In the news today is a report on KDKA from Jon Delano that Lukey Hizzoner Master Lil’ Ravenstahl MC Skate Boi is currently sitting ugly with a scant 19% approval rating.
This data comes from local CMU — “offshoot” is not the right word, but let’s just say it was borne out of CMU — company CivicScience run by CEO John Dick.
Some, including Chris Briem of NullSpace, have questioned the validity of such polls in light of oftentimes unreliable results, drawing a comparison to a KQV poll that showed DeSantis leading Lukey by a 74% to a 34% margin prior to the mayoral election that Lukey, as we know, handily won.
I asked John Dick, who in the KDKA interview is quite adamant that their numbers are accurate, to ask him how he can be so sure about their methodology.
I’ll have you know that I asked for the “elevator speech” and judging by the length of his response, I assume John and I are riding an elevator TO MARS. I kid.
Lots of good info he gave me, but here’s the important-sounding parts as to why the Mayor, it seems, really does have a 19% approval rating when CivicScience’s recent poll was adjusted for gender:
We trust in the law of big numbers. While there is bias in any human sample, those biases can be accounted for when you have large enough data. In the case of the Ravenstahl #s, a 26,000+ person sample included enough people in every key demographic to build reliable models. Consider that a standard polling sample includes 500-1000 people (hence, margins-of-error in the 3-6% range). Throw our data (over 110MM polls and counting) into the machines built by these geniuses here from CMU and you can do things that can’t be done with 500 data points.
The KQV poll was NOT using the machines the geniuses built and also used a sample of about 2,500. Makes sense why perhaps their poll got it so wrong. Plus who knows what the listener demographics are at KQV, another thing to keep in mind. This is why you could put a poll on FoxNews all, “True or False. Barack Obama is an illegal alien.” and it would appear that 95% of Americans think TRUE.
Web connectivity is virtually ubiquitous among all but the oldest and most indigent populations. Compare that with the landline telephone (still the prevailing means of opinion research). Only 74% of US households have a land phone. Among those, at least half subscribe to Do-Not-Call Lists and/or CallerID, rendering them unreachable. The people left over are an extremely biased group (older, suburban/rural, etc). But, by aggregating results from millions of polls over news websites (GenX, Baby Boomers, and Moms), Social networks (Youth, Early Tech Adopters, and increasingly everyone), trade publications (professionals), and other online destinations, we are able to reach a much broader cross-section of the population.
Another reason Lukey might appear to have a higher approval rating in a phone survey, if there is one, could be because of that oldster thingamobobbibbity. You have a better chance of reaching an old person at 2:00 in the afternoon on a Tuesday than you do a 45-year-old working mom.
I could drop the names of some really smart people, including well-known economists, anthropologists, political scientists, consumer researchers, political pollsters and return customers who have validated our work. Numerous independent technical and academic white papers have also been published that substantiate what we do. I will spare you all of that.
Thanks, this is my floor anyhow.
Shit. Mars is COLD.
P.S. Machines the Geniuses Built? BAND NAME!