Election 2016 US Debate Viewership points to Record Voter Turnout
One of the main determining factors, and one of the least covered variables, for the 2016 US Presidential Election is voter turnout. In other posts, we discussed how search activity and interest levels (measured in polls) point to voter turnout that could approach or break the post 19th Amendment record. In this post, we look at how the number of viewers of the US Presidential Debates provides a leading indicator to voter turnout.
Below is the estimated number of viewers of the debates in millions.
Table 1: Estimated Number of Viewers of US Presidential Debates, in millions
| 2016 | 2012 | 2008 | 2004 | 2000 | 1996 | 1992 | 1988 | 1984 | 1980 | |
| First Debate | 84.0 | 67.2 | 52.4 | 62.4 | 46.6 | 46.1 | 62.4 | 65.1 | 65.1 | |
| Second Debate | 66.6 | 65.6 | 63.2 | 46.7 | 37.5 | 36.3 | 69.9 | 67.3 | 67.3 | 80.6 |
| Third Debate | 71.6 | 59.2 | 56.5 | 51.1 | 37.7 | 66.9 | ||||
| VP Debate | 37.0 | 51.4 | 69.9 | 43.5 | 28.5 | 26.6 | 51.2 | 46.9 | 56.7 | |
| Max | 84.0 | 67.2 | 69.9 | 62.4 | 46.6 | 46.1 | 69.9 | 67.3 | 67.3 | 80.6 |
| Streaming | 40.0 | |||||||||
| Max + Streaming | 124.0 | 67.2 | 69.9 | 62.4 | 46.6 | 46.1 | 69.9 | 67.3 | 67.3 | 80.6 |
Source: Wikipedia, Nielson, NPR, Commission for Presidential Debates
The 2016 debates were extremely popular and attracted record number of viewers. The most efficient way of comparing the data over time is likely to look at the maximum level as going further back the data becomes less uniform. For instance, some years had only two debates and in others like in 1980 there is no viewership data for the first debate. Using the maximum shows that 2016 posted a record for the number of viewers. If including the estimated number of viewers watching the debate on-line using a streaming service, then the previous record was not broken it was smashed.
In theory, higher viewership should result in higher voter turnout. In general, interest and enthusiasm levels are good indicators for forecasting turnout – and we can more or less assume that viewership of debates is good indicator of interest.
The following chart compares historical voter turnout in US Presidential Elections and US Presidential Debate viewership.
Chart 1: Comparing Estimated Number of Viewers of US Presidential Debates and Voter Turnout in US Presidential Elections, 1980 – 2016
Source: Wikipedia, Nielson, NPR, Commission for Presidential Debates
The relationship is not perfect but strong none-the-less. In reality, voter turnout is an aggregate ratio of many different age cohorts and demographic groups with each having tendencies towards higher or lower turnout levels, so in order to do a complete study such influences would need to be taken into consideration. For instance, Baby Boomers in 1980 were still in an age group with a lower voter turnout ratio which partially explains high viewership with mediocre turnout. Regardless, even in its aggregate form, voter turnout tends to move more or less in-line with viewership of debates.
Now, by including the streaming viewers in 2016, we can get a better idea of interest level in the debates.
Chart 2: Comparing Estimated Number of Viewers of US Presidential Debates, including Streaming Viewers, and Voter Turnout in US Presidential Elections, 1980 – 2016
Source: Wikipedia, Nielson, NPR, Commission for Presidential Debates
After including the Streaming estimate, the number of viewers of the debates just simply demolished every other election in the sample.
If you can make the relatively easy mental connections between strong viewership of debates => high interest in the election => high voter turnout, then you can more or less assume that voter turnout during 2016 should approach or break previous records.
If this was an isolated data point, then maybe we could leave it alone. But, many other variables confirm the same conclusion. As discussed in other posts, primary turnout, on-line activity, voter enthusiasm, and mega rallies all indicate that the 2016 election is exceptional in terms of interest in the election.
Republicans should benefit disproportionately from the spike higher in voter turnout in the general election. Again, many of these topics are covered in other posts but to finish the analysis let’s go over some of them briefly. Republican primary turnout hit records, Democrat’s did not. Trump is receiving significantly higher interest as measured in social media, Wikipedia page views, and Google search. Multiple polls and surveys show that 2016 interest and enthusiasm indicators are at record breaking levels as compared to previous elections. And, average attendance at Trump rallies is often multiples higher than that for Clinton. Together, these confirm the assumptions that voter turnout will be substantial this election and that it will tend to benefit Republicans.
The level of turnout could smash anything we have seen in the last generation in the US, making 2008 look tame by comparison. The post 19th Amendment voter turnout record is approximately 63%, which judging from the previous chart could be tested or broken this year.
Making forecasts for voter turnout is somewhat difficult. As discussed in other posts, it is amazing that this variable has not received more attention from pollsters and researchers as so many variables point to potential record-setting turnout. Additionally, as turnout was so influential in Obama’s victories, it would be normal to assume that this variable would be on the top-of-mind of researchers going into 2016. However, as it stands currently analysts are not fully incorporating how turnout could significantly impact the election and pollsters have not focused on this important aspect either. Without deeper data, we can mostly make informed estimates based on historical correlations of key variables and on demographic cohort projections.
Pulling data from the other aforementioned variables, it looks like turnout will test or exceed the previous high. Such an ‘unexpected’ jump in turnout levels will likely render almost all of the current polls inaccurate if the demographic mix of voters changes substantially which is the base case.
Summarizing, the US Presidential Debates hit record viewership levels during 2016. Intuitively and historically, high viewership normally equates to stronger voter turnout. When incorporating streaming estimates, which makes sense as the way in which people watch TV has changed, viewership jumps substantially higher implying a very large spike in turnout. Republicans should benefit disproportionately from higher turnout as the main indicators point to lopsided interest in their favor. Current polls do not appear to incorporate a significant spike in turnout which would render them extremely inaccurate in a high turnout scenario.
As an aside, one of the more humorous things to watch is the lack of ability of political analysts to connect the dots. They recognize that viewership of debates broke previous records, that Republican primary vote volume broke previous records, that polls show record-breaking ‘interest in the current election’ responses, and that Trump (and Sanders during the primaries) holds extremely large rallies. Additionally, alternative data such as Google search and social media show very high interest levels, with Trump apparently benefiting. All of these bottom-up variables show that interest is extremely high this election cycle and that such a spike in interest will most likely benefit Trump. However, they decline to incorporate these points into their models and appear to be modeling turnout and ‘likely voters’ based on historic voting patterns which makes sense in a ‘normal election’ but certainly not in 2016.