Margin of Error Explained by a Carny Barker at the State Fair

'Hey! You There! You Are Gonna Vote
Republican! Am I Right? Am I Right?'

Have you ever had a guy at the State Fair verbally accost you, saying he can name your age and if he is wrong, you would win a prize of your choice?

Ever notice what the margins of error are that makes him a winner and you a loser?

“If I guess within 3 years of your age, you lose!”

Depending on your age, that sounds reasonable.  At age 64, your inner egotistical mind might think: “Hey! He might think I am just 50! What do I have to lose other than $1?”

At age 7, the chances of losing are much higher. Most kids’ ages are easily guessed since plus/minus 3 years would put a 7-year old in the range of 4 to 10 years old that most anyone, including the carny barker, could readily guess.

If you were 64, the carny barker would have to guess either 60 and below or 68 and above for you to win. He intuitively has a 95% degree of confidence you are between the age of 61 (3 below 64) and 67 (3 above age 64).

The carny barker has 7 years to choose from, not just the three above or below your real age. He could guess 61, 62, 63, 64, 65, 66 or 67 and still win since he could guess you were 64.

Many public pollsters are the age-guessing carny barkers of the political world. They just don’t tell you they are.

When you see public polls saying, “Joe Biden Is Up 10% over President Trump!”, what you don’t hear are all of the caveats about their sample size and makeup. Many times, poll results are published that have no relationship or bearing on what the final demographic makeup of the voting electorate will be when the counting is done.

Many public pollsters are the age-guessing carny barkers of the political world. They just don’t tell you they are.

A public poll saying Biden is up 10 is trying to make you believe the final result of the election will be 55-45. Any poll is just a snapshot of the mood of the electorate on that date with that selected sample of people. Any political person can write questions and select a favorable sample to produce the results they want to produce to help their side win and depress the other side so they don’t vote.

Since the margin of error is plus or minus 3 for either side, 55% could be on the high side by 3 points and 45% could be on the low side by 3 points, meaning the gap could be just 4 at 52-48, much closer than a 10-point blowout.

Pollsters will say they have a 95% degree of confidence, meaning that they believe that 95 out of 100 times taking the poll with the same sample set of respondents, the result will be in the 55-45 range, including 52-48.

Nothing is ever really certain in politics, which is why we actually have elections and don’t rely on polls to elect our leaders.

Any statewide poll in North Carolina that is not based on interviewing a sample that will approximate a final voting electorate population that is roughly 74% white, 20% black, 5% Hispanic and 1% Asian or other races with appropriate educational level, income status, political affiliation and age distribution is suspect.

Six of the last 10 presidential elections in N.C. have been decided by less than 5 points. Only Bush 43 (twice), Bush 41 and Reagan ran away with double-digit victories in North Carolina.

Chances are there will be a close election on Nov. 3 in many statewide and district races. Emotions are running high amongst partisans on both sides, which guarantees increased enthusiasm to vote.

Just don’t let a carny barker pollster drive you crazy until then.

(first published in North State Journal 10/21/20)

Do You Want Better People to Run for Public Office?
Support the Institute for the Public Trust Today

Visit The Institute for the Public Trust to contribute today