Scholars have shown that social relations between states have a meaningful impact on politics, such as the case of state policy diffusion—when U.S. states “borrow” policies from other U.S. states (Bricker and LaCombe 2020).
One way of measuring social relationships between states is through social media, as individuals are increasingly connecting online. But what factors predict Facebook friendship ties between states? Using State Networks’ data compiled at IPPSR, I explored the various factors that predict the level of friendship between residents of the U.S. states, as based on another powerful research tool, the Social Connectedness Index (SCI) (Bailey et al. 2018).
The State Networks dataset held at IPPSR is a compilation of many state-to-state relational variables, including geographic, political, and demographic information about state governments and their residents. The Social Connectedness Index is a relative measure of Facebook friendship links between any two given counties in the United States. The authors of the Social Connectedness Index developed the measure and explored the geographic, economic, and social features of online connection. I aggregate this measure to the state level (average of all counties in a state) to show how religion, race, and partisanship are associated with this measure of online connection.
First, and as the SCI authors demonstrate (Bailey et al. 2018), the level of Facebook friendship between states is highly associated with the geographical distance between states. A simple statistical model shows that when states share a common region, the pair’s SCI measure increases by an average of more than nine points. States in the Northeast are clustered together, as are states in the West, Midwest, and South. While Texas, Oklahoma, and Arkansas also share closer ties with Midwestern states, states in the South Atlantic are more closely related to each other. Additionally, states such as California and New York are more central to the network than are less populous states such as South Dakota and Nebraska.
Although all states have some degree of SCI connection to each state, some ties are stronger than others. The District of Columbia is the most connected—or most central—in the network, followed by Maryland, Massachusetts, Connecticut, Delaware, and California Each circle represents a U.S. state (abbreviations used in figures, includes the District of Columbia), each line represents the SCI connection between a state pair, and each graph is colored by the covariates of interest: region, religion, race, and partisanship. In each figure below, the top 25 ties—those state pairs who share the highest average Social Connectedness Index—are indicated by darker lines between them.
State religious makeup is also significantly associated with the level of Facebook friendship between states. The next figure visualizes these ties by coloring states according to the population of Christian identifiers in each state. This measure of Christian citizenship includes all denominations of Christianity. As shown in the figure, states with a higher population of Christians tend to appear closer together than those states with a smaller amount of citizens identifying as Christian. Additionally, state pairs with more similar religious makeups — including differences in the number of proportion perhaps of citizens identifying as Jewish, Muslim, no religion, and more — share a statistically higher Social Connectedness Index on average.
Of course, region and religion are not the only measures of similarity between states. The next figures show how similar state race and partisanship relate to Facebook friendship. States that share similar percentages of white citizens tend to cluster together at the exterior of the network; they are less central actors, meaning they are, in total, less socially connected to other states. States with higher populations of non-white citizens are more central and are closely related to each other. With regards to partisanship, states with more Democratic citizenries are closely related, but there is only a weak relationship between state partisanship and social centrality.
While I have demonstrated that these state level attributes are associated with Facebook friendship, the State Networks data allows for exploration of the factors driving social connectedness between states. Additionally, it allows researchers to further examine how these social ties predict policy outcomes. For example, given that SCI is particularly related to religion, we might be interested in understanding how social similarity and connectedness maps onto to the diffusion of religious freedom policies in each state. We might also ask how well this aggregate measure of social connectedness through Facebook friendships relates to citizens’ perceived similarity of other states. We can also compare how citizen-level attributes compare to their state government. The State Networks data allows us to explore how social, geographical, and economic ties predict political and policy outcomes.
Bailey, Michael et al. 2018. “Social Connectedness: Measurement, Determinants, and Effects.” Journal of Economic Perspectives 32(3): 259–80.
Bricker, Christine, and Scott LaCombe. 2020. “The Ties That Bind Us: The Influence of Perceived State Similarity on Policy Diffusion.” Political Research Quarterly: 106591292090661.