Climatic Change

, Volume 145, Issue 3–4, pp 539–548 | Cite as

The spatial distribution of Republican and Democratic climate opinions at state and local scales

  • Matto Mildenberger
  • Jennifer R. Marlon
  • Peter D. Howe
  • Anthony Leiserowitz


Even as US partisan polarization shapes climate and energy attitudes, substantial heterogeneity in climate opinions still exists among both Republicans and Democrats. To date, our understanding of this partisan heterogeneity has been limited to analysis of national- or, less commonly, state-level opinion poll subsamples. However, the dynamics of political representation and issue commitments play out over more finely resolved state and local scales. Here we use previously validated multilevel regression and post-stratification (MRP) models (Howe et al., Nat Clim Chang 5(6):596–603 2015; Mildenberger et al., PLoS One 11(8):e0159774 2016) combined with a novel approach to measuring the distribution of party members to model, for the first time, the spatial distribution of partisan climate and energy opinions. We find substantial geographic variation in Republican climate opinions across states and congressional districts. While Democratic party members consistently think human-caused global warming is happening and support climate policy reforms, the intensity of their climate beliefs also varies spatially at state and local scales. These results have policy-relevant implications for the trajectory of US climate policy reforms.

The distribution of partisan climate opinions is consequential for climate risk mitigation prospects. Among Republicans, officials who once supported climate reforms have moderated their policy positions on the basis of perceived opposition by party voters. Others, such as former Rep. Bob Inglis of South Carolina, have lost their seats to primary challengers opposed to climate action. On the other side of the aisle, many Democrats have embraced climate policies; yet, a vocal party contingent routinely opposed climate reforms throughout the 1990s and 2000s (Mildenberger 2015).

Recent academic analyses have highlighted climate opinion polarization among the general US public (McCright and Dunlap 2011a, b; Guber 2013; Egan and Mullin 2017); yet, this stark political polarization between Republicans and Democrats in the aggregate masks substantial within-party heterogeneity on climate and energy issues. For example, eight high-profile Republicans, including James A. Baker III, George Shultz, and Henry M. Paulson Jr., recently announced support for a US carbon tax.1 A White House briefing notwithstanding, their intervention appears unlikely at this date to reshape Republican climate policy opposition. Still, their efforts caught some observers by surprise in exposing apparent pockets of carbon tax support even within the climate-skeptical Trump administration.2 Climate change may be one of the most politically polarized issue domains in contemporary US politics (c.f. Guber 2013); yet, recent analysis of national public opinion polls finds that nearly half of Trump voters (49%) in the 2016 general election believe that global warming is happening; majorities of Trump voters also support a diverse set of climate change mitigation policies (Leiserowitz et al. 2017).

What is the extent of this partisan climate opinion heterogeneity? Are Republican elites who push “conservative climate solutions” systematically out of step with Republican voters? Are there still substantial pockets of climate policy support within the party that may support a reorientation of the party’s climate policy goals? Are Democratic leaders championing climate reforms in step with their party’s base? And what are the implications of heterogeneity in Democratic climate beliefs on that party’s approach to climate policymaking?

These are ultimately empirical questions that require an accurate understanding of how partisan climate opinions vary across relevant political geographies. Yet, current analyses of partisan climate opinions are restricted to partisan subsamples of nationally representative polls or, in some cases, state-level polls that include a limited number of climate-related questions. The Republican party may be one of the most important political actors organized against climate policy action in the world; yet we still lack any detailed understanding of the distribution of Republican climate opinions at geographic scales relevant to party politics. For instance, we do not have a systematic way to measure the climate opinions of Republican party primary voters within Congressional districts. Likewise, we do not have a detailed understanding of the distribution of voices in favor of climate reforms within the Democratic party.

Understanding the distribution of state and local opinions about climate change and energy reforms has taken on new importance in the current policymaking context. Despite the Trump administration’s withdrawal from the Paris Climate Agreement in mid-2017, state and local governments continue to act as critical sites of climate and energy policymaking in the USA (Raymond 2016; Rabe 2004; Carley 2011). Accelerating levels of state and local policymaking highlight the need for public opinion and policy preference data at these subnational spatial scales.

This research letter helps fill this critical empirical gap. We combine previously validated climate-oriented multilevel regression and post-stratification (MRP) models (c.f. Howe et al. 2015; Mildenberger et al. 2016) with novel methods to construct party cross-tabs at local geographies in order to estimate the distribution of partisan opinions for every state and congressional district in the USA. Beyond its direct contributions to science and policymaking in the climate domain, this is also the first spatially resolved opinion model to systematically estimate the partisan distribution of opinions below the US state level. We are publicly releasing our model output as a companion to this research letter. The results should prove highly useful for other researchers and for publics and policymakers engaged in climate risk mitigation efforts.

1 Methods

To date, the primary barrier to models of partisan opinion distribution has been the absence of high-quality cross-tabs that measure the number of partisans with particular demographic attributes who reside in particular state and local geographies. Critically, neither the US Census nor the American Community Survey (ACS) collects political orientation data. We overcome this challenge by using custom-generated cross-tabs derived from analyses of state voter files that estimate the distribution of political partisans by race, gender, age, and geographic location. We combine these custom cross-tabs with unique large-n survey samples of Republican (n= 5696) and Democratic (n= 6383) climate and energy opinions collected by the Yale Program on Climate Change Communication and George Mason Center for Climate Change Communication between 2008 and 2016. These surveys were conducted by GfK Knowledge Networks using probability-based online samples. Survey questions are provided in supplementary material, along with question coding details.

This combined dataset allows us to estimate the spatial distribution of Republican and Democratic climate and energy opinions using a MRP approach. MRPs are increasingly used to model spatial distributions of public opinion (for more detailed treatments, see Park et al. 2006; Lax and Phillips 2009; Warshaw and Rodden 2012; Buttice and Highton 2013). An MRP analysis involves two steps. First, individual survey responses are modeled as a function of both individual-level demographics and geography-level covariates (the “multilevel regression model”). Second, using a fitted version of this regression model, population-weighted opinion estimates for demographic-geographic subtypes are aggregated based on the subtype population distribution within each geographic subunit (“post-stratification”). We extensively detail our approach to generating an MRP model from our partisan opinion datasets in an appendix to this letter. We also validate our MRP results for individuals with known party identification using an internal cross-validation technique and an external validation against an independent dataset, the Cooperative Congressional Election Study.

2 Results

Our estimates reveal substantial heterogeneity in partisan climate and energy opinions across the USA (Figs. 1 and 2). For example, much as we see with general population models of climate opinions (Pew Research Center 2016), a much larger fraction of Republicans think that global warming is happening than think that it is human-caused. Nonetheless, more than a quarter of Republicans in most states think that global warming is both happening and human-caused (Fig. 1a, b). Despite consistently high belief in climate change and support for climate reforms among Democrats, we also find that the intensity of Democratic beliefs varies spatially at the state level (Fig. 2a, b).
Fig. 1

Projected % of Republicans who hold selected climate opinions and policy preferences by state. States with plus symbol do not provide party registration data; party identification in these states is modeled

Fig. 2

Projected % of Democrats who hold selected climate opinions and policy preferences by state. States with plus symbol do not provide party registration data; party identification in these states is modeled

Since our survey includes a category for respondents to say they “don’t know” if global warming is happening, we can identify the number of states where there is a plurality rather than majority belief among a given partisan subgroup. To do this, we separately model the opinion that global warming is not happening. In the case of Republicans, we find that a plurality of Republicans in each of the 50 states think that global warming is happening. The largest difference exists in New York, where Republicans are 43 percentage points more likely to think that global warming is happening than not. Moreover, these effects are not limited strictly to traditional “blue states.” The third largest gap is Alaska, where Republicans are 32 percentage points more likely to believe that global warming is happening than to believe it is not. By contrast, a plurality of Republicans in every state are not convinced that global warming is mostly human-caused.

How do these climate opinions translate into policy preferences? Again we find heterogeneity in partisan opinions. For example, Republican support for Renewable Portfolio Standard (RPS) policies is mixed across states, with clear variation by region (majorities of Republicans in most Northeastern states support RPS standards, with more variable support across other regions) (Fig. 1c). By contrast, there is majority support among Republicans for regulating carbon dioxide as a pollutant in all states but Wyoming, though the size of this majority varies (Fig. 1d). Among Democrats, there is majority support in all states.

In Fig. 3, we include a sample plot of RPS policy support that offers more details on these state-level estimates with 95% confidence intervals for states that provide full party registration data (see Appendix for details). States with existing policies tend to also show higher levels of partisan policy support, though there are notable exceptions. For instance, Florida is the state with the largest fraction of Republican voters in support of an RPS policy that lacks such a policy at present. Figure 3 also offers companion data charting support for RPS policies among state-level Democrats. Perhaps not surprisingly, we find that a majority of Democrats in all states with party registration data support RPS standards.
Fig. 3

Projected % of registered Republicans (left) and Democrats (right) supporting state-level Renewable Portfolio Standards by state (states with party registration ID only). Asterisk indicates states with current Renewable or Alternate Energy Portfolio Standards

Partisan heterogeneity is equally dramatic when we consider variation in climate change opinions across congressional districts, although given our sample size, these estimates also have larger uncertainty intervals. In Fig. 4, we display absolute (left) and relative differences from the national average (right) in Republican and Democratic beliefs that global warming is happening for all congressional districts. The relative difference map highlights areas of the country where there are relatively more or fewer Republicans who believe that global warming is happening, compared to the national average for all Republicans, which is 53%.
Fig. 4

Projected % of registered Republicans (top row) and Democrats (bottom row) who believe global warming is happening by congressional district. Left panels map absolute estimates and right panels map relative differences from the national average for Republicans and Democrats, respectively. Congressional districts in states with plus symbol do not provide party registration data; party identification in these districts is modeled

Importantly, we find heterogeneity in congressional-district level opinions among both Democrats and Republicans. For climate attitude items, the standard deviation of district-level opinions is similar among both Democrats and Republicans. For these items, the primary difference is that the Democratic spread is centered around a much higher partisan average. By contrast, given the more even divisions among Republicans, opinion heterogeneity appears to offer more substantial political implications. However, for some policy items, we see evidence of substantially less heterogeneity among Democrats as compared to Republicans. For instance, the standard deviation across congressional districts in support for carbon pollution regulation is only 2.8 percentage points among Democratic partisans but double that amount (5.4 percentage points) among Republicans. Similarly, the standard deviation across congressional districts in support for RPS policies is 5.5 percentage points among Republicans and 3.6 percentage points among Democrats. This trend illustrates that, despite variation in Democratic intensity of support, even Democrats in the lower tail of the congressional district distribution still hold absolute levels of support compatible with policy reform.

Model map estimates illustrate substantive variation, even for questions where the average level of support among Republicans or Democrats is low. They also emphasize that patterns of geographic variation in Republican opinions do not simply follow existing political gradients (e.g., with blue state Republicans holding consistently higher climate beliefs than red-state Republicans). For instance, Florida Republicans are found to particularly support regulating carbon pollution compared to the average national party member (see Fig. 5). Republicans in southern Florida districts are also much more supportive of regulating carbon dioxide as a pollutant, even compared to more traditionally liberal parts of the USA. By contrast, there is more homogeneity among Democrats in support for this policy.
Fig. 5

Projected % of Republicans (top) and Democrats (bottom) who support regulating CO2 as a pollutant (expressed as difference from the national average), by congressional district

The distribution of Republican climate and energy opinions has particular salience in the context of primary elections. In Table 1, we subset Congressional districts according to the 241 congressional districts that are currently held by Republican legislators in the 115th Congress and the 194 districts currently held by Democrats. We then count the number of districts in which a plurality of registered Republicans support a climate-related statement. We find substantial heterogeneity in the opinions of Republican registered voters. While a plurality of Republicans in nearly every Republican-held congressional district think that global warming is happening, there are few districts where a plurality of Republican voters think that it is human-caused or that it will harm the USA. By contrast, Republican voters still support many policies. We find plurality support for funding research on renewables in every single Republican-held district. We also find strong Republican support for regulating carbon pollution. We find plurality support in 111 districts for RPS standards.
Table 1

Number of Republican-held and Democratic-held congressional districts with plurality agreement among Republicans for select climate and energy opinions, based on total 241 districts held by Republicans and 194 districts held by Democrats in the 115th Congress as of September, 2017


# Republican-held districts

# Democrat-held districts

GW happening



GW human-caused



GW worried



GW harm US



Regulate carbon pollution



Support RPS



Fund renewables



Even where this plurality does not exist, the opinion distribution of registered Republicans suggests more variation than one might expect. For example, there is less than a 5 percentage point gap between Republicans who support and oppose RPS policies in an additional 50 Republican-held congressional districts. Likewise, in an additional 11 Republican-held congressional districts, the gap between Republican voters who think that global warming is human-caused and not human-caused is less than 5 percentage points.

By contrast, Republicans living in Democratic-held congressional districts tend to be somewhat more concerned with climate change. For example, Republican voters in all congressional districts currently held by Democrats show plurality belief that global warming will harm the USA. Likewise, a plurality of Republicans in 152 Democratic-held districts support RPS policies.

3 Conclusion

Overall our analysis finds that partisan climate opinions are characterized by a policy-relevant degree of heterogeneity. Both the belief that global warming is happening as well as support for particular climate mitigation policies varies widely across both states and congressional districts. There remain some states and congressional districts where a majority of registered Republicans think that human-caused global warming is happening or support climate reforms; for instance, a plurality of Republicans in every US state think that global warming is happening (although beliefs that it is human-caused are considerably lower). Republicans also hold heterogeneous climate policy preferences, including widespread support for renewable energy funding and carbon pollution regulation. We find similar levels of heterogeneity among Democrats for many climate attitudes, though substantially lower levels of heterogeneity for some policy support questions. However, since the partisan average is substantially higher among Democrats, this variation may have less substantial implications for climate policymaking. Democrats in every district and every state think that global warming is happening and support climate and energy policies.

As the authors collect additional data over the coming years, the precision of these congressional district estimates will increase accordingly. At present, we view the data as being appropriate for aggregated analysis of responsiveness (as in the data provided in Table 1), or for research efforts, for example, to regress roll call votes on public opinion. However, some caution is appropriate, given estimate uncertainty, in making strong descriptive claims about particular congressional districts.

Broadly, this novel geographic dimension of partisan climate opinions in the USA should greatly inform decisionmakers, educators, and communicators working at subnational geographic scales. While subsets of the Republican voting public do not support climate policies and hold views consistent with party elite, Republican climate and energy opinions are more varied than might be presumed from political discourse. Similarly, the results emphasize consistent support among Democrats for climate and energy policies, despite variation in belief intensity. More broadly, these results find substantial heterogeneity in partisan climate opinions and policy preferences and demonstrate the ability of MRP approaches to document these differences.


  1. 1.

    John Schwartz. 7 February 2017. “A ‘conservative climate solution’: Republican group calls for carbon tax. New York Times. Available online at

  2. 2.

    Josh Dawsey, Annie Karni and Andrew Restuccia. 21 March 2017. “Carbon tax debate exposed rift among Trump’s aides.” Politico. Available online at



Thanks to Baobao Zhang, Chris Warshaw, Lyle Scruggs, and two anonymous reviewers for comments on an earlier draft of this paper.

Supplementary material

10584_2017_2103_MOESM1_ESM.pdf (46.1 mb)
(PDF 46.1 MB)


  1. Buttice MK, Highton B (2013) How does multilevel regression and poststratification perform with conventional national surveys?. Polit Anal 21(4):449–467CrossRefGoogle Scholar
  2. Carley S (2011) The era of state energy policy innovation: a review of policy instruments. Rev Policy Res 28(3):265–294CrossRefGoogle Scholar
  3. Egan PJ, Mullin M (2017) Climate change: US public opinion. Annu Rev Polit Sci 20:209–227CrossRefGoogle Scholar
  4. Guber DL (2013) A cooling climate for change? party polarization and the politics of global warming. Am Behav Sci 57(1):93–115CrossRefGoogle Scholar
  5. Howe P, Mildenberger M, Marlon J, Leiserowitz A (2015) Geographic variation in opinions on climate change at state and local scales in the USA. Nat Clim Chang 5(6):596–603CrossRefGoogle Scholar
  6. Lax JR, Phillips JH (2009) How should we estimate public opinion in the states?. Am J Polit Sci 53(1):107–121CrossRefGoogle Scholar
  7. Leiserowitz T, Maibach E, ReserRenouf C, Rosenthal S, Cutler M (2017) Trump voters and global warming. Yale University and George Mason Project on Climate Change CommunicationGoogle Scholar
  8. McCright AM, Dunlap RE (2011a) Cool dudes: the denial of climate change among conservative white males in the United States. Glob Environ Chang 21(4):1163–1172Google Scholar
  9. McCright AM, Dunlap RE (2011b) The politicization of climate change and polarization in the American public’s views of global warming, 2001–2010. Sociol Q 52 (2):155–194Google Scholar
  10. Mildenberger M (2015) Fiddling while the world burns: the logic of double representation in comparative climate policymaking. PhD thesis, Yale University, New HavenGoogle Scholar
  11. Mildenberger M, Howe P, Lachapelle E, Stokes L, Marlon J, Gravelle T (2016) The distribution of climate change public opinion in Canada. PLoS One 11 (8):e0159774CrossRefGoogle Scholar
  12. Park DK, Gelman A, Bafumi J (2006) State-level opinions from national surveys: poststratification using multilevel logistic regression. In: Cohen JE (ed) Public opinion in state politics. Stanford University Press. 209–228Google Scholar
  13. Pew Research Center (2016) The politics of climateGoogle Scholar
  14. Rabe BG (2004) Statehouse and greenhouse: the emerging politics of American climate change policy. Brookings Institution Press, Washington, D.C.Google Scholar
  15. Raymond L (2016) Reclaiming the atmospheric commons: the regional greenhouse gas initiative and a new model of emissions trading. MIT Press, CambridgeGoogle Scholar
  16. Warshaw C, Rodden J (2012) How should we measure district-level public opinion on individual issues?. J Polit 74(01):203–219CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2017

Authors and Affiliations

  1. 1.Department of Political ScienceUniversity of California Santa BarbaraSanta BarbaraUSA
  2. 2.School of Forestry and Environmental StudiesYale UniversityNew HavenUSA
  3. 3.Department of Environment and Society, Quinney College of Natural ResourcesUtah State UniversityLoganUSA

Personalised recommendations