37

Are the surveys in the gameshow 'Family Feud' (and its many international derivatives) systematically fake?

When I used to watch the British version, 'Family Fortunes', I couldn't imagine that they'd bother with all the trouble, time and expense of conducting endless trivial surveys, each of which had to include 100 participants; not least because of the unreliable answers they would have got.

On the show, the surveys' results always comprised a small number of repeated answers, and I don't recall any of the survey answers to be stupid. Whereas I think in real surveys, the number of different answers would be much higher, and would include all sorts of wrong and stupid answers.

Has anyone ever participated in one of these surveys?

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
EmmaV
  • 517
  • 1
  • 4
  • 10
  • 6
    As for "small number of repeated answers", I don't think it unlikely that they'd shoehorn the more diverse ones into neat little categories. For example: "Name things kids don't like". "brushing teeth" could also include "using mouthwash", "using floss" or even "washing their used toothpaste down the sink". – user4843 Jun 12 '15 at 08:10
  • 2
    At least in the late 90's they did disclose the point of the survey by the end of the call. I was surveyed by them once in college. – Bill Jun 12 '15 at 21:06
  • In fact, I have seen cases where Family Feud answers were wrong answers. – GEdgar Jul 20 '16 at 15:17
  • 3
    They're not technically wrong as far as the show is concerned since the show clearly states "we surveyed some people and this is what they said". So, the people answering the survey may be wrong but that's irrelevant. The intent is to entertain, not educate. –  Jul 21 '16 at 10:55

1 Answers1

55

According to this article, And the Survey Says... (by Carl Bialik, 2008) in the Wall Street Journal, the Family Feud show does conduct real surveys.

Surprisingly, they don't just conduct polls of the studio audience (which seems like it would be an easy source of survey participants and people that would be happy to do the survey). Instead, they conduct telephone surveys using an outside polling agency.

The agency chooses participants at random and asks 30 - 40 questions in each call out of a list of around 100 from the show's writers. They do not reveal that they are conducting the survey on behalf of the Family Feud show, so no one would know if they were a part of a Family Feud initiated survey. In the show's early years, they did use self-selected participants from the show's mailing list.

Granted, polling only 100 people does not lend much statistical validity to the survey (the article says that the results have an error margin of plus or minus 10), but still, their polling methodology is more rigorous than I would have expected for a TV game show.

The WSJ article was written back in 2008, so it's possible that they are no longer using the same methodology, but I couldn't find any more recent references.

ChrisW
  • 26,552
  • 5
  • 108
  • 141
Johnny
  • 2,929
  • 1
  • 26
  • 24
  • 5
    Interesting, did it say if these polls were *specifically* and *exclusively* commissioned for the show? I wonder if they might be part of wider "lifestyle trends" surveys which the polling company routinely produces and sells, or something the polling company uses for publicity? – user56reinstatemonica8 Jun 12 '15 at 10:07
  • It says, `The surveyers don't disclose that the questions are for "Family Feud."` – ChrisW Jun 12 '15 at 10:27
  • 13
    I remember the host usually saying something like "the top five answers are on the board," suggesting that outlying answers are discarded if they are stupid, inappropriate, or only appear once. I'm also pretty sure that many boards didn't actually equal 100 points, but were typically closer to 85 points or so. – phyrfox Jun 12 '15 at 16:18
  • 3
    According to the Wikipedia article on Family Feud "Only answers that receive two or more responses can appear on the board." That would explain why there aren't so many diverse answers and why they don't add to 100. Also, I believe they group similar answers. If asked Name an animal with 4 legs, if someone says Poodle, that would count in the answer "dog". – user1008646 Jun 12 '15 at 19:46
  • 1
    @user1008646: So then [Picard's Flute](http://vignette3.wikia.nocookie.net/memoryalpha/images/0/02/Picards_Flute.jpg/revision/latest?cb=20100104143941&path-prefix=en) could never be an answer. – Matt Burland Jun 12 '15 at 20:12