The Curious Case of Black Conservatives: Assessing the Validity of the Liberal-Conservative Scale

My paper, "The Curious Case of Black Conservatives:
Assessing the Validity of the Liberal-Conservative Scale among Black Americans" is now accepted at Public Opinion Quarterly. I'm stoked.

Science is wild, y'all.

A 🧵 on the paper + some thoughts. https://t.co/vvfMftaWGO pic.twitter.com/KDNXTDjeyg

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Although the overwhelming majority of Black Americans identify with the Democratic Party, surveys suggest that up to one-half of Black respondents identify as conservative.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

What accounts for this apparent disconnect, especially given general patterns of increased partisan and ideological sorting among the American public?

Look at plot of the correlation coefficiens between party id & ideology for white (black line) and Black Ams (red line). Wild! pic.twitter.com/tA9YfvUM2m

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Existing accounts posit that the group-centric nature of Black politics explains this so-called partisanship-ideology paradox. According to these accounts, Black Ams’ commitment to their racial group & the fear of social sanction from ingroup members explain this mismatch.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

I offer an alternative explanation. Using data from @electionstudies, I demonstrate that the misalliance between Black Americans’ partisan & ideological identities manifests because the liberal-conservative scale lacks construct validity for a large share of Black respondents.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

As scores on a measure of what I call “liberal-conservative familiarity” highlight, the terms “liberal” and “conservative” appear to resonate less with Black respondents relative to their white counterparts, thereby compromising the usefulness and validity of the measure.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

When we account for variation in respondents’ familiarity with these ideological terms, we observe that Black respondents familiar with the terms select ideological labels consistent with their expressed partisan preferences, thus resolving the partisanship-ideology paradox. pic.twitter.com/euCI3AM3SN

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

The findings I present have significant consequences for the social sciences and survey research. The standard measure of ideology has been included on the @electionstudies since 1972. Today, it features prominently in scholarship across academic disciplines.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

In political science, the liberal-conservative scale is one of the most frequently used measures in the study of American public opinion and forms the core of contemporary scholarship on polarization in the United States.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Even scholars not interested in ideology or polarization, per se, construct models of mass behavior that often include the measure. However, despite its popularity, the liberal-conservative scale cannot be used to understand the complexity and diversity of Black public opinion.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

A couple of important notes! pic.twitter.com/VRovjZbnSo

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

To be clear, the insights presented here matter beyond the study of Black politics. They must be considered by anyone who relies on the liberal-conservative scale in their scholarship or in their commentary about American politics.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

For example, to the extent that they intend to say something about polarization beyond the politics of white Americans, scholars must reassess the robustness of their claims in light of the evidence I present here.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Beyond polarization, the findings I highlight go a long way in helping us understand why other classic frameworks, including the Moral Foundations Theory, fail to replicate among Black samples.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

So, what to do? When using the standard measure to proxy for ideology, researchers should consider describing to respondents what “liberal” and “conservative” mean.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Alternatively, researchers should consider using respondents’ policy positions to generate measures of ideology, which allows for an examination of Black politics unconstrained by the liberal-conservative framework.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

To be sure, all of this is much harder than using the single-item liberal-conservative scale we are accustomed to using in our research.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

Still, this reminder from Mondak (2001, 224–25) is worth remembering here, too: “It does us no good to draft a survey that is easy for interviewers to administer and for respondents to complete if ambiguity exists regarding the meaning of the resulting data.”

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

At best, the liberal-conservative scale produces ambiguous results for Black respondents. At worst, it produces wholly misleading results. Regardless, researchers interested in the politics of Black people should take Black people at their word.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

When it comes to identifying as liberal or conservative, many simply “haven't thought much about it.” https://t.co/vvfMftaWGO

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023

I won't dwell on it, but the crazy review processes this paper went through reminded me of how absolutely wed to existing constructs & measures we are as social scientists, even in the face of an avalanche of data. I mean, I wish I could share the revision memos & responses here.

— Hakeem Jefferson (@hakeemjefferson) March 8, 2023