Unlikely Voter

Conservative views on polls, science, technology, and policy

On the polling by The Conservative Journal

A website called The Conservative Journal is releasing polls now, with an archive available on the website.

The polls are definitely making the rounds on the right, thanks to results that are relatively good for Republicans. But are they worth the pixels they’re printed on? It’s hard to say.

I’m sorry if that last line came off badly, but when I first ran into these polls, I approached them the same way I would any other poll that had no background data, no methodological details, and strange-looking margins of error. For example, getting an MoE of 5 from 500 LVs sounds a little high to me; I would expect an MoE of about 4.2.

But that was before I ran into this initial announcement of this polling series, which makes it clear the whole thing is a big experiment not to be taken too seriously:

Here’s a little background on the TCJ Research group. The research part of the group is carried out by a great friend of mine who has been in marketing his whole life and has experience conducting real polls. He didn’t join The Conservative Journal until this summer when I contacted him about conducting these polls. I play some roll in deciding what polls we conduct and when we conduct them, but for the most part my main function with this part of the site is to relay the information to you guys and present it in a clear way. We’re no Rasmussen, and we probably won’t be included in the RealClear Politics average anytime soon, but we thought this would be fun for us and all of you. You never know, maybe we’ll get some stuff right!

If TCJ Research passed off its polls with very little data behind them as serious, polished efforts demanding respect, I’d come down on them like a hammer. But given that they’re open about being new to this, and perhaps not entirely accurate yet, I give them a pass.

That pass doesn’t include counting the polls in my own projections yet, but it does include some advice: Post some crosstabs. Explain your LV screening. Tell us how you’re calculating your Margin of Error. Maybe even tell us how you’re cooking up your call list. Transparency builds credibility more than accuracy.


No Responses to “On the polling by The Conservative Journal”

Write a Comment

Comments are closed.