Sunday, October 30, 2022

Reporters Continue to Misrepresent Polling Margins of Error

Last week, Rasmussen Reports released a poll showing Republican Herschel Walker with 48% support among Georgia likely voters, besting Senator Raphael Warnock who had 43%.  The poll had a 3% margin of error.   Media outlets reporting on the poll said Walker's lead was outside the margin of error.

Wrong.  The lead is inside.

One of my pet peeves about political reporting is that reporters continually mischaracterize how margins of error work.  A margin of error is a statistical measurement that accounts for the difference between actual and projected results in a survey sample.  Applied to political polling, the margin of error refers to how much the measurement of a candidate's support could be off just due to the fact that the sample, despite the best efforts of the pollsters, might not be representative of the population polled.  

Let me explain using the aforementioned Rasmussen poll.  According to the 3% margin of error, Walker could have as much as 51% support or as low as 45%.  Warnock's possible range is 46% to 40%.  So one result of the poll is that Walker could lead by as many as 11 points (51-40) while in the scenario on the other side of the possibilities, he trails by 1 point (45-46).

If you're looking to see whether a lead is outside the margin of error, one needs to double the margin of error.  In the Rasmussen example, Walker has a 5 point lead while the doubled margin of error is 6.  Again, Walker's lead is inside.

As a result of reporters continually misreporting how margins of error in polling operate, the public has come to expect an accuracy in polling results that no sane pollster would try to claim or has been able to live up to.

No comments: