Talk:2569: Hypothesis Generation
I removed a couple of paragraphs that seem to be more about the the editor's opinions on education and the philosophy of science than the comic. -- Nitpicking (talk) (please sign your comments with ~~~~)
- I had a look at what you removed:
- I think this is a personal tutorial, or else bad teaching/learning by the intense Lenhart/Cueball interaction to the exclusion of all those unpictured.
- It is indeed important not to do tests and then work out what kind of hypothesis would lead to the inevitably significant results you have already seen to stand out.
- That said, the bits could have been less wordy.
- (I'd be tempted to add Reversion To The Mean as a key problem of reacting wth ideas to prior 'random' test-results, but that wouldn't help to shorten anything!) 172.70.91.126 17:38, 18 January 2022 (UTC)
Also, what the heck happens on 31 January, as mentioned in the top image at xkcd.com? -- Nitpicking (talk) (please sign your comments with ~~~~)
- Haha, this is Randall teasing us about our hypotheses about the Countdown over at https://www.explainxkcd.com/wiki/index.php/Countdown_in_header_text Sure (talk) 13:01, 18 January 2022 (UTC)
- I just added this link Countdown in header text at the top of the comic aws for the previous four comics. See there. (and sign you comments ;-) --Kynde (talk) 13:03, 18 January 2022 (UTC)
- Oh, and I just read the title text. Totally makes sense from Randall's perspective! Sure (talk) 13:04, 18 January 2022 (UTC)
- I really do not think this comic is a comment on the countdown. But I can see why it could be funny for us here at explain. But I'm quite sure that Randall stays away from all these fora discussing his comics. We put way to much importance to this site if we think he actually comes here... ;-) --Kynde (talk) 13:14, 18 January 2022 (UTC)
- Oh, and I just read the title text. Totally makes sense from Randall's perspective! Sure (talk) 13:04, 18 January 2022 (UTC)
- I just added this link Countdown in header text at the top of the comic aws for the previous four comics. See there. (and sign you comments ;-) --Kynde (talk) 13:03, 18 January 2022 (UTC)
I'm too shy to do the edit myself, but shouldn't the explanation mention how the comic is going meta ? Like, Cueball is generating a hypothesis about hypothesis generation! (btw, how would you test that kind of hypothesis?…) --Anonymous reader 172.70.86.22 (talk) (please sign your comments with ~~~~)
- You are never more anonymous than your IP, so please sign anyway. I think the meta of this is already implicit in the current explanation when it says: "He has made a hypothesis about how to generate a hypothesis. " --Kynde (talk) 16:23, 18 January 2022 (UTC)
- "You are never more anonymous than your IP, so please sign anyway." All right, sorry. Let's say I never commented yet, and I don't mean to start doing this regularly (although… we'll see!) (Understand: I don't have an account, and I don't want to register atm.) By the way, the IP signoff seems broken: I have nothing to do with "172.70.86.22", and it's faaaar from where I live, according to GeoIP. Back to my original point, I though it would be better to make an *explicit* reference to the meta level. Since it's an explanation, it ought to be made as clear as can be. But here's why I didn't make the edit myself: I don't usually edit wikis, so I'll trust your better jugement. --Not So Anonymous reader
How science works in primary school:
- Consider a question and generate a hypothesis
- Study the literature to understand the problem space
- Design a methodology to test the hypothesis
- Perform the experiment and record the results
- Analyse and interpret your results
- Draw conclusions and write a report
How science works in practice:
- The sponsor comes to you with an invention he wants to sell
- Design a methodology to showcase the invention's advantages
- Run the experiment and hope the undocumented fragile DAQ system works
- Start with the test cases where the invention is expected to perform best
- Scrap the remaining test cases once it falls behind existing tech
- Graph the data and write a paper merely describing what the graphs show
- Skim any prior literature on the design to add citations
