How to read a political poll…

I watched MSNBC this morning and the MSNBC personality was discussing a recent Wisconsin Poll which show Biden 47 percent and Trump 43 percent and the personality stated that the poll was within the margin of error. There were 2 polls. One of registered voters and the other of likely voters. The 47/43 number are for likely voters. My analysis is for registered voters which was 46/40 numbers. But the concept is the same.

Here is the poll that was being discussed. Read the full poll with a click here

From the Marquette Law School Release

A statistics how to

So the poll is Biden 46 and Trump 40 with a margin of error of 4 points… so the range for Biden is 42 to 50 and the range for Trump is 36 to 44

The likely number measuring the population for Biden will fall between 42 and 50 and the likely number measuring the population for Trump will fall between 36 and 44 among registered voters… so if Biden performs at his best and Trump performs at his best then it would be 50 to 44 or a 6 point margin. If Biden performs at his worse and Trump at his worse, it would be 42 to 36 or 6 points. If Trump performs at his best or 44 and Biden performs at his worse it would be Trump 44 and Biden 42…

So basically, in order for Trump to win in Wisconsin, Trump must do his best and Biden must do his worse…

THE MSNBC personality this morning didn’t say this… he simply said the poll is in the margin of error…

If you really think about it, most polls that measure the election will be within the margin of error.

What a pollster is attempting to do is measure the outcome of the population by measuring a sample of that population. So this poll attempts to identify the voting totals of Joe Biden and Mr. Trump on November 3, 2020 with a sample of the population of voters in August and September.

This poll suggest that the population vote total for Joe Biden will be 46 percent of the total plus or minus 4 points… or a potential high of 50 percent and with a potential low of 42 percent of Registered Voters and Mr. Trump will have a potential high of 44 percent and a potential low of 36 percent of Registered Voters.

Usually there is a confidence value… 95 percent or 90 percent or some other value and I couldn’t find a confidence value… based on my reading the confidence value suggest that if the poll is repeated there is a 95 percent chance the results of the new poll will fall within that plus or minus 4 percent value, including the final poll of the population on November 3, 2020.

Sometimes Polls are wrong…

The idea behind confidence levels and margins of error is that any survey or poll will differ from the true population by a certain amount. However, confidence intervals and margins of error reflect the fact that there is room for error, so although 95% or 98% confidence with a 2 percent Margin of Error might sound like a very good statistic, room for error is built in, which means sometimes statistics are wrong. For example, a Gallup poll in 2012 (incorrectly) stated that Romney would win the 2012 election with Romney at 49% and Obama at 48%. The stated confidence level was 95% with a margin of error of +/- 2, which means that the results were calculated to be accurate to within 2 percentages points 95% of the time. 

The real results from the election were: Obama 51%, Romney 47%, which was actually even outside the range of the Gallup poll’s margin of error (2 percent), showing that not only can statistics be wrong, but polls can be too.

A poll might report that a certain candidate is going to win an election with 51 percent of the vote; The confidence level is 95 percent and the error is 4 percent. Let’s say the poll was repeated using the same techniques. The pollsters would expect the results to be within 4 percent of the stated result (51 percent) 95 percent of the time. In other words, 95 percent of the time they would expect the results to be between:

  • 51 – 4 = 47 percent and 
  • 51 + 4 = 55 percent.

So when a TV personality says a poll is within the margin of error, what does that TV personality mean, and does he know what he or she mean?