- Party alone can correctly predict Trump approval 81% of the time
- Additional demographics improve predictions only slightly
- Effects of race are apparent among Republicans, independents
WASHINGTON, D.C. -- It is well-established that President Donald Trump's ratings are the most politically polarized Gallup has measured for a president. Trump's support also has shown consistent differences by gender and by race and education, among other characteristics. What has been less clear is whether those subgroup differences mainly reflect the party leanings of those groups or if they demonstrate an appeal Trump has to certain groups that transcends partisanship.
A new Gallup analysis of 2019 Trump job approval data indicates that subgroup differences largely originate in the party orientation of those groups. Once party is taken into account, there is little to no variation among demographic or attitudinal subgroups. To the extent subgroup differences exist, they are more apparent among independents than among Republicans or Democrats.
The analysis uses statistical models to predict whether a respondent approves or disapproves of Trump. A baseline prediction of 58% correct is used to evaluate the effectiveness of various alternative models. The 58% figure corresponds to how often a model would be correct in predicting Trump job approval not taking into account any prior information about respondents. Instead, the baseline model merely predicts the most common response (disapproval) for every respondent.
A second model including only a measure of respondents' party identification as Republican, Democrat or independent correctly predicts 81% of respondents' opinions of Trump. In other words, by just knowing a person's partisanship, one achieves a substantial increase in accuracy of more than 20 percentage points over the baseline model.
A third model -- which includes party identification but also other respondent characteristics such as political ideology, age, race, gender, education, region, marital status, income and urban/suburban/rural residence -- can provide insight into how much factors beyond party predict Trump job approval. This model only improves prediction accuracy by five percentage points over the party-only model, correctly predicting 86% of opinions. In other words, after the effects of party are taken into account, few other characteristics have an independent effect on opinions of Trump.
|Party identification only||81|
|Full demographic model||86|
|Note: Analysis uses combined data from five 2019 Gallup polls that interviewed more than 6,000 U.S. adults. The baseline naive model includes no predictive variables and predicts the modal category (disapproval) for all respondents. The party-only model includes a variable for party identification as a Republican, independent or Democrat. The full model includes variables for party identification, gender, age, race, ideology, education, region, marital status, household income and place of residence.|
The third model does identify ideology and race as the factors that have the strongest relationships to Trump job approval after party identification. This finding is consistent with prior Gallup analysis of Trump job approval.
Little Variation in Approval for Subgroups of Democrats and Republicans
The relative effects of party, race, ideology and other characteristics on views of the president are apparent when looking at Trump job approval within party subgroups. Almost all Republican subgroups show Trump's approval ratings between 85% and 90%. The major exceptions are by race and ideology. Twenty points separate the approval ratings of white (92%) and nonwhite (72%) Republicans. Likewise, conservative Republicans are more likely than moderate or liberal Republicans to approve of the job Trump is doing.
Democratic subgroups tend to exhibit even less variation in their opinions of Trump than Republican subgroups do, with nearly every Democratic subgroup registering approval in the single digits. Ideology is a factor in the sole Democratic exception, with 11% of conservative Democrats approving of Trump.
In instances in which demographics are related to opinions of Trump, they are most apparent among independents. For example, the gender gap among independents is 14 percentage points (40% of men and 26% of women approve of Trump). The gender gap is only six points among Democrats, while there is no gender gap among Republicans. There are substantial differences in Trump approval ratings among independents of different ideologies -- conservative independents (57%) are more than twice as likely as moderate (28%) and liberal (15%) independents to approve of the job Trump is doing. Race has a similarly strong relationship to Trump job approval among independents, as 44% of whites but only 18% of nonwhites think he is doing a good job.
|All U.S. adults||Republicans||Independents||Democrats|
|All U.S. adults||41||89||34||5|
|Race by education|
|White college grad||38||88||34||4|
|White college nongrad||59||94||49||8|
|Place of residence|
|Note: January-March 2019 Gallup polls; Income definitions are as follows: Upper income (annual household income of $100,000 or more), Middle income (annual household income of $40,000-$99,999), Lower income (annual household income of less than $40,000)|
Subgroup Members' Opinions of Trump Consistent With Their Party Leanings
Further evidence that partisanship appears to be the primary driver of subgroup differences is found when looking at the party leanings of groups that are much more approving or much less approving of Trump. For example, one of Trump's strongest groups is whites without a college degree -- 59% of noncollege whites approve of Trump. As shown above, there is little variation within party groups among whites with and without a college degree. But noncollege whites are one of the most Republican subgroups, with 42% identifying as Republicans and only 21% as Democrats.
The gender gap in views of Trump is also largely driven by the Democratic tilt to women's underlying party identification (39% Democrat, 26% Republican), in addition to stronger support for Trump among independent men than independent women.
Differences in Trump's approval by marital status and household income also appear to stem largely from those subgroups' partisan leanings.
|All U.S. adults||28||40||31|
|Race by education|
|White college grad||28||37||35|
|White college nongrad||42||37||21|
|Place of residence|
|Note: Combined January-March 2019 Gallup poll data|
Partisanship has always been strongly related to presidential approval, but it has become even more so over time, and the divisions in job approval ratings by party have been larger than ever during the Trump administration. Extreme party polarization explains the stability in Trump's job approval ratings to date, which have mostly held around 40%, while never exceeding 45% or going lower than 35%.
These patterns suggest that the GOP is now defined by President Trump. But it is not clear if certain subgroups such as noncollege whites have become Republican because of their affinity for Trump, or if it comes from a longer-standing loyalty to the GOP and those subgroups have come to embrace Trump as the leader of the party. A future article will explore changes in party identification by subgroup between the Obama and Trump years to attempt to answer that question.
Explore President Trump's approval ratings and compare them with those of past presidents in the Gallup Presidential Job Approval Center.
Learn more about how the Gallup Poll Social Series works.