html xmlns="" xml:lang="en" lang="en"> From the archives: I was wrong about Kieran. I apologize.

Friday, July 20, 2007

I was wrong about Kieran. I apologize.

Kieran said...
Kieran’s tone rides a mean edge, and the comments on his posts are markedly meaner than others.

Not that one is able to remove the plank from one's own eye, but I'd disagree with that characterization. I'd say that your view is colored via a selection effect: the awful comments on that post about you and kids (the post itself was not at all mean), and the couple of posts I've written this week in sharp disagreement with someone else.
10:35 AM

I answered:

1. Seems like there is data enough that the hypothesis is testable. Some grad student could code the posts and comments for a year of Crooked Timber, whom we presume are a common pool, and see whether the effect I see is really there.

2. Do your co-bloggers get that same effect?
11:30 AM

After saying that Kieran was mean and getting that brutal response from him, I got to thinking that I would prove it to him. I would prove it to him using numbers and science, just like I wrote. After all, if you’re faced with numerical proof that you’re a big meanie, there isn’t much you can do but admit it and change your ways. Armed with a hypothesis (Kieran’s posts are meaner and elicit a meaner response) and some vague memories of my methods classes, I spent a few hours coding nearly three hundred Crooked Timber threads (Jan-April 2007) by number of ad hominem attacks. My dear readers, as a special treat, today we have data.

The thing that immediately jumps out at you, the startling outlier, is, of course, Kieran’s outrageously high score. That score is based entirely on the number of fights he incites as a blogger, the fights that are a reflection on his personal nature, the inner viciousness his blogging has inadvertently revealed to the world, spotlighted by my startling perceptiveness and willingness to Call It Like I See It.

Except that I was wrong. There is nothing remarkable about Kieran’s score. I knew that early in the project, because when I re-read Crooked Timber posts and comment threads as a block, it became obvious that Kieran’s posts are not mean, nor close to mean. His comments threads don’t stand out. Kieran’s explanation, that selection effect was coloring my memory of his posts, is right. In addition to being wrong about the nature of his posts, and wrong to impute that to his character, I was wrong to do all of that before going back to check his posts and comments, which were available to me. Kieran, I owe you an apology; I am sorry to have libeled you in the previous post. I take it back, of course, and will prominently correct that in the post itself. I owe my readers an apology as well; if any of you got a bad impression of Kieran from what I wrote, I am sorry that I led you wrong. My regret over this will be a sharp reminder to be very slow to accuse people, especially when I am emotionally involved. I am sorry, y’all.

P.S. In our email exchange where I apologized to him directly, that son of a bitch called me a bad blogger:
"you know, all this openness to data and willingness to recheck your assumptions and basic generosity --- are you sure you're really cut out to be a blogger?"


Blogger LizardBreath said...

That Excel chart is everything I love about geeks. "Disagree with me, will you? There's objective data that will settle this!" [Work work work, think think think] "And it establishes that you're absolutely right. Good thing you spoke up there." It's a lovely post to see on a day when I was disliking people.

3:12 PM  
Blogger Megan said...

DATA! Data will prove my point!! Except when my point is dead wrong.

3:15 PM  
Anonymous Francis said...

megan, you are awesome.

(except when you ignore me -- are you really moving to Oakland? why?)

3:40 PM  
Blogger Megan said...

Francis, cut that out. Serious.

3:44 PM  
Blogger justus said...

Not to disparage your efforts but your data doesn't address the statement Kieran quoted. There were two parts: "Kieran's tone rides a mean edge". I don't think "ad hominems" is a very good proxy for that. The second part: "the comments on his posts are markedly meaner than others" isn't even potentially addressed.

4:04 PM  
Blogger Megan said...

I knew we'd get to methods.

Before we discuss them, I want to make clear that my current qualitative assessment of his tone and the comments it evokes are that it is not mean. Reading several months of CT convinced me that it is the usual bemused and neutral blogger fare. With that out of the way:

I didn't code for tone of blogger at all. That was just too hard. I figured if the effect was there, it would be consistent for each blogger and be amplified in the comments and easier to catch.

I did think that # of ad hominem attacks was a reasonable (and quantitative!) proxy for meanness of comment thread. There were some that were hard to categorize; they had an overall snippy tone, but didn't actually call each other assholes. Those probably got underestimated.

So I'd say my second point: "the comments on his posts are markedly meaner than others", was roughly addressed (and disproved) and the connection to tone of blogger still up in the air.

But dudes, I'm not gonna do more looking. That's for some grad student to do.

4:13 PM  
Anonymous Anonymous said...

Why did you code threads with 1-4 ad hominem attacks as "1" and then add up the coded numbers? Wouldn't adding up the actual numbers of ad hominem attacks produce more precise data?

I'm not a very quantitative thinker, so I'm probably missing something here, but given that almost all the threads at which you looked had between 1 and 4 ad hominem attacks in them, wouldn't it have made sense to differentiate between the 1's and the 4's?

6:15 PM  
Blogger Kieran said...

Wait, so you're saying I'm not special?

6:34 PM  
Anonymous D said...

This? Is why you are always worth reading.

6:44 PM  
Blogger Dizzy said...

Hahahaha - ok, he may not be statistically more mean. But he sure can find the soft spots... That was quite the email response :)

9:21 PM  
Blogger Megan said...


I believe the DATA show us who the special person at Crooked Timber is. But I'm not messing with him.

10:11 PM  
Blogger Megan said...

This comment has been removed by the author.

10:11 PM  
Blogger jens said...

Oh....go for it, Megan.

Clearly the data for Montagu Norman proves that - since being "not mean" like Kieran is obviously not enough - the only way to avoid ad hominems in comments without pruning them manually, people being people, is to not have any comments at all!

2:21 PM  
Anonymous Anonymous said...

While this post is certainly evidence that Megan is a good engineer, we now know for a fact that she's a terrible economist. She clearly doesn't get the idea of cost/benefit analysis, for example.

She could have been making lattice-top pies with all the time spent doing that research.

6:13 PM  
Anonymous emir said...

That "special person" isn't reluctant to police comments, anyway ...

2:56 AM  
Blogger Megan said...

I saw that. He did an awesome job.

9:39 AM  
Anonymous belle waring said...

jesus, I've got to start making more ad hominem attacks.

7:03 AM  
Blogger Megan said...

Step it up, Belle.

I'm sure you can bring the best out of your commenters.

3:01 PM  

Post a Comment

<< Home