I 💗 Science

XKCD:





7 comments:

sykes.1 said...

Tufte recommends you first erase any fitted lines and just look at the raw data. In this case, there may be (may) some mild uptrend. But really there is no pattern.

Grad students have a very hard time believing there is no trend or pattern in some data.

Douglas2 said...

A note about XKCD comics:

All but about two of the near 2400 that have been released have a "mouseover" title (on the web, if you hover the mouse pointer over the image additional text is displayed, on mobile you have to click on an alt text button)

By posting an image of the comment that leaves out the alt text, you're leaving out a significant part the comic. This one is available here:

https://xkcd.com/2048/

A second note about XKCD:
There's an unofficial group wiki built for the benefit of people who don't understand either the joke or some of the background necessary to get the joke. It is at explainxkcd.com

So this one is at explainxkcd.com/2048

E Hines said...

Nobody did any Cook's D analysis on any of the blatantly obvious (say I) outlier data points....

Eric Hines

Tom said...

Also, I believe if you put the mouseover text from the original comic into the title field of the image, the text will appear when you mouseover the comic, pretty much like the original. In HTML, it would look like this:

img src="https://xkcd.com/2048/" title="Cauchy-Lorentz: 'Something alarminly mathematical is happeneing, and you should probably pause to Google my name and check what field I originally worked in.'"

Of course, you'd need the /> and < signs at end and beginning. I left them off because I don't know how they'd be interpreted in the comments area.

Tom said...

Meh, posted without editing that title text. Ah well. You get the idea. It's easy in Blogger to switch into the HTML editor, find the image line, and just add title="text you want" to it.

douglas said...

"Grad students have a very hard time believing there is no trend or pattern in some data."

That's because one of humans great strengths is recognizing patterns from incomplete data- we're *really* good at it. Because we are, it's really difficult to think we can't find one- so we make one fit through rationalization.

Texan99 said...

That link works a whole lot better. For one you, you can actually read the text now.