Orest Biloskursky, the head of the Kyiv24 Applied Research Think Tank, offers an exclusive analysis of the U.S. presidential election campaign for kyiv24.news. With over 20 years of experience in television research, Biloskursky has been crucial in developing modern TV measurement systems in Ukraine and is widely recognized as an expert in TV ratings. His previous positions include Director of Applied Research at StarLightMedia and head of the research department at the Ukrainian Public Broadcaster. Also, he is the founder of the independent research agency Boy Research.
A detailed analysis of U.S. polling data indicates that the outcome will hinge on just two states: Pennsylvania—home to the Scranton setting of The Office—and Wisconsin, a state that may be less familiar to Ukrainians.
Before you read further, take a moment to look at the expected results map. By the end of this text, you’ll gain a clearer understanding of what could unfold on Wednesday.
Currently, there are three states where polling data does not indicate a clear winner due to the narrow margins between candidates. Nevada, known for its Las Vegas and Reno counties, has six electoral votes, but these are not decisive in this scenario. The real competition lies in Pennsylvania and Wisconsin.
While a win in Wisconsin alone will not guarantee a Democratic victory, it is sufficient for Republicans. On the other hand, a victory in Pennsylvania secures the necessary electoral votes for the Republicans. The Democrats, however, need to win both states, which, upon further analysis, seems unlikely.
So, why is that the case? Let’s take a closer look at the polling data and the results from the 2020 election. Here’s how the comparison looked for Trump, showing the monthly differences between average polling data and the final election results in Pennsylvania:
Throughout the year, polls consistently underestimated Trump’s support by an average of 4.4 percentage points, indicating he performed 4.4 points better than the annual polling average suggested. A more precise figure is that he exceeded the fall 2020 polls by 3.8 points. What does this imply for our predictions? Although pollsters have made slight improvements and their methods are more effective in the U.S. now, the systemic discrepancy has not been fully resolved. Therefore, we can anticipate that actual results in fall 2024 will likely be 1.9 percentage points better than what the polls indicate.
Looking at the same analysis for Biden, we see:
On average, the polls underestimated him by 1.2 points. However, it’s noteworthy that the fall polls overestimated his results by 0.9 points. This suggests that in 2024, we can expect a 0.5-point overestimation for the Democratic candidate in Pennsylvania, which is an important consideration for us.
Now, moving to Wisconsin. Does anyone know the state’s largest city (Milwaukee) or capital (Madison)?
On average, polling errors in this context were 5 percentage points, reaching as high as 5.4 percentage points during the fall months. In October, the error was 5.1 points. What does this indicate? We can reasonably anticipate an underestimation of 2.6 points in 2024.
What about the Democrats?
What can we learn from Pennsylvania?
Among the most accurate polling companies, four stand out: Susquehanna Polling & Research, Trafalgar Group, Civiqs, and AtlasIntel. The latter is particularly noteworthy for its close alignment with actual results, although it has only conducted one survey. To maintain objectivity, it’s best to examine the data from each company individually rather than averaging them. I won’t go into these details now, but if you’re interested, I’d be happy to share the information, though there isn’t much available. This is something lacking in Wisconsin.
Here’s an overview of how these companies performed in that context:
It’s challenging to determine who performed best, but we can focus on three companies that were the most accurate in predicting Trump’s support: Civiqs, Ipsos, and YouGov. Interestingly, Civiqs appears to understand both states better than the others.
What can we expect in 2024 based on this data?
The monthly polling data for 2024 shows the following:
A change in the candidate positively affected the situation before fall, allowing us to comprehend how Pennsylvania actually votes based on the initial data.
We observe a clear Republican victory in Pennsylvania, projected at 49.45% to 47.94%—a margin significant enough to avoid any recount, even though initial analysis suggested a much closer result.
Wisconsin is the second state with a slightly smaller margin, but the expected results still indicate a confident victory for Trump.
Here are our final projected results:
In conclusion, a detailed analysis of polling data and historical trends suggests that the outcome of the upcoming U.S. election will heavily depend on results from two key states: Pennsylvania and Wisconsin. Although initial data indicates a tight race, considering potential systemic polling errors and adjustments for 2024, it appears that Republicans may have an advantage. This could result in a victory with a sufficient margin to avoid any recounts.
Orest Biloskursky
