Talk:1572: xkcd Survey

Explain xkcd: It's 'cause you're dumb.
Revision as of 15:12, 2 September 2015 by Dgbrt (talk | contribs) (Not a transcript)
Jump to: navigation, search

Mildly interesting to note that the ordering of most of the checkbox/radiobutton lists randomise each time the survery is loaded. Also, there is at least one other comic where Randall comments about not having figured out HTML imagemaps. Anyone remember which? --Pudder (talk) 10:52, 2 September 2015 (UTC)

It was in one of his "under the logo" news bars, about him starting What If, iirc --Aescula (talk) 11:28, 2 September 2015 (UTC)

I wonder how many people, on reading 'Type "cat" here:', typed '"cat" here:'? I know I did... 141.101.98.214 11:58, 2 September 2015 (UTC)

Guilty...--Pudder (talk) 12:08, 2 September 2015 (UTC)
Me too... However you could have typed '"cat" here:', as well... (/edit: I wonder how many different entries the survey's result will reveal) (/edit2: I did not read properly... sorry. I typed '"cat"' not '"cat" here:' -.-)Elektrizikekswerk (talk) 12:27, 2 September 2015 (UTC)
I typed meow -141.101.105.86 12:41, 2 September 2015 (UTC)

Not sure if it was just me, but the comic wasn't a link at all! The cursor changed into a No cursor for me everytime I mouseover'd the comic. I went to survey using the "Bonus Link!" below the comic page. Brilliantnut (talk) 12:01, 2 September 2015 (UTC)

Never mind, this was probably due to the WebComics reader extension that I have in my browser. Brilliantnut (talk) 12:03, 2 September 2015 (UTC)

From hearing people on reddit comment about not being able to completely fill the text box (not just the visual box) with the error "Answer too long", it's caused by a 10k character limit. Presumably by Google Docs. 108.162.249.192 13:18, 2 September 2015 (UTC)

The validation choices are interesting.

  • "Enter a number between 1 and 100" rejects numbers outside this range (e.g. -1) but also reject valid responses (e.g. π).
  • "Enter your age" and "Enter the number of $SIBLING" accept invalid responses such as -1. Paddles (talk) 13:29, 2 September 2015 (UTC)

"it's possible that someone may be able to identify you by looking at your responses" Then why send those informations to Google ? I find the idea of thee survey interesting but why Google doc ? There are other options like Lime Survey. Seipas (talk) 13:37, 2 September 2015 (UTC)

Given the stated intention to make the collected dataset available publicly, there's no information-security reason to prefer another survey tool over Google. Paddles (talk) 14:21, 2 September 2015 (UTC)

Not a transcript

This is mildly interesting, but it is not a transcript. Transcripts are meant mainly for blind people and search engines. Different letter sizes and a frame are not needed. Xhfz (talk) 12:54, 2 September 2015 (UTC)

Introducing

THE XKCD SURVEY

A search for weird correlations
Note: This survey is anonymous, but
all responses will be posted publicly
so people can play with the data.
Click here to
take the survey

Or click here, or here.
The whole comic is a link,
because I still haven't gotten
the hang of HTML imagemaps.

The transcript is not only for blind people. And an enhanced layout doesn't harm them but instead it would help them. A speech synthesizer would tell them something like "headline" or "small text at bottom" so that the impaired people would get a much better feeling of the comic. --Dgbrt (talk) 15:12, 2 September 2015 (UTC)


I wonder how many people included "battery, horse, staple, correct" in the five random words box.

I typed ');drop table survey; -- at the end of the random characters text box. I must have been the first person to think of that because the survey was still working. Jeremyp (talk) 13:46, 2 September 2015 (UTC)

Tables Vs Bulleted List

The list of questions and possible responses has been added to the explanation by myself and xhfz, in different formats. I went for a wikitable, xhfz used a bulleted list. Rather than just overwrite each other, I think we need to have a discussion on which is the best choice. The reasons I believe a wikitable is the best option:

  • Far better expandability, in anticipation of survey results
  • More structured and neater presentation

In general I tend to lean towards tables, but it is probably a constructive discussion to have for the wiki as a whole. I would be interested to hear opinions of bulleted list vs tables in these types of situation.--Pudder (talk) 13:42, 2 September 2015 (UTC)

If we have a table we need colspan instead of rowspan. On the other hand a table is very difficult to maintain. In addition, the table didn't have space for explanations (another column, maybe). Xhfz (talk) 13:44, 2 September 2015 (UTC)

If you used colspan, questions like "How many of these 20 words do you know" would be excessively wide. Perhaps a combination of colspan and rowspan, or simply a single cell with the responses listed as comma seperated list. As far as adding a column for explanations, its pretty trivial. What I'm getting at is that perhaps the format of a table would need to be optimised, but that is entirely feasible.
I agree with you as far as tables being more intimidating to edit and maintain, but once set up they aren't that bad.
Bulleted lists (to me at least) look messy, and tend to lack a coherent structure. As more information is added, sub-levels and sub-sub-levels are added without much thought as to the overall intent. --Pudder (talk) 14:14, 2 September 2015 (UTC)

By colspan I mean this:

Question Possible Answers
When you think about stuff on the internet, where do you picture it being physically located? Even if you know it's not really how things work, is there a place you imagine websites and social media posts sitting before you look at them? If so, where is it?
Multi-line text box
Which of these words do you know the meaning of?
Slickle
Rife
Soliloquy
Fination
Stipple
Peristeronic
Modicum
Trephony
Tribution

Xhfz (talk) 14:44, 2 September 2015 (UTC)

Missing questions

Something notably missing which would have greatly helped later analysis was a question about where someone is - Country and/or State. Some of the questions and answers will be differently understood because of that (eg meaning if 'sandwich') --141.101.98.65 14:23, 2 September 2015 (UTC)

I think that is the whole point though, to provide a data set that actively attempts to prevent the obvious simple analysis. There are plenty of statistics on how people from place A are more likely to do thing B. What I want to know is "How many people who would class a taco as a sandwich and can drive stick shift are able to juggle?". Also, is it true that most people think they are above average drivers? --Pudder (talk) 15:09, 2 September 2015 (UTC)