WitCH 18: Making Serena Pointless

We have lots of catching up to do, WitCHes to burn and whatnot. However, we’ll first try to get in a few quick topical posts (give or take a couple weeks …). This first one is half-post, half-WitCH. We had planned it as a post, but it then seemed worth letting readers have a first whack at it; as always, readers are welcome and encouraged to comment below.

Serena Williams was back at Wimbledon this year, for the ninety-fifth time, almost grabbing her eighth title. This phenomenal athlete was also the subject of some media fluff, of the type that always accompanies these events. It was reported, pretty much everywhere, that

“One in eight men think that they could score a point off Serena Williams”.

Oh, those silly, silly men.

Twitter, of course, lit up over these “delusional” men and the media gleefully reported the ridicule, and more often than not piled on. A rare few articles gave tepid consideration to the idea that the men weren’t delusional, and none more than that.

The Serena fluff came courtesy of British polling firm YouGov, popular with those comforted by the illusion that someone cares what they think. Specifically, YouGov asked:

“Do you think if you were playing your very best tennis, you could win a point off Serena Williams?”

YouGov announced the result of the poll on Twitter, with catchy headline and accompanying graph:

“One in eight men (12%) say they could win a point in a game of tennis against 23 time grand slam winner Serena Williams”

Note that 3% of women also answered that they could win a point; we could see nothing in the media reports questioning, much less ridiculing, this percentage. (The missing percentages correspond to people who answered “don’t know”.)

On the YouGov website, the poll is also broken down by age and so on, but there is little information on the nature of the polling. All we are told is:

“1732 [Great Britain] adults were questioned on 13 Jul 2019. Results are weighted to be representative of the GB population.”

OK, so now the WitCH aspect. What is wrong with the poll? What is wrong with the reaction to it and the reporting of it? As always, feel free to respond in the comments. (You might try to keep your answers brief, but it won’t be easy.)

Finally, to state explicitly what should be obvious, we are not in any way having a go at Serena Williams. She is a great athlete, and throughout her career she’s had to put up with all manner of sexist and racist garbage. We just don’t believe the YouGov poll is such an example, or at least so clearly so.

7 Replies to “WitCH 18: Making Serena Pointless”

  1. 12% in this case is quite ambiguous. 12 out of 86 is not 12%. Much closer to 14%. Even though we know the missing 14%, “Don’t know”, it is not really a great conclusion to say 12% think they could.

    The study is biased to begin with, since only a rather narrow portion of the population would likely have answered the survey (this may or may not be true, but likely true and for a fair study should have been made explicit…)

    Even if you do accept the 12 out of 86 (or 100, or whatever denominator the publisher thought convenient), this is a point estimate (and a rather bad one). Taking into account the sample size and other issues, any reasonable confidence interval… OK I’m over this now… back to class.

  2. And, on a side note, if she double faults (which statistically happens in a lot of her matches), does this count as winning a point? If so, 12% seems like an underestimate and 3% for the women also seems quite a long way under.

  3. To start, the question is clunky and could be misinterpreted. What does “your best tennis” mean? It is vague and could mean playing a good game at your current level of ability or a level you think you can achieve with practice.

    Next the question says “win a point”. Well this is also open to interpretation, as you may think that through your tennis abilities to you could win just one point or just through chance (Serena could make an unforced error). Winning one point doesn’t seem unreasonable against any tennis player, especially if someone is decent at tennis themselves.

    So the data collected is not valid due to the poorly worded question.

      1. Or indeed just playing the one game (and ‘game’ is the word used in the question), that is, a minimum of 4 points (if so, who’s serving?) Personally I wouldn’t fancy my chances under those circumstances …..

        Not sure how YouGov collects its samples, but if it’s an ‘opt in’ survey (where the reward in this case is intrinsic rather than extrinsic) then I agree with Number 8 that there’s a high likelihood of bias (I’d imagine most people would opt in because think they *can* win a point, particularly if they think that they are playing over at least two sets – chances of a double fault over two sets …?)

  4. Is there a difference between “thinking” you can win a point and “saying” you can win a point? The two seem to be used interchangeably in the reports.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.