Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Debunking Princeton

facebook.com

Post reply on HN
Print view View original

61–70 of 150 posts

Re: Debunking Princeton

#61

Post by ajju »

yuvadam wrote: post #57
ajju wrote: post #42

Earlier quoted context omitted.

You mean the actual metrics linked to in the 2nd sentence of the post? https://www.facebook.com/photo.php?fbid=10101235005815241&se... (Oh, you missed that? Is that because you didn't actually read the post?)

Those look like numbers from Facebook's SEC filings, which says all there is to say about which metrics they're choosing to publish. (Read the post, didn't miss it.)

Not sure what you expect them to publish beyond MAUs and DAUs by region. Those are pretty standard ways to measure engagement and if I were Facebook I'd be loathe to publish more granular data just to "quiet the critics".

Re: Debunking Princeton

#62

Post by STRiDEX »

hkmurakami wrote: post #51
hooande wrote: post #13

I hope this becomes a bigger story than the original. I can't tell you how many times someone has thrown a junk science article in my face, thinking that the issue in question was now settled. A survey with a small sample size, outsized extrapolations and numbers that don't match the accompanying conjecture. There should be consequences for the people who publish these things. People have a tendency to believe anythi…

The worst part for me is that the two authors are Ph.D. candidates in MAE (Mechanical and Aerospace Engineering) [1]. They aren't even tenure track faculty at the school and aren't really in a position to carry the name of the school on their shoulders. Yet from what it seems, it will be the school itself that will be hit with most of the public shaming burden (ex: the FB post's headline). As a Princeton alum, I'm a…

The action of the two PH.D. candidates. The action of few employees. Yet they both represent the group to which they belong.

Re: Debunking Princeton

#63

Post by poopsintub »

jakewalker wrote: post #15
poopsintub wrote: post #10

Earlier quoted context omitted.

My investment account seems to disagree too. The stock was down more than $1 before the market even opened based on that sham study by Princeton.

Why are you so sure that was the cause of the price decline?

The other big news was that Facebook was testing ads outside of their own platform. The possibility of taking on Google's Adwords and adding more than a few billion dollars in revenue should have increased the price?

Re: Debunking Princeton

#64

Post by lotso »

ajju wrote: post #42
yuvadam wrote: post #36

Hey, Facebook, wanna prove Princeton wrong? How about publishing actual metrics proving that user engagement is at an all-time high? (Oh, you don't have metrics to prove that? Is that because users are jumping ship?)

You mean the actual metrics linked to in the 2nd sentence of the post? https://www.facebook.com/photo.php?fbid=10101235005815241&se... (Oh, you missed that? Is that because you didn't actually read the post?)

I'm not sure DAU is the best stat for looking at engagement.

Re: Debunking Princeton

#65

Post by chaz »

devindotcom wrote: post #39

Seems like Facebook is responding to the media interpretation of the Princeton paper rather than the paper itself. I saw no problems with the paper: it applied an epidemiologically-inspired statistical technique to Facebook using a defunct precursor as exemplar. The limitations of this technique are obvious, but it's an interesting idea. People do "research" like this all the time - you throw shit at the wall and see…

The problem is less about the technique, but about the flawed assumptions about the data that they used, and how they assumed that it was a valid basis for a paper. That undermines the whole thing, no matter the rigor of the analysis.

The medium matters, too. I'm all for people discussing crazy ideas and seeing what sticks, but when they wrote a paper about it and submitted it to arXiv.org ("Submissions to arXiv should conform to Cornell University academic standards."), they staked their reputation. Why wasn't this just a blog post?

Re: Debunking Princeton

#66

Post by minimaxir »

S4M wrote: post #23

Call yourself "Data Scientist" and publish graphics without units. Congrats!

The exact values for the axes are proprietary to Facebook. (in fact, in ggplot2 which he used for the first 2 graphs, you explicitly have to tell it to not show the units on the axes, and it's very unintuitive)

Re: Debunking Princeton

#68

Post by 1stop »

rogerbinns wrote: post #59
devindotcom wrote: post #39

Seems like Facebook is responding to the media interpretation of the Princeton paper rather than the paper itself. I saw no problems with the paper: it applied an epidemiologically-inspired statistical technique to Facebook using a defunct precursor as exemplar. The limitations of this technique are obvious, but it's an interesting idea. People do "research" like this all the time - you throw shit at the wall and see…

Alex St John has good details on why using existing epidemiologically formulas don't work: http://www.alexstjohn.com/WP/2014/01/23/forecasting-demise-f... Example: "The amount of time you spend sick with the flu is constant no matter how many of your friends ALSO contract the flu"

Not if they contract a different flu... then there is a higher chance you'll get that too, and be sick longer.

Re: Debunking Princeton

#69

Post by sabbatic13 »

STRiDEX wrote: post #62
hkmurakami wrote: post #51

Earlier quoted context omitted.

The worst part for me is that the two authors are Ph.D. candidates in MAE (Mechanical and Aerospace Engineering) [1]. They aren't even tenure track faculty at the school and aren't really in a position to carry the name of the school on their shoulders. Yet from what it seems, it will be the school itself that will be hit with most of the public shaming burden (ex: the FB post's headline). As a Princeton alum, I'm a…

The action of the two PH.D. candidates. The action of few employees. Yet they both represent the group to which they belong.

Bullshit.

4500 undergrads, 1500 graduate students, and hundreds of faculty, and you think that every little paper they give or study they conduct is a product of the whole in a meaningful sense?

The study was practically just a thought experiment by two grad students looking at something outside of their expertise. It _only_ got any attention, because it was about Facebook, and so the media saw a chance to sell more ads. For the same reason, they used the metaphor of pars pro toto as a rhetorical ploy, because Princeton vs. Facebook sells more ads. Like many among the unthinking masses, you got punked by it.

Re: Debunking Princeton

#70

Post by scholia »

abus wrote: post #60

I see this as Facebook's willingness to use its unfair influence to discredit anyone who dares challenge it. In other words, "if you have a page here, we won't hesitate to use it against you".

Facebook takes a lot of flack: it's today's standard whipping boy. In this case, it did respond to a fundamentally flawed research paper, but I doubt it would have bothered if that paper hadn't been so widely and so shoddily reported.

I though it was quite generous of Facebook to resort to gentle mockery rather than shred its numerous stupidities.

Post reply on HN