2618: Selection Bias

Explain xkcd: It's 'cause you're dumb.
Jump to: navigation, search
Selection Bias
We carefully sampled the general population and found that most people are familiar with acquiescence bias.
Title text: We carefully sampled the general population and found that most people are familiar with acquiescence bias.

Explanation[edit]

Blondie is giving a talk at the conference "Statistics Conference 2022." She asks for a show of hands from those attending the conference on whether they are familiar with selection bias. She uses this as part of her presentation, concluding that most people are therefore aware of what selection bias is.

Selection bias is when a survey or poll of some sort comes up with incorrect results due to those who were asked. For example, if you asked a group of people how many acres of land they own, your average number will be higher if you ask a group of farmers rather than a group of city residents.

The joke is that she is thus falling for the very thing she's trying to explain. A statistics conference is likely to have an audience consisting of professional statisticians, or at least people interested in the subject, and it is expected that most of them would thus be familiar with any mainstream statistical term, like selection bias. Had she asked a random sample of people in the street, many of them would likely not be sure what selection bias is. This effect is also the subject of 2357: Polls vs the Street.

This joke also ties into how statistics as a whole can be highly counter-intuitive and sometimes almost paradoxical, where things like the Monty Hall problem and survivorship bias lead people into thinking the answer to a problem is definitely in a place it's not. That Blondie, presumably a statistician herself, made this kind of (potentially deliberate) error is professionally embarrassing but not unprecedented.

The title text refers to Acquiescence bias, which is the tendency of people to respond positively to positive questions, for example, "Are you familiar with the famous webcomic xkcd ?" is more likely to generate the answer yes than "Are you familiar with that webcomic for engineers that nobody else understands until they go to Explain xkcd?" Acquiescence bias is not a widely known concept,[citation needed] making the results of this poll suspect; similar to the selection bias example above, the reason that the general public seems familiar with acquiescence bias may be because the surveyor themself fell victim to promoting acquiescence bias.

Transcript[edit]

[Blondie is standing on a podium behind a lectern with a microphone. She is standing under a hanging sign with large text. In front of the podium is an audience of five seated persons all with their hands raised above their heads. The audience includes two guys that look like Cueball, Hairbun, and two other persons with dark and blonde hair.]
Sign: Statistics Conference 2022
Blondie: Raise your hand if you’re familiar with selection bias.
Blondie: As you can see, it’s a term most people know...


comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!

Discussion

Why is the site giving errors so much? 172.70.130.121 21:19, 11 May 2022 (UTC)

It just went down for about ten minutes. Once in a while I was able to get pages to load, but it was rare (probably 3-5 times). Most of the pages that did load had no CSS. 172.70.130.161 22:03, 11 May 2022 (UTC)
I blame a secret global infiltration by brain-eating extraterrestrials bent on steadily eroding our infrastructure while we kill each other in stimulated warfare. Do you agree with this common hypothesis? 172.70.114.241 01:00, 12 May 2022 (UTC)
No. 141.101.104.240 05:56, 22 May 2022 (UTC)

I love the current explanation of the title text, thanks Kev ;-) --Kynde (talk) 06:38, 12 May 2022 (UTC)

I am the individual on mobile who revised the referenced text so as to mention engineers. This above comment by Kynde was added after my revision, but links to the previous version which was not current at the time of their post. I do not know why they call it current. 162.158.62.123 07:07, 13 May 2022 (UTC)
Hear, hear! A true gem! 172.69.134.17 07:37, 12 May 2022 (UTC)
Thank you kindly, meanwhile I return to find a much improved BALLPIT APPARTMENT and have come to say I got that reference and heartilly approve'Italic text'Kev (talk) 22:27, 12 May 2022 (UTC)
I agree wholeheartedly, but what does it have to do with selection bias??? 172.70.114.241 01:06, 13 May 2022 (UTC)Bumpf
I don't get it at all, which kind of defeats the point of Explain XKCD when people are adding inside jokes. 172.70.130.121 05:06, 13 May 2022 (UTC)
it's a reference to 150: Grownups and 162: Angular Momentum, two very well known xkcd comics (so much so they are the footer comics!)172.69.69.250 13:35, 13 May 2022 (UTC)Bumpf
Neither of those have anything whatsoever to do with this comic, and I never look at the footer, assuming I even read comics on the XKCD website rather than here. 172.70.126.221 18:07, 13 May 2022 (UTC)
I wonder if they are practicing familiarity with xkcd ! Anything drawn from xkcd makes you sound like a nettie cause randal half parroted nettie culture to make stuff. 172.70.110.209 00:00, 14 May 2022 (UTC)
They're randomly drawn phrases out of a brain filled with XKCD, so there may be some bias in what words are selected. There may be a risk of becoming so meta even for any acronyms Kev (talk) 17:20, 13 May 2022 (UTC)

Acquiescence bias is often used in sales to try to induce a bigger sale. Asking, "Will there be anything else?" instead of "Is that all?" can make a significant difference in the amount of a sale, even though you are still leaving the decision up to the customer. Nutster (talk) 06:30, 13 May 2022 (UTC)

WHY is the explanation unsure if she made the error accidentally??!? It's clearly intentional and part of her presentation! Ugh, one thing that always bugs me about this site, how often people express uncertainty where there's certainty... Don't be so timid! LOL! NiceGuy1 (talk) 05:13, 14 May 2022 (UTC)

Of course she makes the mistake and is not aware of it. This becomes clear when they do the same in the title text. The joke is that she is not aware of it, or else it would not even be fun. --Kynde (talk) 18:45, 15 May 2022 (UTC)
She is 100% aware of it, she isn't making any mistake, she is DEMONSTRATING Selection Bias completely on purpose as part of her presentation on Selection Bias. It is the AUDIENCE who is participating by accident, not realizing that by voting that they are demonstrating the bias. THAT is the joke, THAT is the fun, that they are pawns in her presentation. She is cleverly using them without their knowledge or realization. Her next line could be "See? I got you!". :) NiceGuy1 (talk) 05:44, 16 May 2022 (UTC)
I disagree. If Blondie were not speaking so as to give Randall an xkcd punchline, she might have followed up by saying, perhaps, "Yes, pretty much everyone here is familiar with selection bias, but this is a statistics conference, so it's probably not representative of the general public. So now I'll tell you about some recent studies where the authors didn't realize that they had problems with selection bias ...." If the audience had failed to raise their hands, Blondie would not have known whether the audience really didn't understand selection bias, or was misleading her, or misheard/misunderstood her. --172.70.131.104 18:48, 1 October 2022 (UTC)
*sigh* Again, no mistake, no uncertainty. This is a statistics conference. To be at ANY conference you must be significantly involved in the field. Therefore, Blondie is making a completely safe bet that literally the entire audience is familiar with selection bias, as anyone involved in statistics must be. There is no chance whatsoever that anybody in the audience would fail to raise their hands, if they're being honest. Well, people who dislike participating in polls, but this is a simpler universe. Which also dissolves any possibility of not understanding, or misleading, or mishearing. Honestly, mine is the correct interpretation, the audience are pawns in Blondie's clever speech. Without the slightest doubt. Check into the concepts of K.I.S.S. and Occam's Razor, Randall doesn't include such hidden complications. NiceGuy1 (talk) 08:16, 12 November 2022 (UTC)

> 'Are you familiar with that webcomic for engineers' --

I always assumed it was mainly for scientists...

> Acquiescence bias is not a widely known concept

Depends who you are asking. 141.101.76.29 19:54, 22 July 2022 (UTC)

xkcd is ABSOLUTELY for scientists. Also for mathematicians. STATISTICIANS! Not so much for engineers. Some scientists are engineers. Some engineers are scientists and statisticians--but far too few. Whether she made the mistake of selection bias intentionally is fairly irrelevant. Someone making the mistake of selection bias INTENTIONALLY is committing some kind of fraud--but that doesn't mean they will/should be prosecuted for it. Hopefully their paper submission is rejected, and the MSM reports that it is invalid. (Hopefully.) 172.69.42.115 01:48, 11 November 2022 (UTC)

Why would the Manhattan School of Music have any reason to get involved? 172.71.178.187 03:44, 11 November 2022 (UTC)