A long-held apocryphal belief is that individual’s political attributes change substantially over the life course and become more conservative with age. However, scholarly evidence in the literature indicates that the opposite is true: People’s politics are remarkably constant throughout the life course. This paper seeks to explore the reasoning behind the lack of political development and heterogeneity throughout the adult life course. This phenomenon is explored through the lens of homophily, the notion that people have a propensity to surround themselves with others who share similar attributes. When people are involved in politically homophilous groups, their exposure to diverse political perspectives is lower and the likelihood, therefore, of an ideological shift is less likely. With the advent of social media, there unquestionably are many more accessible sources of diverse political perspectives. Given the availability of opportunity, do people engage in politically diverse groups, or do they selectively expose themselves only to groups and views consonant with their own political views? Using 155.6 million communications from 207,556 unique online communities on the social media platform, Reddit, I develop a model using principal component analysis and linear regression to predict user politics and measure the political homophily of each user’s self-selected network. This model is an improvement of past political classifiers, wherein I treat politics as a continuous variable on a spectrum as opposed to hardline categorical variables. Results indicated that users are far more likely to self-select into groups consonant with their own politics, disproportionately skewing the diversity of their political exposure.