- reverse pyromania
- Posts
- On the street and the epitome of vague
On the street and the epitome of vague
Don't believe everything you read about polls, with an example
mister mix
March 08, 2025
I think the New Republic generally practices good journalism, and I’m a pretty regular listener to Greg Sargeant’s Daily Blast podcast. That said, I’m going to use one of their recent pieces as an example of how journalists sometimes cite polls to make a point when the poll doesn’t make anything like that point.
The headline is “Democrats’ Efforts to Stand up to Trump Torched in Devastating Poll”.
Democrats’ so-called resistance to Donald Trump is getting bad reviews among voters.
A new survey from Blueprint, a liberal research firm, found that registered voters have lost the plot on what—if anything—Democratic lawmakers are doing to oppose Trump’s sweeping agenda to shrink the federal workforce, slash essential government programs, and roll back regulations and rights.
A whopping 40 percent of respondents said that the Democratic Party doesn’t have any strategy at all for responding to Trump. Meanwhile, 24 percent of respondents said that the party did have a strategy but that it wasn’t working, and only 10 percent of respondents said that it had a good strategy.
So, that sounds bad, right? I mean, there are polls out that say that Democrats themselves disapprove of the way the Democrats are doing things, and I was interested to read any new data from this poll. So I followed the link in the piece and looked at the poll details. And it’s an immediate womp womp. First, the poll was in the field two weeks ago (the TNR story is from Thursday). Then, there’s this:
So, you’re telling me that a sample that is 47% approving of Trump is 40% disapproving of Democrats’ response to Trump? Let me show you the opposite of my shocked face.
The pollster, Blueprint, is supposedly a liberal polling outfit. And, there is some interesting information in the poll. For example, this sample’s biggest approval percentages are for immigration issues. 72% approve of Trump’s stay in Mexico prior to asylum policy, and a net 10% approve of Trump’s border security policy. I temper my interest in that result by noting that the poll was in the field when Trump was making big noise about immigration.
Anyway, I find these kind of polls not very interesting. They’re a quick sample of generic opinion of a small sample of a broad swath of America that can be influenced a lot by the way that questions are worded. Campaigns don’t use polls like this so far away from an election. Instead, they test different wordings of the same message and judge response. They also adjust their samples to areas that may have different socio-ethnic-economic characteristics. Their polls look to see where public opinion can go, not where it is.
The poll that’s the subject of the TNR piece is just an ephemeral sample of people’s off-the-cuff answers. I’d be interested in the results if, over multiple months, the same polling methodology yielded some consistent results. Otherwise, it’s just a hook for a story to get a click.
Reply