Why do two different polls on the SAFE Act have such different results?

Posted at 1:54 PM, Mar 19, 2018
and last updated 2018-03-19 14:06:58-04

This morning, we posted the results of a new Siena College poll on the SAFE Act which showed that support for the gun law had increased across the state, with 61% of voters supporting the law. Immediately, hundreds of people took to our Facebook page to tell us just how wrong we were.

So, we created a poll, asking your thoughts, and so far the results have been overwhelming. As of 2 p.m. Tuesday, more than 80% say that they oppose the SAFE Act.

So how can this be? How can these two polls be so sharply divided? One has to be "Fake News," right?

Not exactly...


How Siena College, and most scientific polling firms, handle their polling

When Siena College holds a statewide poll, first they try to find the right people to ask. What determines the "right people?" They look for a mix of people that roughly match the voting electorate of the state.

That means since about 50% of the state are registered Democrats and 22.7% are registered Republicans, according to the state's latest voter registration totals, Siena College tries to have their respondents match: that means 50% Democratic voters, 23% Republican voters, and the other 27% from the other political parties, including Independent.

They also look to match by geography: Since about 42% of the state lives in New York City, about 42% of those participating in the poll should also be from New York City. There are a lot of other factors as well -- but the basics are to create a group of people that closely match those expected to vote in the next election.

Even in national polls, the actual number of people polled are quite low, and the margin of error is usually quite high. (In this poll, it was +/- 4%, which means the poll could be as much as 8% closer) That's why when a poll is "close," like in the last Presidential Election, they can often be found to be 'wrong,' even though the final results end up in that margin of error.


The one thing not being reported in Siena's poll... the breakdown by region -- and we're guilty of this too, although we've since gone back and updated our initial story. According to Siena, support in "upstate" New York, of which we are a part, is at 48% and opposition at 42%. That's much different from the 61-28 split statewide.

But it still shows more support than opposition, and that's not what we found at all in our Facebook poll.


Where our Facebook poll differs

It differs in a lot of ways; first, since the majority of our Facebook fans live in Western New York, it's not a statewide sample. And we've seen that Western New York does have a different voting makeup than New York City.

But, not all of our fans are from NY and many have shared it with your friends who may or may not live in New York -- so there's no real way to tell how representative our poll is of true Western New Yorkers.

Secondly, one reason it may be so one-sided is due to how it's shared. If you voted against the SAFE Act, for example, and shared that poll with your friends -- it's more likely than not that a majority of your friends think similarly to you and would vote the same way. The poll started off in the first half-hour about 70-30 opposing the SAFE Act, it's since gone even further to that side. This would be one reason why that could happen.

Even though as of this writing we're over 3,000 voters -- more than 3x the Siena College poll -- it's not as representative of all voters. Even in the Siena Poll, 28% statewide oppose the safe act. That's 5.5 million voters. Even a 10 or 20-thousand vote poll, done unscientifically, could find either a huge groundswell of supporters or a huge groundswell of detractors.


So what is the truth about SAFE Act support in Western New York?

It's probably somewhere in the middle. There has been a lot of criticism of polling since the last Presidential Election, (and how it's reported) including that this type of polling is largely done only by phone on landlines during the day, which tends to lead to a certain group of voters over another.

And a 4% margin of error is huge. It means support could fall between 44% and 52% in "upstate" New York in this poll if Siena's polling is correct, and opposition between 38% and 46%. Those are huge ranges -- and a possibility that the opposition really is the bigger group.

As with any poll, it's a good idea to take it with a bit of healthy skepticism. (Which many of you have done!) But I wouldn't dismiss it either, even if you are certain that it has to be wrong. The current system for polling is far from perfect but it's still the best way we have of predicting what a larger group of people think on a given topic.