Polls Mislead Most Of The Time. Here’s Why.

Polls mislead most of the time, even though we design/use them to gain more clarity. If that sounds surprising or confusing to you, do not worry. By the end of this essay, you will have understood logically why this is the case.

The story begins with me fiddling with polling algorithms and polling codes for an application that I had been working on. While creating the reporting section, I kept running into a bizarre scenario where my own conclusions made no sense to me.

Being a fan of the scientific method, I looked into some literature on this issue. Lo and behold, it is a well-known fallacy in mathematical circles. It’s just that I did not know back then.

In this essay, I will be explaining why polls mislead us using a simple hypothetical example. To this end, we will only need basic mathematics and logical reasoning. Let us begin.

This essay is supported by Generatebg

The Company Poll — Expenses or Debt?

Imagine that you run a modest tech company. Your business has been suffering recently and you are now struggling to keep the boat afloat. You are the kind of leader who listens to your people before making decisions. Naturally, you decide to engage your employees via a poll.

In your poll, you offer your employees the following options to counter the situation:

Option A: Save on company expenses for the next few months.

Option B: Take on debt, but thereby risk the company’s health even further.

At the end of the poll, 76% of your employees are in favour of cutting down on company expenses, whereas only 24% are in favour of taking external debt. Based on this, you decide to cut down your company’s expenses. But that brings you to the next question:

“Where can you cut company expenditure?”

Being the good leader that you are, you once again turn to your employees for suggestions and directions.

---

Where Did the Majority Go?! — Polls Mislead Most of the Time

To gather consensus on where to cut expenses, you set up yet another poll. This time, the poll options are as follows:

Option A: Stop the free meals that employees currently enjoy.

Option B: Exchange the current fancy company smartphone for a simpler model.

Option C: Stop the transportation allowances that employees currently enjoy.

Your polls are blind polls. That is, everyone gets to know the results only after the polls complete. Once the results arrive, you are quite confused:

Expressed otherwise, 91% of your employees do not wish to give up free food, 83% do not wish to give up their fancy company smartphones, and 77% dislike the idea of cutting down transport allowances. Just what is going on here?

Clear the Forest but Don’t Cut the Trees!

What you have on your hands is a contradictory situation. The majority of your company wishes to cut down on expenses. But when asked which part to cut down on, the majority is against each provided option.

Could it be that you missed a magical option where there might be consensus? Well, I would highly doubt that; we will see why in a bit.

But first, how could it be that your employees treated the whole (cutting down expenses) rationally and the parts (where to cut expenses) irrationally? To quote a Harris poll:

“Many people seem to want to cut down the forest but to keep the trees.”

Unfortunate as the situation might be, you HAVE to make a decision, for your company’s life hangs in the balance. Before we finalise your decision, let us understand this problematic phenomenon a little deeper.

Polls Mislead Most of the Time. Here’s Why.

You see, polls work very well when voters are deciding between two options. This is the reason why your first poll gave you an efficient result — to cut down company expenses rather than take on external debt.

However, as soon as the number of options increases to more than two, polls start losing effectiveness. Your employees did not suddenly become irrational.

They just wish to keep benefits that they find useful/necessary — any rational human being would do the same. It is what you are looking for that is the real problem — a consensus.

You cannot throw a bunch of options to a crowd and “expect” the people to converge on one option. That is not how human beings work (historically speaking).

In rare cases, when there is overwhelming “human” agreement, it might work. But most of the time, the result will be akin to your poll result — no clear majority.

In practice, analytics firms, corporate managers, politicians, etc., take advantage of this phenomenon to push their own agenda. Think about it; You can interpret your own poll result in at least three different ways.

Depending upon the idea you like the most, you could formulate your decision-statement accordingly. In fact, that is exactly what we are going to do!

---

The Final Decision

Your employees are clearly divided, in that, there is no clear majority support for any option. If you wish to handle the situation cleverly, you should turn the situation on its head. In other words, you should take advantage of the lack of a consensus.

How do you do that? Well, here is one possibility:

“Dear employees,

We have decided not to touch the free meal system we all love and cherish in our company. We give each employee the choice between switching their expensive smartphone for a less expensive one or give up their travel expenses.

Please communicate your choice to your manager, and we will take care of the rest. This situation will only last for a few months. So, do not worry. When we have crossed the tide, we will restore all functions just like before.

Yours truly,

CEO.”

What you have done here is pick two options simultaneously and request each employee to choose between the two. Let’s be honest. Based on the poll results, we can safely say that no matter which option you chose, the majority of your company would have been dissatisfied.

By choosing two options and introducing a choice element, you are maximising divided opinions even further. Although such a strategy is often used to manipulate people, in your case, you use it for a fair cause: to keep your company afloat. So, you can be proud of your decision.

Final Thoughts

Whenever you read reports or conclusions based on poll results, be sceptical. Read into the data, and draw your own conclusions. It is most likely that news agencies, analytics firms, etc., have their own agenda that they try to push with their reports/conclusions.

Furthermore, most polls have implicit assumptions and unavoidable biases baked into their design. It helps to be aware of why the poll was designed and what exactly it is that you are looking for.

Do not fall into the trap of blindly trusting false/contradictory conclusions based on polls. Because polls mislead most of the time!

---

Reference and credit: Jordon Ellenberg.

If you’d like to get notified when interesting content gets published here, consider subscribing.

Further reading that might interest you:

• How To Really Benefit From Curves Of Constant Width?
• The Additive Palindrome Conjecture Is A Risky Venture.
• How Easy Is It Really To Predict The Future?

If you would like to support me as an author, consider contributing on Patreon.