Climate Change Deniers vs Political Affiliation

Tom Courtney

5 min readNov 8, 2022

Tom Courtney

2022–11–01

ASK

What is the main purpose of this Case Study?

Is political affiliation correlated with Belief/Non-Belief in Climate Change?

Why is this important?

Having a deeper understanding may enable more successful efforts to convert non-believers.

PREPARE

Where will Climate Denier Data come from?

The data underlying the maps come from a large national survey dataset ( >28,000 respondents) collected between 2008 through 2021 as part of the Climate Change in the American Mind project led by the Yale Program on Climate Change Communication and the George Mason University Center for Climate Change Communication. (https://climatecommunication.yale.edu/visualizations-data/ycom-us/)

Where will Political Affiliation Data come from?

Data comes from the 2020 Presidential voting data found here. (https://dataverse.harvard.edu/file.xhtml?fileId=6104822&version=10.0)

Note on granularity: Both data sets contain data at the County level.

County data should provide better granularity than Statewide data. Some states have significant differences depending on proximity to major cities. Therefore, County-wise data will be used for all analysis.

PROCESS

Several cleaning steps were required to merge the two data sources into one data frame.

Denier Data Set

County Identification was standardized to fips. Federal Information Processing Standard.
fips data coerced to character data type.
Data filtered for only County data. (Omit State level data)
Data filtered for the field “happening”. This field is proportion of county that believe Climate Change in happening.

denier_data <-read_csv("YCOM6.0_2021_Data_.csv")

#Lets use fips for all County identifiers. 
denier_data_counties <- denier_data %>% 
    rename(
    fips=GEOID,
    ) %>%
  filter(GeoType=="County")%>%
  select(fips,GeoName , happening) 

denier_data_counties$fips<-as.character(denier_data_counties$fips)

Visualize - Let’s look at Climate Denier data vs County

plot_usmap(data=denier_data_counties,values="happening")+
    scale_fill_continuous(low="red",high="green",
                         name = "% believers Climate Change", 
                        limits = c(50,85))+
  ggtitle("Climate Change Belief by County")+
  theme(legend.position="top")

Some key Observations from the visualization

Some states have very high variation within the state. For example, look at Florida or Texas.
Central and midwest deny much more than California
South West Texas has counties with believers. Perhaps due to high Hispanic communities?
Most counties are > 50%, indicating practically all counties have majority of believers.

Let’s now import and clean Political Affiliation Data set.

Political Affiliation Data Set

In Excel, I filtered data only from the 2020 Presidential Election
In Excel, proportion of voters voting for Trump calculated and saved.
In Excel, proporation of Voters voting from Biden calculated and saved.
County Identification was standardized to fips. Federal Information Processing Standard
fips data coerced to character data type.

Visualize — Let’s look at Political affiliation vs County

plot_usmap(data=political_data_counties,values="DEMOCRAT")+
    scale_fill_continuous(low="white",high="blue",
                         name = "% democrats", 
                        limits = c(0,100))+
  ggtitle("Political Affiliation Map - Democrats")+
  theme(legend.position="top")

plot_usmap(data=political_data_counties,values="REPUBLICAN")+
    scale_fill_continuous(low="white",high="red",
                         name = "% republicans", 
                        limits = c(0,100))+
  ggtitle("Political Affiliation Map - Republicans")+
  theme(legend.position="top")

Some key Observations from the Democratic visualization

Highest democratic counties are those in urban cities.
Note high democratic counties in southwest texas. Hispanic region.
California overall quite blue.
Data not available for all Alaska Counties

Some key Observations from the Republican visualization

Note concentration in central plains
Essentially an inverse of the democratic map.
Data not available for all Alaska counties

Now it’s time to merge the Denier data with the Political data

First, need to reshape political data from wide to long.
Rename variables to improve interpretation
Merge the two dataframes using fips as the KEY.

# Need to reshape political data from wide to long

df<-melt(political_data_counties, id="fips")

df<-df %>% 
  rename(party=variable) %>%
rename(political_value=value) 

#Now, let's merge the denier data
df<-merge(denier_data_counties,df,by="fips")

Visualize — Let’s see political affiliation of Climate Deniers.

ggplot(data=df,aes(political_value,happening,color=party))+
  geom_point()+
  scale_color_manual(values=c("blue","red"))+
  labs(title="Climate Change Believers vs Political Affiliation - by County")+
  xlab("Political Leaning %") +
  ylab("Believe Climate Change is happening %" )+
  geom_smooth()

Climate Change Believers vs Political Affiliation — by County

ANALYZE

Key takeaways from Scatter plot

WOW! Political Affiliation is VERY STRONGLY CORRELATED WITH CLIMATE BELIEFS
Look at lower right
Red strong Republican Counties: Only 50%-60% believe Climate Change is happening!
Look at Upper Right Democratic Counties:
Blue strong Democratic Counties: 70%-80% believe Climate Change is happening!
Why are there more red dots than blue dots for political leaning % greater than 50%
Since democrats are located in highly populated urban areas, there are fewer counties.
Since republicans are located in rural areas, there are more republican counties.

SUMMARY

I really didn’t expect the correlation to be that strong! What’s going on here? Does political affiliation influence people’s ability to interpret scientific information? It would seem so. Why are so many republicans skeptical of scientific reasoning? Perhaps there’s another variable that is cross-correlated. Perhaps education is the more dominant variable. See Figure 1 at end of document. From this figure one can see Democrats have higher percentage of college degrees than Republicans. It’s also interesting to not how that trend is getting even stronger over time. It’s also interesting to note that prior to 2009, Republicans were slightly more educated than Democrats.

ACT

What actions can be taken convert more people to believe in Climate Change?

In order to save our planet, we need Innovations.

Technical Innovations are necessary! Taxing and setting targets won’t solve the problem.
Innovations are expensive and do not always provide a return on investment.
Therefore, government funding of Innovations is necessary.
Private sector can not absorb those risks.
Politicians determine funding to Innovations.
Politicians follow the BELIEFS OF THEIR CONSTITUENTS.
Therefore, the BELIEFS and WILL of the people will ultimately determine whether we solve this problem.

Actions

Deeper education, especially in Republican counties.
How?
Make the education personal.
Have people affected by Climate Change to come and tell their stories.
Talking theory and science is a waste of time to the republican sector. Don’t even bother.
Make more popular movies (not documentaries!) to give people a sense of what the future world will look like.
How can you make a difference?
Tell your friends this…

Both our children and our grandchildren are counting on us. Don’t let them down.

Final Note: This Case Study was also prepared as a requirement for the Google Data Analytics Capstone Project to complete the Coursera “Google Data Analytics Professional Certificate” program.

https://www.coursera.org/professional-certificates/google-data-analytics