The Democratic and Republican parties are sometimes thought of as mirror images of each other on the right and left, but the coalitions of voters that make up each party are quite different. In this post, we compare the demographic profiles of those who identify with each party drawing on data from the Grinnell College National Poll (GNCP) conducted in November of 2018.
Any attempt to describe the members of a political party involves a variety of challenges. On the one hand, research has generally found that an individual’s party identification is a social identity that is highly stable over time (Green, Palmquist and Schickler 2002). Once a young adult identifies as either a Democrat or Republican, that choice is likely to endure for the remainder of their life. On the other hand, parties themselves are continually in a state of change as older generations exit the electorate and new generations come of age (Miller and Shanks 1996). Major political events can also impact parties by breaking apart the coalitions that make up the parties.
Both parties have changed dramatically over time – a fact well illustrated by the history of the Democratic party. In the Civil War era, it defended the institution of slavery, favored a weak national government, and after the war, opposed protecting the rights of emancipated African Americans (Foner 1990, Key 1984). In the New Deal era, Democrats were divided between a northern and southern wing. The old Democratic party of the Civil War era endured in the south, while the northern wing of the party took on the cause of workers and advocated for a powerful national government to intervene in the nation’s economy (Polsby 2004). In time, northern Democrats became supporters of civil rights for African Americans, fracturing the party in two and ultimately leading southern Democrats to abandon the party (Schickler 2016). Today, Democrats are known as the party of economic and cultural liberalism – a complete reversal from their positions of the 1860s (Brownstein 2007).
Changes in the coalitions of voters supporting each party reflect this evolution. For example, African American voters were more likely to support the Republican party in the wake of the Civil War, but shifted their support in favor of Democrats as Democrats became identified with support for civil rights (Valelly 2004). Today, the Democratic party is more likely to be viewed as the party that is receptive to emerging claims of rights from historically disenfranchised groups. Parties – and the voters that make them up – change over time.
Observers have argued that the election of Donald Trump as president on a platform of opposition to immigration and economic nationalism has sparked new movement in the coalitions of voters supporting each party (Sides 2017). In this account, voters with lower levels of education and income, and who are opposed to immigration, shifted toward the Republican party in 2016. These changes have injected a new element of uncertainty into the political landscape, as analysts try to assess which party has the advantage in a changing electorate.
Here, we use data from the GCNP to capture a brief snapshot of the two parties as they stood in November of 2018. While we cannot compare these data with how the parties looked in the past, we can draw a demographic portrait of each party as it stood at that moment that we can use as a basis of comparison for the future.
The Grinnell College National Poll was conducted November 24 to 27, 2018, for Grinnell College by Selzer & Co. of Des Moines, IA. It is based on telephone interviews with 1,000 U.S. adults ages 18 or older, including 828 likely voters in the 2020 election and 769 self-identified voters in the 2018 midterm election.
Party identification was determined using Question 102 of the poll: “In politics as of today, do you consider yourself a Republican, Democrat, or independent?” Respondents who answered with “independent” received a follow up question: “Do you lean more toward the Republicans, or more toward the Democrats, or are you totally independent?” Respondents who answered the initial question with either “Democrat” or “Republican” were asked a separate follow up question: “Do you consider yourself a strong Democrat/Republican?”
Previous research finds that voters who “lean” toward one party have similar voting behavior as those who directly identify with a party. For that reason, we categorize respondents who identified with a major party or who “lean” toward that party as a member of that party. Respondents who stated they are “totally independent” are categorized as independents.
We begin by presenting an overview of the two parties, comparing the major demographic characteristics of those who affiliate with each party. Overall, 40 percent of respondents in the data are Republicans and 40 percent of respondents are Democrats. A total of 20 percent of respondents are independent.
Our first step was to conduct a logistic regression to assess the relationship between a respondent’s party identification and a standard set of demographic variables. For this analysis, our dependent variable was party identification, with Republicans coded as 0 and Democrats coded as 1 (independents were excluded from this analysis). We do not present the full results of the regression analysis here. We found statistically significant differences between the parties for five variables. Respondents with high levels of education were likely to be Democrats, while respondents with lower levels of education were likely to be Republicans. African Americans were more likely to identify as Democrats than white respondents, but we did not find statistically significant relationships to party identification for respondents who identified with other groups. Respondents who described their political ideology as either “Moderate” or “Liberal” were more likely than Conservatives to identify with Democrats rather than Republicans. And, respondents who scored high on a nativism variable we constructed for this analysis were more likely to identify as Republicans than Democrats. We describe this variable in more detail below.
Next, we turn to a more detailed discussion of our findings for each variable that had a statistically significant relationship with party identification.
Education has five categories, with the lowest indicating a high school education or less and the highest indicating post-graduate education. Our data show that Democrats are more educated on average than Republicans. According to the graph above, 48 percent of Republicans have a high school education or less, compared to 35 percent of Democrats. Additionally, 36 percent of Democrats have a bachelor’s degree or above, compared to 28 percent of Republicans.
Income has five categories, with the lowest indicating an income $25,000 a year or under, and the highest indicating an income of $100,000 a year or over. Our data show that Republicans have higher incomes than Democrats on average. Nearly 29 percent of Democrats have incomes under $25,000, compared to 13 percent of Republicans. In addition, 26 percent of Republicans have incomes above $100,000, compared to just under 20 percent of Democrats.
The GCNP asks respondents to choose “what racial or ethnic group” they identify with most. Here, we report our findings for groups with over 100 respondents. We consolidate any group with less than 100 respondents, which includes respondents who identified as Asian, Native American, multi-racial, other, or refused, into a group we call “Additional.” Small samples sizes for each of these groups mean that specific findings about them cannot be reported with confidence.
Our data show that the Democratic party is more racially heterogeneous than the Republican party. Nearly 46 percent of Democrats self-identify as members of racial minority groups or fall into the “Additional” category, compared to 29 percent of Republicans. Twenty percent of Democrats identify as black and 15 percent identify as Latinx. By contrast, 4 percent of Republicans identify as black while 18 percent identify as Latinx.
Ideology has three categories: conservative, moderate, and liberal. Our data show more ideological heterogeneity among Democrats than Republicans. Nearly 70 percent of Republicans say they are conservative, compared to 43 percent of Democrats who say they are liberal. In addition, 45 percent of Democrats are moderate, compared to only 23 percent of Republicans.
Finally, recent research (Sides 2017) finds that nativist attitudes among voters are associated with support for Donald Trump and a political agenda of limiting immigration to the United States. To assess the degree of nativist sentiment among Democrats and Republicans, we follow an approach modeled after Citrin and Wright (2009). They describe two competing models of the way that Americans think about American national identity. The first is the civic model, in which respondents define being American as adhering to “certain fundamental values and institutions” (3). The second is an ethnic model in which being an American is defined by an individual’s descriptive characteristics, such as their religion or language. The GCNP asks a battery of questions about the kinds of characteristics respondents associate with being a “real” American to assess whether their attitudes better fit one model or the other.
We created a variable called “nativism” by combining the responses from four questions in which respondents are asked to evaluate the importance of a set of descriptive characteristics to being a “real” American: being born in the United States; living in the United States for most of one’s life; the ability to speak English; and to be of the Christian religion. Respondents who rated having all of these characteristics as “very important” scored highest on the nativism variable. Respondents who rated these characteristics as “not important” scored lowest. The rescaled variable ranges from 0 to 1, with 0 indicating the lowest level of nativism and 1 the highest.
Consistent with past research, our data show a statistically significant difference in the level of nativist sentiment measured among Republicans and Democrats. The mean nativism score for Republicans was 0.52. compared to 0.33 for the Democratic Party, indicating that Republicans were more likely than Democrats to identify characteristics associated with the ethnic model of American national identify as being important to being a “real” American.
Using the November 2018 Grinnell College National Poll data, we present a brief overview of the two political parties and analyze their major differences. Specifically, we analyze how education, income, racial demographic, ideology, and nativism differ between two parties. We find that Democrats are more educated, racially diverse, and ideologically heterogeneous than Republicans. Republicans are more likely to have higher levels of income than Democrats, lower levels of education, and are more likely to adopt nativist views of what it means to be a “real” American.
Brownstein, Ronald. 2007. The Second Civil War: How Extreme Partisanship Has Paralyzed Washington and Polarized America. New York: Penguin Press.
Citrin, Jack and Matthew Wright. 2009. “Defining the Circle of We: American Identity and Immigration Policy.” The Forum, Volume 7, Issue 3.
Foner, Eric. 1990. A Short History of Reconstruction, 1863-1877. New York: Harper and Row.
Green, Donald, Bradley Palmquist and Eric Schickler. 2002. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven: Yale University Press.
Key, V.O. 1984. Southern Politics in State and Nation. Knoxville: University of Tennessee Press.
Miller, Warren and J. Merrill Shanks. 1996. The New American Voter. Cambridge: Harvard University Press.
Polsby, Nelson. 2004. How Congress Evolves: Social Bases of Institutional Change. New York: Oxford University Press.
Schickler, Eric. 2016. Racial Realignment: The Transformation of American Liberalism, 1932-1965. Princeton: Princeton University Press.
Sides, John. 2017. “Race, Religion, and Immigration in 2016: How the Debate over American Identity Shaped the Election and What It Means for a Trump Presidency.” Democracy Fund Voter Study Group. https://www.voterstudygroup.org/publication/race-religion-immigration-2016
Valelly, Richard. 2004. The Two Reconstructions: The Struggle for Black Enfranchisement. Chicago: University of Chicago Press.
Peter Hanson is associate professor of political science at Grinnell College.
Yuejun Chen ’20 is a double major in economics and mathematics at Grinnell College, and a research assistant with the Grinnell College National Poll.