Americans 'support' the idea of tuition-free college: an exploration of sentiment and political identity signals otherwise. Issue 3 (16th March 2019)

• Record Type: Journal Article
• Authors: Collier, Daniel A.; Mishra, Shubhanshu; Houston, Derek A.; Hensley, Brandon O.; Hartlep, Nicholas D.
• Is Part Of: Journal of further and higher education. Volume 43:Issue 3(2019)
• Page Start: 347
• Page End: 362
• Publication Date: 2019-03-16
• Subjects: Sentiment analysis -- social media discourse -- tuition-free college -- political identity theory -- higher education -- identity; Education, Higher -- Great Britain -- Periodicals; Education -- Periodicals; 374.941
• DOI: 10.1080/0309877X.2017.1361516
• Languages: English
• ISSNs: 0309-877X

Abstract

Abstract: In January 2015, President Obama captured headlines in the US by announcing America's College Promise (ACP), a policy that would reverse four decades of privatisation in higher education by making community colleges 'tuition-free'. This research explores the conversation that unfolded across the Facebook pages of various media sources. Key phrases associated with sentiments and communication styles are uncovered using a Bag of Words (BoW) technique. Next, guided by political identity theory, the researchers employ logistic regression to explore variable effects (e.g. source, gender, race, age and political leaning) on communicating: (1) Against ACP, (2) in a Civil manner, and (3) Against/Civil/On-Topic. BoW models suggest those against the policy utilise anti-free-ride, privatisation rhetoric, whereas those using uncivil language attack commentators and groups who counter personally-held political beliefs – while also introducing non-sequiturs from other policies (e.g. healthcare and immigration). The combined communication styles unlocked tokens not found in larger sentiments, such as concerns over student loan debt. Logistic regression illustrates that, depending on sentiment or communication style explored, political identities and memberships associated with source and political alignment significantly affected likelihoods of communicating in the conversation. These findings are linked back to political identity theory.