter is reacting to already posted content, and verified accounts are more likely to post original content than non-verified accounts.