Back in 1936, it was a simpler, albeit difficult, time. President Franklin Roosevelt was finishing his first term in office, the Great Depression was still dominating the nation, and our neighboring state of Kansas had a nice guy named Alfred Mossman Landon as governor. Alf, as he was known (although, how cool would it have been if he went by his middle name?) was quite popular in his state and was, in 1934, the only Republican governor running for reelection to win. Two years later Alf found himself the GOP nominee for president, running against the first term FDR. As you can imagine, lots of people wondered who would win.

At this same time, there was a very popular national magazine called the “Literary Digest.” The magazine came out each week, and it examined news stories and offered opinions. It was a respected journal, providing important political analysis. Oh, and it was published by Funk & Wagnall, so if you are old enough to remember the 1960s TV comedy “Laugh In,” you are amused. Otherwise, you are just wondering why I made that obscure reference (ed: yes, exactly).

So, with the election looming, the Digest embarked on creating the definitive election poll. The publishers wanted to make sure they did the largest and most comprehensive poll ever conducted. The Digest was, after all, the magazine of record when it came to American politics. And so, they set out to interview, via a mailed survey, a truly massive collection of Americans. The Digest mailed out forms to no less than 10 million folks and got roughly 2.4 million sample ballots returned. With that massive collection of data, the Digest confidently predicted that Alf Landon would win the presidency by a landslide. So there!

Well…

As it turned out, of course, it was FDR that won in a landslide, taking 62% of the vote nationally. Landon only won two states, Maine and Vermont. Landon was utterly crushed. He returned to Kansas a defeated man but would be respected until the day he died at the remarkable age of 100 in 1987. But the Literary Digest was so mocked and ridiculed that it ended up going out of business.

So, what did the Digest do wrong? As it turns out, the size of the sample is not the critical factor in conducting a good poll (trust me on the math, I have a Ph.D. in this stuff. Yup, I’m even nerdier than you thought). Rather, the absolutely key thing is to get a truly representational sample of the electorate. The Digest, thinking only that size matters, looked for ways to get a massive sample, and they settled on people listed in phone books and people with automobile registrations on file. So, they mailed ballots to those folks. But think for a second about the types of people that had phones and cars in 1936, in the throes of the Depression? Yup, rich people, who were, in fact, overwhelmingly pro Alf. But the huge majority of Americans were then (and now) not rich, and the Digest poll missed that group entirely.

Which, of course, brings me to a recent Colorado Politics story about the race in Colorado’s 3rd Congressional District.

The Democrat running to defeat Lauren Boebert, a successful businessman named Adam Frisch, released a new poll that shows the race to be a dead heat, with both candidates’ popularities within the poll’s margin of error.

If accurate, that is great news for Colorado and, frankly, for the nation. Boebert is an embarrassment for a number of reasons. A quick internet search for the term “Boebert is an embarrassment” returns 154,000 results. But CD3 is seen by most analysts as being quite red. So what, if anything, has changed?

Well, the poll’s main point is that there has been a real shift away from Boebert in the district’s GOP leaning independents. Boebert also has problems with unaffiliated voters, with 62% of them holding an unfavorable opinion of her. So, is it possible that moderate Frisch has a shot? I hope so, but frankly, it’s tough to tell.

Even as recently as my own failed run for Congress back in 2008, it was possible to conduct a fairly accurate poll by calling peoples’ home phone. Though there certainly were lots of cell phones back then, most folks still had a home landline. But today, that is just not true. So how do you correctly and properly sample a representative group of “likely voters?” Well, it’s really tough. I mean, do you accept calls on your cell that show an unknown number?

Rather than bore you with terms like “oversample” and “regression to the mean” I’ll just say polling is quite a challenge today. Polls must be blinded and unbiased, and they must also reach the right folks. Back in 2008, it cost me $16,000 for a single poll to be done and those costs have now risen dramatically.

I surely hope that the Frisch poll is correct. He seems like a good guy, and it would be great if he can replace the truly horrible representative now holding that seat. But I worry that the challenges of modern polling may have created a bit of a false picture. We’ll know come election day. Until then, let’s hope the lessons of the Literary Digest are not being repeated.

Hal Bidlack is a retired professor of political science and a retired Air Force lieutenant colonel who taught more than 17 years at the U.S. Air Force Academy in Colorado Springs.