Polling Is Not an Exact Science

November 24, 2020

After the 2016 election, many asked, “What is wrong with polling?” All the polls showed that nationally, Hillary Clinton had a commanding lead over Donald Trump. However, when the results of the Electoral College were tabulated, Donald Trump was elected the next president. Many thought this was a one-time aberration. However, in 2020, the polling predicted a landslide by Joe Biden. While Biden did receive over five million more votes than Trump, many states appeared to be razor thin. Most notably, Wisconsin was much closer than any poll showed.

What can we as broadcasters learn from this? Polling is not an exact science.

Polling, like any survey technique, is extremely difficult to conduct. Numerous missteps can lead to misleading data. However, every day broadcasters rely on a survey – Nielsen. As we all know, each of our careers and our financial well-being are dependent on what Nielsen reports.

The good news is that broadcasters are not dependent on any one single survey. Most major markets get snapshots 12 to 13 times per year. This is helpful, but problems can still occur.

So what do broadcasters do?

First, they must understand and track Nielsen’s performance. On a regular basis, Research Director, Inc. tracks Nielsen’s performance. This goes beyond just overall in-tab, to track in-tab and proportionality by:

• Demographic group
• Gender
• Ethnicity
• Geography

We do not expect Nielsen to be perfect, but broadcasters need to know where underperforming sample groups may affect the released data. With this knowledge, smart programmers can avoid making programming decisions using data that is not a true reflection of listeners’ habits.

As they work with the data, broadcasters need to pay attention to how they are segmenting the data. The more they reduce the sample size, the more unexplained fluctuations might occur. While Research Director, Inc. has various tools, like our Hot Zip and Exact Age Reports, that segment audiences, we are constantly advising our clients on how to properly use this data.

One cannot just say, “This Nielsen data is wrong, or does not make sense.” As we always say, facts matter. If you feel that the Nielsen (or any survey) data is not a true reflection of your stations or market, look into how the survey was performed.

Understanding the quality of an individual survey can make you smarter and help you make better programming decisions.

To discuss your market’s sampling, please reach out to us here.

-Charlie Sislen, Partner

• Tweet