I think that’s a joke, and if someone is dumb enough to enter the establishment because they think a dentist’s opinion means anything in terms of culinary taste, they deserve it.
I listen to a podcast and the host regularly makes outlandish claims when pitching his sponsors. I know he’s joking, and laugh at anyone that buys something actually believing it.
It’s nothing because false advertising requires a reasonable person’s reading. The fact that you looked at it and immediately recognized it as a ludicrous joke probably protects them providing that you’re reasonable. If I were to invent a drink called ‘AstroSoda’ and put on it ‘Favored by Martians everywhere’ you would likely not be misled that Martians actually enjoyed my drink and so it would not be false advertising.
Something like 500 people, IIRC.
Just speaking from a research best-practices standpoint, 100 respondents is about the minimum I’d want to see in order to feel that the results were even directionally accurate. The nature of how statistical sampling works is that, the more interviews you do, the more confident you can be that the results of your survey will be pretty close to the results you’d see if you did a census (i.e., if you surveyed every single person in the relevant population).
So, this means that, generally speaking, conducting more surveys gives you more accurate results. But, the power of additional surveys falls off as you add more. For example, going from a sample of 50 to a sample of 100 adds quite a bit of statistical accuracy, but going from 100 to 150 doesn’t add quite as much, going from 150 to 200 adds even less, etc.
Generally, you don’t get a whole lot more statistical accuracy by going above 300, and definitely not a lot more once you go above 500 (which is why a lot of political polls are done with sample sizes in the 300 to 500 range). About the only reason you’d do more surveys than that would be if you want to be able to look at sub-groups within your sample with high accuracy (i.e., men and women, younger respondents and older respondents, etc.)
Also, I should note that the information I shared about claim substantiation research in the U.S. isn’t, as I understand it, something that’s codified in a FTC regulation. Rather, it’s considered to be best practices / guidelines for advertisers, based on the history of claims that have been challenged in front of the FTC, the claims that have been upheld (versus the ones that have been found to be impermissible), and the detains of the research that had been done to support successfully-upheld claims.
Also: IANAL. 
Except there is no reason to take restaurant advice from dentists. It’s a joke, and we don’t call jokes lying, even when they involve statements that aren’t factual.
For those replying with the implication that I don’t know it is a joke, allow me to respond. I know it is a joke. I found it funny (see OP). The question is more if there is some jerk out there who wanted to stir up trouble could they based on false advertising? It seems like it being obviously humour the answer is no. Which is good, because I don’t like jerks having the ability to stir up trouble based on stupidity.
For population of 300M people (or any large population over about 250,000), a confidence level of 95%, and a confidence interval of ±5%, you would survey 384 people. This means that if you surveyed 384 people and got the results that 80% of people prefer Crust® Toothpaste, there is a 95% chance that between 75-85% of the entire population prefers Crust® Toothpaste. A 95% confidence level is pretty much the standard for this type of thing; I’m not sure of the typical confidence interval.
The takeaway is that you can survey surprisingly few people to get a statistically valid result of a large population.
Absolutely so. One will frequently see criticisms of political polls: “How can they say the results are real if they only polled 500 people?” “I’ve never been asked to take a political poll, so they can’t be valid!”
They’re understandable questions, but not easily answered without trying to teach someone more about stats than they’d be willing to sit through.
Frankly, there seems to nearly always a jerk who wants to stir up trouble. In this case, I’m not sure that said jerk would be able to stir up a whole lot of trouble, but that’s just my own somewhat-informed opinion.
Except it doesn’t; it’s clearly a joke.
Because Pepsi did not actually have to give away a Harrier Jet then I would say this is not false advertising.
Well, they could certainly try. A jerk can stir up trouble based on nothing at all. You can sue someone for pretty much any reason.
Or it could mean the folks from the dental clinic next door went to lunch, and when someone said “Where are we going”, 4 out of 5 of those dentists said “Let’s go to that restaurant.”
if you assume it was a valid statistical poll of many dentists (more than 5) you are jumping to conclusions.
Local restaurants near me (Seattle area) make grandiose claims frequently. You can always tell which claims are serious because they’ll be specific and substantiated. A claim like “Voted #1 Burger in [local] Magazine” will be accompanied by an excerpt from that magazine. One that just says “Best Burgers in Town” without specificity can mean the proprietors have confidence in their burgers, but that’s it.
Either they did the survey or they didn’t.
Follow their defense for these two cases.
If they didn’t do it, their defense is that absurd that they surveyed dentists…
If they did do it, then its an absurdly small number.
I think its absurd for one or two reasons, therefore.
Also, I’m one of australia’s best dentists. Well there’s at least 5 people worse than me, so I’m not the worst, and therefore, one of the set of the “best”… (top 25 million out of 25 million … + 5 )
There are numerous subtle things going on here that change it from lying to a joke:
Basically, the whole joke is applying a well known advertisement claim in a context where it doesn’t fit. The claim does need to be substantiated when it’s used in a relevant context, but I think most people would see it and recognize the phrasing from gum or toothpaste commercials and chuckle.
As an example, here’s a few claims that probably wouldn’t fly:
Which are all specific claims where the expertise of the group ostensibly being polled lines up with the claim being made.
It is completely inappropriate to solicit an opinion from the comatose. I would exclude them and change the marketing to 100% of dentists polled prefer …
And I’m surprised no one has brought up this suitagainst a misleading film title yet.
Chung King Chinese Food used to have really funny commercials in the early 70s. In one, they claimed that four out of five doctors preferred their food. Cut to a panel of five doctors, four Asian and one white.
I assume that this only works if you make some effort to be sure that the 384 people are chosen from a wide variety of types. If, for example, you did your survey in old people’s homes, then it would not be valid. The same would apply if you did it on Kidzworld.
Yes, I meant to say that. Random selection for surveys is a science unto itself.