Aggregate statistics can sometimes mask important information.

The title of this article is a quote from Ben Bernanke--former Chair of the Federal Reserve. I felt the need to put my statistical brain at ease and gather a few thoughts about political polls, missed opportunities, and learning how to pause and welcome a little granularity and arched eyebrows.

But for the people, what does an addiction to polls, however cockeyed their results, signify other than an obsession with public opinion? It will come as no surprise to those familiar with Hannah Arendt’s way of thinking that she believed polls, like senseless calls for donations, signify either having lost or been swindled out of one’s own opinion.--Jerome Kohn in the introduction to Thinking Without a Banister by Hannah Arendt

In the hopes of revealing my selection bias I will tell you that I am in full agreement with Hannah Arendt regarding polls.

As a member of the National Press Club in Washington DC I am shocked that the story as reported has climbed into the microphones of talking heads across the political news desks.

As data people, we should be surprised that aggregated data was allowed to sell newspapers, grab attention, and not do the important service to the electorate.

First of all, polls are by nature not reliable. We have seen this time and time again. They are the noise, not the signal. Next time you watch the talking heads discuss the results make note of how clearly you are guided regarding the number of respondents, who was polled, and who actually sponsored the poll.

The Siena poll has a history of being trusted so I am blaming those that should know how important it is to disaggregate data. Especially if you are looking for something more reliable than having 3,662 people (of which large chunks didn’t even vote in 2020) sway the perception of the electability of the current president and the likelihood of a former president returning to power.

I am not interested in telling you how to vote. My interest is in showing you how to decipher misleading headlines. The video below is included to share the depth of the questions, the thoughtful answers, and how aggregating to the highest level silences voices and masks real opportunity for engagement.

https://www.nytimes.com/interactive/2023/11/07/us/elections/times-siena-battlegrounds-registered-voters.html

If you are going to tell stories with data this should be a cautionary tale. You need to disaggregate data to find the granularity in survey data.

There is definitely a pattern we see based on white, uneducated males of a certain age. There is another pattern seen in likelihood to vote and who actually voted in the last election.

Ask yourself a few other questions. How likely are you to answer a phone survey originating from an unknown number? Are we to believe that education status correlates with knowledge regarding the world at large?

Is there a chance that the wording of these questions might be interpreted differently by individuals?

Do you think the United States is on the right track, or is it headed in the wrong direction?

I encourage you to read the cross tab for a fascinating glimpse into issues troubling your fellow Americans.

I have a few additional questions.

• Can you tell me what the consumer price index measures?
• How well regarded is it as a measure of inflation (you know, among economists not the guy that makes your latte in the morning).
• What is the basket of goods and services? What is in that basket? Hint: over 94,000 items.
• Are you familiar with the Bureau of Labor Statistics (BLS) and its publicly reported monthly reports?

Do you think less than 4,000 fellow Americans that opted to answer their phones represent the 170 million registered voters closely enough that we should see this as a harbinger of anything?