Which team do you think will win the Super Bowl?

Who wore it better on the red carpet?

Who are you going to vote for for mayor?

Public opinion polls are everywhere.

Important decision makers in American government

have long relied on public opinion polls

throughout elections and important legislation.

The problem is public opinion isn't easy to track

and, often times, isn't even right.

In 1948, the Chicago Daily Tribune ran a now famous headline:

"Dewey Defeats Truman,"

they cried in big, bold, black and white letters.

The problem is

that Dewey hadn't defeated Truman.

The Tribune had relied on polls to come to their conclusion.

Whoops!

This happens all the time because public opinion polls

are either inaccurate or misleading.

So, why are they wrong?

And why do we keep using them?

First, let's start with an important term:

sample.

A sample is the group of people

that respond to questions during a public opinion poll.

A poll's quality rests largely on its sample,

and a sample can be bad in a few key ways.

It can be too small,

too narrow,

or the poll itself can be too difficult.

Polls that are too small are bad for obvious reasons.

And while you can't possibly ask

every single person in America for their opinion,

the more people you ask,

the more accurate your prediction.

Polls that are too narrow,

that only ask a certain type of person a question,

are bad, too.

Consider a poll about whether or not

the potato is the best vegetable in America.

If you only asked people in Idaho,

where the state food is the potato,

chances are that you would get a much different answer

than if you asked people in the state of New Mexico,

where the state vegetable is beans.

Getting the right kind of diversity in your sample

means making sure that your sample has a range

of ages,

races,

genders,

and geographic regions,

just to name a few.

Finally, polls that are too hard can't tell you much either.

If you're asking people for their opinions on things

about which they have no prior knowledge,

the results will be pointless.

You're better off shaking a Magic 8 ball.

It's not just the people you're asking

that can cause bias, though.

The person doing the asking is part of the problem, too.

That's called interviewer bias.

Interviewer bias is all about the effect

that the person asking the questions

has on the sample.

Humans generally don't like confrontation.

People worry that their answers may make them look bad.

Therefore, we find that people tend to give

socially desirable responses,

not necessarily their honest opinions,

because they don't want to come across

as heartless,

racist,

or bigoted.

And the way we word our questions matters too.

When polls purposely sway the answers one way or the other,

it's called a push poll

because it pushes people to answer a certain way.

"Would you vote for candidate Smith?"

is a perfectly normal question.

"Would you vote for candidate Smith

if you knew that he robs senior citizens?"

is a push poll.

So, if polls are open

to all sorts of manipulation and inaccuracies,

why are they still so prevalent?

Despite their flaws, public opinion polls provide us

with some sense of the thoughts and moods

of large groups of people.

They offer politicians the chance to pass legislation

they think a majority of Americans will support.

They help fashionistas on TV

know which star wore the dress better on the red carpet.

Finally, they make us,

the people who get polled,

feel as though our voice has been heard.

So, next time you get a phone call asking your opinion,

or if you see a poll online,

take some time to think about who is asking

and why they're asking.

Then, take that poll, and its results,

with a grain of salt

or a potato.