I tried. I really tried to read this article from Lauren Leader, CEO of All In Together and darling of gun control advocates, without rolling my eyes or cursing inventively. It would be a gross misstatement to say I failed. To put it bluntly, Leader and All In Together created a poll that was so flawed, so slanted to “prove” women, even Republican women, want more gun control that it’s not laughable. It’s an insult to any person with a single working brain cell.

Leader starts with the “it’s us against them” attack point, with “them” being a small cabal of Republican men against the rest of American when it comes to the issue. After all, according to her, there are gun control measures  “. . . that have a majority of support from voters. . . .”. Of course, she doesn’t explain what those ways are, mainly because she then alleges that ” there’s a significant disconnect between Republican women and Republican men on guns.”

Oookay. I’ll keep reading because I just have to see where she’s going with this and what facts she has to support it. Besides, best to read it and do my cussing before church, where I can then ask for forgiveness for my thoughts. Right?

It would be easy to take the article paragraph by paragraph, even sentence by sentence and snark at its over-simplification of a very important issue, not to mention expose every fault and fallacy in the polling process. Instead, let’s just look at some of the highlights, starting with the poll itself.

To do so, you have to go beyond the article itself. The questions, the manner in which the poll was conducted, the manner in which the answers were tallied and deciphered are mysteriously absent for the most part in the article. I doubt this was an oversight. Leader’s been doing this sort of thing long enough to know most people reading her article are already of the “right sort” of mindset. Another large subset of readers will read the article but not take the time to click through to the “report” on the survey results.

I’m neither of those. So I clicked through and, yes, the rolling of eyes and the cursing continued.

The survey was fielded online from May 8-10, 2023 in English among a sample of 1,277 voters in the Likely Electorate (LV) nationwide. The base sample was 856 voters in the Likely Electorate and an additional oversample of 421 women to achieve a total of 836 women respondents in the sample, using non-probability sampling.

Wow, how many issues can you see with the above quote? Without thinking too hard and operating on my first cup of coffee, I found three serious ones right off the bat.

First, it was an online survey. So how did they choose their respondents? Were these regular supporter of Leader and her organization or members of the NRA? How did the respondents learn about the survey and did they volunteer to answer it or were they chosen by Leader et al?

Second–and I’ve already expanded the number of issues I saw at a quick glance–it was a survey conducted in English. What about those who are members of the “Likely Electorate” but who aren’t fluent in English? Hmmm, is it okay to throw out entire groups of people for the ease of tilting a survey to meet your desired outcome?

Third, you have 1,277, of which 836 were women. The description even makes it clear there was a concerted effort to make sure the survey was heavily skewed so many more women responded to the poll than men.

Fourth, noticeably absent in the description above is party breakdown. How many Democrats, Independents, or Republicans did they get responses from and how many of those were men or women.

Fifth, where in the country were these respondents from?

Finally, this was a non-probability sampling.

Combine the first and fifth concerns and you have a poll with serious issues right out of the box.

The Pew Research Center listed a number of dangers of using online polls and how to mitigate some of those dangers back in 2020. Yet, neither in the Politico article nor in the article giving more concrete (and I use that term loosely) results from the poll, we see very little to show Leader and her polling partners took those dangers and concerns into consideration.

On the scientific front, the dangers of non-probability samplings have also been noted:

A proportion of the individuals would not have been contactable by email, and this group may have different attitudes from the rest. The nature of the responses of those individuals who failed to reply to the questionnaire, the majority, is unknown. They might have differed from respondents if, for instance, busier or more stressed psychiatrists were less likely to participate. As a result, the sample identified by the researcher may not have been representative and the findings cannot be safely generalised to all those working in this field. This is a non-probability sample and, as such, statistical inferences cannot be validly made from the results.

Now, add in the other concerns with Leader’s poll and you have yourself a hot mess of a survey with questionable results.

Now, let’s take a look at some of the so-called results. Before we do, however, consider how Leader believes we have a “tyranny of the minority” (and she doesn’t mean certain Democratic leaders and powerbrokers who have been whittling away at our Constitutional Rights over the years).

You can find the “report” here in it’s glorious Powerpointesque glory.

The very first data slide prove the lengths the pollsters are willing to go to make sure Leader’s conclusions are supported.

The answers are constructed in such a way to hit the fear factor the media has carefully cultivated concerning the potential for mass shootings. There is also no definition given for what a “mass shooting” might be. Are we talking about two shootings or ten? Are we talking about a single location or multiple locations? There is no option for someone taking the poll to say they are not concerned at all. The closest is unsure which is most definitely not the same as not being concerned about such an event happening. So the basic structure of the answers is flawed and slants toward a desired result.

Another “result” claims that GOP women are less likely to think a mass shooting event will happen close to them than their Democratic or “Independent” women. The same fault with how they could answer the questions as listed above apply. But so does another concern. There is nothing indicating where these women were from. Someone living in rural Iowa will more than likely be less concerned with a mass shooting happening near them than someone in Chicago.

The crowning idiocy when it comes to “findings” is when Leader et al claim there’s a gender divide in the Republican Party when it comes to “some gun policies”. As noted above, the poll was artificially constructed so there would be more women than men answering. Specifically, if there were 1,277 respondents and 836 of them were women, of course there is going to be a “divide”. Especially if other factors were at play in making sure the sampling would support what feels more and more like preordained findings.

In other words, this so-called survey and the results coming from it are about as reliable as most other political or social surveys. Take them with a grain of salt–and possibly a stiff drink. In this case, look at who commissioned the survey, what her aims are, what her prior statements have been, and the way the survey was put together and the “data” interpreted before accepting what she says as true. Do not let her, or anyone else, convince you to give up a Constitutional Right simply because they managed to taint the results through careful wording, sampling and more.

(And, yes, I know “Constitutional Right” shouldn’t be capitalized but it is important enough in this instance to deserve it.)

Featured image: A Minuteman against the flag, created by Amanda S. Green using Leonardo.AI.