# This says it all, really

From PhD Comics.

It’s just like a game of telephone! While it doesn’t exactly fit, this cartoon reminded me of the hubub over the announcement that will CHANGE EVERYTHING that is going to be made this morning.

1. #1 Sigmund
May 19, 2009

Which journal accepted that P value?
p=0.56 ?

2. #2 Tony P
May 19, 2009

Now take that little game of telephone and extend it back say, a bit less than a couple thousand years and tell me the Bible we have today isn’t severely distorted.

3. #3 Andy
May 19, 2009

At risk of sounding like “that guy” (you know, the one who gripes in the middle of the Star Trek movie about how it’s not physically possible for the ship to have done a particular maneuver), I think it’s actually a “rho” value, which is an accepted symbol for the correlation coefficient.

4. #4 Blake Stacey
May 19, 2009

Yeah, it’s a ρ value (a Pearson coefficient) indicating a moderate but potentially interesting positive correlation between A and B.

5. #5 Drew
May 19, 2009

Not looking forward to having to explain, yet again, what a crap term “missing link” is, and how the most important fossil is “all of them, and how they all relate to each other” not any one find.

6. #6 cm
May 19, 2009

And anyway the cartoon doesn’t say anything about the result being published in a peer-reviewed journal, just that there is a result. Even more apt!

7. #7 Lilian Nattel
May 19, 2009

Great cartoon. Can’t you imagine the Dr. Seuss book explaining it?

8. #8 Mary
May 19, 2009

Ok, that is a keeper. Thanks for that

9. #9 alex
May 20, 2009

guys it is called a p-value, nothing to do with correlation. it isn’t meant to be a rho, or a Pearson coefficient. It simply gives you the probability between two values is due to chance, so in this case the correlation is 56% likely to be due to chance.

http://en.wikipedia.org/wiki/P_value

i’m quite amazed no one here has a background involving reading scientific papers (or biology ones at least, they come up everywhere for us)

10. #10 David Canzi
May 20, 2009

guys it is called a p-value, nothing to do with correlation. it isn’t meant to be a rho, or a Pearson coefficient.

I checked it with a screen magnifier. I also went to PhD Comics for an image closer to the original and enlarged it in Firefox. It looks like a rho, not a p.

Oh, noes! Could it be… might I have… too much time on my hands? Say it ain’t so.

11. #11 David Canzi
May 20, 2009

I went back to PhD Comics, found the lower case ‘p’ in the word ‘picked’, magnified it and examined it carefully. I can now state authoritatively that I have too much time on my hands.

12. #12 Julia
May 20, 2009

alex, that is totally a rho; you don’t even need to use David’s superior powers of investigation to notice that the letter is in a different font from the rest of the text (a serif p would not have a rounded top).
not to mention, it’s inane to imagine that a) the readers of this blog would not already know what a freaking p-value is or that b) the writer of PhD Comics would make such a stupid mistake.
sheesh.
/spleen vent

13. #13 someguy
May 21, 2009

Sure it looks like a rho, so obvious it’s been photoshopped.

14. #14 Richard
May 22, 2009

Alex, I am sad at the state of statistics education that you have had to think that’s a p-value.

The Greek character rho represents the population value of a Pearson’s correlation coefficient (r). Usually you don’t claim to have determined rho unless you did something to summarize across studies (like meta-analysis) where you have done some sort of correction for range restriction and/or attenuation of the predictor and/or criterion.

Also, @Sigmund, the lack of publication of non-statistically-significant results is one of the most problematic aspects of academic publishing today. Just because you didn’t find a relationship or difference doesn’t mean that your study is bad. It might just mean that there really isn’t a difference (or at the least, there is only a very small one).

15. #15 Richard
May 22, 2009

ALSO, your definition of p-value is just wrong. Even the Wikipedia page you cited gives you a different definition on the VERY FIRST LINE: “the p-value is the probability of obtaining a result at least as extreme as the one that was actually observed, assuming that the null hypothesis is true.”

If the comic said p=.56 (which it DIDN’T, but IF IT DID), it would be translated as “if we were to assume the population correlation is zero, there is a .56 probability that we would see a correlation of the observed size or larger in any particular randomly-obtained sample from that population.” – NOT “the correlation is 56% likely to be due to chance.” That is a horrible, uninformed, uneducated, and outright misleading interpretation.

/rant

(Sorry – I just HATE people spouting off their elementary understanding of statistics as if they know what they’re talking about, ESPECIALLY when they’re as condescending about it as alex just was.)

16. #16 says it all?
May 22, 2009