Red Oak Strategic

  • Home
  • About Us
  • Services
  • Blog
Contact Us
Patrick Stewart
Thursday, 17 November 2016 / Published in Analytics, Data, Political Analytics

How Our Poll Got It Right

This year, the Red Oak Strategic team decided to undertake a new challenge in the world of polling and analytics : conduct a bi-weekly, public, national survey that we would execute and release using the Google Surveys platform. Beginning in August and continuing through the final week of the 2016 election, our partnership with GCS led to the introduction of fresh and innovative polling methods which rendered great results.

Keeping in line with the Red Oak Strategic mindset of continually searching for the best, most innovative approach to research, our analysts tested this survey tool, which differs from traditional survey methods in that it targets voters through the channels of websites and digital media, rather than traditional phone calls. We then analyzed our results, released publicly in full, against results gathered using other tools and platforms.

Below is the summary of result comparisons, which you will see, proved the Red Oak Strategic tracker to be very accurate.

Red Oak Strategic vs. The Poll Averages

After each survey, we compared our results to those of two different national polling averages including FiveThirtyEight and RealClearPolitics. We found that as Election Day approached, our polling numbers were very aligned with that of the national average, especially in regards to the support for the two major candidates, Trump and Clinton.

Trump & Clinton Support vs. Averages

In our final survey going into Election Day, the Red Oak Strategic survey showed Hillary Clinton at 45.3%, just 0.1% above the national average of both RealClearPolitics and FiveThirtyEight.

Clinton Support v. Averages

We predicted Trump to earn 44.7% of the vote, which was a 1.9% difference from the RealClearPolitics estimation, and a 2.7% difference from FiveThirtyEight.

Trump Support v. Averages

Red Oak Strategic vs. the Popular Vote

Our survey, compared to the results of the popular national vote, proved to be very accurate — predicting a 0.6% Clinton advantage which in actual vote, is likely to end up at approximately 1% to 1.5%.

As of today, November 17, with 99% of the vote reporting, Clinton has edged out Trump in the popular vote by 1.0%, or 1,339,019 votes.

What the Polls Got Right in 2016

Many pundits and poll evaluators will study what happened with the 2016 election results and why many state/national polls did not pick up on late movement. It’s clear that pushing the world of survey research closer to the world of analytics gives better results. Our survey was designed with this rigor in mind and the accuracy bears this out.

We’re proud that our national survey and the work that we did with the Google Surveys team proved to be one of the most accurate national surveys released this cycle — likely to have predicted the national popular vote within 1%.


  • Tweet
Tagged under: 2016 Election, Data Science, Politics, Polling

What you can read next

Blockchain from a Data Science Perspective
Could Google and Xbox solve all of polling’s problems?
Data Processing with RegEx in Python 3

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Recent Posts

  • Tracking Coronavirus: Building Parameterized Reports to Analyze Changing Data Sources

    The pace of our modern world, and the impressiv...
  • Draw Rotatable 3D Charts in R Shiny with Highcharts and JQuery

    While it might be tempting to liven up a report...
  • Customizing Click Events: How to Capture and Store Data from JavaScript Objects in R Variables

    Interaction Design for Data Exploration Visuali...
  • Analytics at Scale: h2o, Apache Spark and R on AWS EMR

    At Red Oak Strategic, we utilize a number of ma...
  • Time Series Forecasting with Machine Learning: An example from Kaggle

    Introduction This post will demonstrate how to ...

Categories

  • Analytics
  • Blockchain
  • Code
  • Data
  • Data Science
  • Data Visualizations
  • Databases
  • Excel
  • Financial Analytics
  • Forecasting
  • JavaScript
  • Machine Learning
  • Political Analytics
  • Predictive Analytics
  • Python
  • R
  • R-Shiny
  • Tutorials
  • Uncategorized
TOP

Insert/edit link

Enter the destination URL

Or link to existing content

    No search term specified. Showing recent items. Search or use up and down arrow keys to select an item.