Covid-19, Blame, and Trump Support

October 24, 2020

Introduction

Covid-19 is obviously at the forfront of many voters minds today. As proven in the final presidential debate, a lot of ambiguity still surrounds the question of different vaccines and the potential end of the pandemic. Also discussed in the debate was President Trump’s success or failure with handling the virus, depending on your perspective. This question of “did Trump do a good job with Covid?” draws attention to a larger point of interest: do voters blame Trump for the continued pandemic?

In this blog post I will investigate this question through looking at the relationship of average poll support for Trump and different measures of Covid-19 progression such as death rate and cases per capita.

Background: how did Covid-19 affect different states?

For some brief context of the historical and current rates of Covid-19 in each state, see the visual below that shows Covid-19 cases per 100,000 people in each state.

ani

Discussion

It is apparent that cases per capita is continuing to grow as we near the election

Southern states seem to be hit relatively hard on a per capita basis

Battleground states such as Arizona, Georgia, and Florida seem to have some of the highest case rates for their populations

Is there a relationship between polls and worsening of Covid-19?

If voters do blame Trump for the pandemic, it’s possible that it would be reflected in the polls. I would expect that if this were the case, a worsening of Covid-19 conditions would lead to a decrease in support for Trump. The graffic below shows the relationship between shifts in polls and cases per 100,000 across states. As you can see, there doesn’t appear to be a relationship between the two variables, and running a regression shows no statistically significant relationship between the two variables as well.

Discussion

This figure suggests that voters do not punish Trump simply because of an increase in Covid-19 cases

It makes sense that the pandemic would not shift the polls in strongly blue or strongly red states, so this could be masking the relationship of poll changes in battleground states

However, running a regression that accounts for battleground status does not reveal for any statistically significant relationships either

Covid in Battleground States

It is interesting to look at analysis of Covid-19 metrics and poll support for Trump in individual battleground states.

Cases per Capita

First, we will look at the same metric as above– cases per 100,000 in each battleground state. The graphics below show the evolution of that metric in each state and the relationship between Trump Support in each state and that number.

Discussion

It is evident that the rate at which Covid-19 is growing hasn’t slowed in some states

In some states there is a positive correlation between cases and Trump support

In some states there is a negative correlaiton between cases and Trump support

There doesn’t appear to be a strong linear relationship across states and even in individual states

This implies that voters have not directly punished Trump for growth in cases

Thus, it would be difficult to estimate how Covid-19 will effect the election from this metric

Death Rate

Death rate is another way to look at how damaging Covid-19 is in a certain state. Perhaps it is the case that voters will respond more to death than just illness and that relationship will be reflected in the polls. The graphic below shows the change in death rate over time in battleground states as well as the relationship between poll support for Trump and death rate.

Discussion

There does not seem to be a consistent linear relationship between death rate and poll support for Trump

As with looking at cases per capita, some states show a negative correlation and other states show a positive correlation

This inconsistency makes it difficult to predict how Covid-19 will affect the election across states or even within states

Conclusion

I did not find any consistent relationship between Covid-19 rates and support for Trump. While this could suggest that voters don’t punish Trump simply because Covid-19 exists, I believe that voters will still punish Trump indirectly for consequences of Covid-19 in areas such as the economy. Alternatively, it is possible that voters will reward Trump for pandemic aid or stimulus checks, as discussed in a previous blog post. However, without a clear relationship of how shifts in Covid-19 affect support for Trump, I think it is best to leave pure Covid-19 data out of my predictive model.