« Ahmadinejad's rough week | Main | 300 »

March 11, 2007

Comments

300

Dammit!

Jesurgislac and dr ngo: Your comments beggar belief. Let me refresh your memory:

1. I make a comment pointing out several problems with the study, only one of which is the fact that it spuriously claims to have shown anything about whether Bush is "first."

2. Hilzoy responds only on the "first" point (wrongly suggesting that the study didn't make any such claim), and not at all on the other points.

3. Jesurgislac responds to me by pointing out that the real question isn't whether Bush was "first."

4. I specifically responded to Jesurgislac on this point. I said: "I've pointed out several other criticisms of the study; I haven't limited myself to the criticism that the study unjustifiably claims that Bush is the 'first.'"

5. Pat also points out the same thing: "Hilzoy, in my post at Stubborn Facts, I deliberately left the criticism of the 'study' cited by Krugman about the 'first administration' claim to last. It's the most minor point."

6. I point out yet again that Hilzoy quite unjustifiably assumed that the "only flaw alleged with the study is whether it claimed that the Bush administration was 'first' to engage in political witchhunts."


By now, even the most inattentive reader might have realized that we critics of the "study" have been complaining about a whole boatload of things other than the issue of whether Bush was "first."

Yet even after all of that, Jesurgislac parachutes in and pretends yet again that the only criticism at issue is whether the study properly claimed that Bush was the "first.":

None of these new commenters have paid attention to the refutation, but continue to buzz about Bush possibly not being the first, as if that mattered.

Nobody has "continued to buzz" about this issue, except insofar as I pointed out that it is a telling example of either stupidity or dishonesty (take your pick).

Yet Jesurgislac claims that "bad faith" on my part has been "proved"! She is the one who ignores other people's posts, pretends that their only criticism is one that they themselves repeatedly describe as "minor," and refuses to acknowledge the existence of their more significant criticisms. And we're the ones arguing in bad faith? What

What amazing chutzpah, that should have said.

John Doe | March 13, 2007 at 12:59 AM : it's bullshit to put out a study claiming that the Bush administration is the "first" to do something when the study didn't even look at anything pre-Bush

John Doe | March 13, 2007 at 12:13 PM: It apparently bears repeating: You can't know whether something is "first" unless you look at what came before. You can't know whether Tuesday is hotter than Monday if you only know Tuesday's temperature. You can't know whether Los Angeles has more people than New York if you only know the population of Los Angeles. Is there anyone who still doesn't grasp this basic point?

John Doe | March 13, 2007 at 01:06 PM: I haven't limited myself to the criticism that the study unjustifiably claims that Bush is the "first" (although this should be a good clue to the discerning reader that mischief is afoot, as it proves that the study's authors are either too stupid to know what the word "first" means or too dishonest to care).

John Doe | March 13, 2007 at 05:10 PM: The authors of the study are clearly dishonest -- as shown by their claim to have proven that the Bush administration is the "first" to do [something] even without looking at any pre-Bush evidence.

John Doe | March 13, 2007 at 08:22 PM: Well, we've already established beyond a shadow of a doubt that these particular professors are willing to lie about their results (i.e., by claiming that they've proved that Bush is the "first" to politicize prosecutions, when they didn't even study anything pre-Bush).

John Doe | March 13, 2007 at 11:39 PM : I'm one of the few people who seems to have actually looked at the study (which is why I don't make embarrassing errors like claiming that the study had a "random" sample, or that it did not claim anything about Bush having been the "first" to engage in political prosecutions).

John Doe | March 14, 2007 at 09:02 AM to me: There's been a long discussion of points other than the one you seem hung up on, i.e., whether Bush was "first." Of all the criticisms that can be made of the so-called "study," that doesn't even make the top 10. Why are you still pretending that it's the only criticism that's been made?

John Doe | March 14, 2007 at 01:40 PM: Nobody has "continued to buzz" about this issue, except insofar as I pointed out that it is a telling example of either stupidity or dishonesty (take your pick).

And we're the ones arguing in bad faith?

Yes.

John Doe | March 13, 2007 at 12:59 AM : it's bullshit to put out a study claiming that the Bush administration is the "first" to do something when the study didn't even look at anything pre-Bush

John Doe | March 13, 2007 at 12:13 PM: It apparently bears repeating: You can't know whether something is "first" unless you look at what came before. You can't know whether Tuesday is hotter than Monday if you only know Tuesday's temperature. You can't know whether Los Angeles has more people than New York if you only know the population of Los Angeles. Is there anyone who still doesn't grasp this basic point?

John Doe | March 13, 2007 at 01:06 PM: I haven't limited myself to the criticism that the study unjustifiably claims that Bush is the "first" (although this should be a good clue to the discerning reader that mischief is afoot, as it proves that the study's authors are either too stupid to know what the word "first" means or too dishonest to care).

John Doe | March 13, 2007 at 05:10 PM: The authors of the study are clearly dishonest -- as shown by their claim to have proven that the Bush administration is the "first" to do [something] even without looking at any pre-Bush evidence.

John Doe | March 13, 2007 at 08:22 PM: Well, we've already established beyond a shadow of a doubt that these particular professors are willing to lie about their results (i.e., by claiming that they've proved that Bush is the "first" to politicize prosecutions, when they didn't even study anything pre-Bush).

John Doe | March 13, 2007 at 11:39 PM : I'm one of the few people who seems to have actually looked at the study (which is why I don't make embarrassing errors like claiming that the study had a "random" sample, or that it did not claim anything about Bush having been the "first" to engage in political prosecutions).

John Doe | March 14, 2007 at 09:02 AM to me: There's been a long discussion of points other than the one you seem hung up on, i.e., whether Bush was "first." Of all the criticisms that can be made of the so-called "study," that doesn't even make the top 10. Why are you still pretending that it's the only criticism that's been made?

John Doe | March 14, 2007 at 01:40 PM: Nobody has "continued to buzz" about this issue, except insofar as I pointed out that it is a telling example of either stupidity or dishonesty (take your pick).

And we're the ones arguing in bad faith?

Yes.

This comment to apologize for the double post gives me the opportunity to note that John Doe wins the amazing chutzpah award of the night (though not the mensch award of the day) and to add that I am glad to be off to Belgium on holiday tomorrow. With any luck, once I'm back online on my usual basis, this crowd of wingnuts will have drifted away again.

Also, maybe Bush will have been impeached. I'll hope for it.

Anyone want chocolate? Or would they rather have beer? Belgian chocolate and Belgian beer are two of the good things of this earth.

Apologies for the double post.

Belgian Beer! Hmmm Goulden Carolus!

Oddly enough, you don't seem to have read the posts that you quoted. Viz:

There's been a long discussion of points other than the one you seem hung up on, i.e., whether Bush was "first." Of all the criticisms that can be made of the so-called "study," that doesn't even make the top 10. Why are you still pretending that it's the only criticism that's been made?

Indeed.

And this is exactly correct:

"Nobody has 'continued to buzz' about this issue, except insofar as I pointed out that it is a telling example of either stupidity or dishonesty (take your pick)."

Neither you nor Hilzoy nor anyone else has a good answer to this.

To the contrary, the only point that you and Hilzoy have made is that even if the study wrongly claims that Bush is the "first," that doesn't (in and of itself) disprove the rest of the study, nor does it prove anything about the attorney firings.

That's all fine and dandy. I don't disagree with that.

But at the same time, the fact that the authors are willing to make a partisan accusation that can be disproved just by reading the rest of the page doesn't speak well for the reliability of the study. If they are so untrustworthy as to easily-checkable matters, why should anyone trust their research on points that are less easily checked? Any rational person would be suspicious of a study if the author announced, "Based on my study that focuses exclusively on Ohio, I think Ohio has a higher rate of [whatever] than Oregon."

Not only have you not "refuted" this point, you haven't even tried.

Jes,

Belgian beer for me, too. I prefer the fruit lambics.

So now the hair-splitting is not about the study, but who said what about the study. But I think it's about over, although I thought that this morning.

Beer for me. Thanks. (Although I'm lucky enough to live outside of Philadelphia, which I know has quite a few Belgian bars. I've heard it's about the best city for those who love Belgian beer, but I don't know that to be true. I did a quick study that showed Philadelphia had 7 times as many Belgian bars as the city with the next highest total.)

hairshirthedonist,

Philly also had the late and lamented Cuvee Notredame, which had wonderful Belgian cuisine.

'Tis kind of amusing to watch some make a big deal of something, have it debunked, and then whine that someone paid enough attention to debunk it and therefore the debunker must be saying something else.

I have addressed only the bogus study, which is Crap with a capital C. If anything claimed in the study is for real, it's not because the study showed it to be so.

And for the record, I've been reading here for quite a long time, I just don't usually bother to comment unless it's to address total BS being touted as reality.

Yeah, John, some folks think that "refutation" consists of saying "IS NOT!" repeatedly. Pfui. Another reason I don't comment here much--arguing with True Believers of either wing bores the hell out of me (which must make me a saint by now, having had all that hell bored right out).

If you're ordering beer, you really need a fridge that launches.

Darn: I saw that Gary had commented on this thread, and thought: oh. my. god. Perfect storm. Unstoppable cannonball, immovable object. Gary vs. the study fiends!!!

What a disappointment ;)

Saw that, Gary. I'm resisting as it would put me into permanent sofa spud-dom...or maybe I just wanna build my own with a better distance-guaging remote....

"What a disappointment ;)"

I'm more in a sit back and have a beer, and some popcorn, mood, I'm afraid.

Besides, the Belgians really do make some excellent beers.

My life is otherwise much happier by ignoring the whole this-study debate; there's no shortage of actual news, and analysis.

Besides, I'm contemplating whether I should write a long long post reviewing the very first season of Mission Impossible.

You know, important stuff.

Belgian Beer! Hmmm Goulden Carolus!

Lol, my spouse spent years complaining how hard it is to find it in the Netherlands (though in Dutch you should write 'gouden') - only to find that our little liquershop at the corner of our square has a room filled with special beers like the Gouden Carolus ;)

I go to Brussels regularly, my MIL lives there, but I cherish the Peking Duck above all else. Since the Marks and Sparks retreated with their food section there is NO Chinese in the Netherlands where you can eat it properly :(

This is pretty far afield, glasnost, but what's your reason for thinking that in a 5-year period where around 6,000 officials were actually prosecuted by the federal government, the number of investigated officials who were elected (rather than appointed) is anywhere near 375?

I've said this more than once now, John, and Pat brought it up himself.

Your figures from the DoJ report include both elected and unelected officials. The study in question is explicitly limited to elected officials. Your figure of "6000" includes a larger set of officials than is being investigated by the study. We don't know how many elected officials were indicted by DoJ.

Not that more is needed, but on top of that,
you yourself admitted in your first post that it's not clear if all those figures stand for new investigations.

If you cut out federal and state officials, and look at local ones, which is where the 7:1 ratio comes from, you'll see that there were only about 1400 indictments of both elected and unelected officials during the Bush Admin to date.

I don't know that 375 investigations represents all or most of the investigations of local officials under Bush DoJ. It simply seemed plausible, and I had no reason to discount it. Your contrary numbers do not clearly enough represent what we're talking about to convince me that it is implausible now.

I'll have to concede you whatever further points you wish to make.

Pat, a final point.

You seem prepared to admit that the extent of the study may well be flawed, that the number is not 7 to 1 as claimed, but instead wish to reclassify its basic point as being that there is some bias. Well, that's not the claim the authors make. They claim it's 7 to 1, not "some." Their conclusion is defeated if the ratio is only 6 to 1, or 5 to 1. Your conclusion of bias may or may not be correct, but it's not made by the study. You've asked me to disprove not the study but your own conclusion of some bias.

Of course, without knowing the methodology, I can't rule out with certainty that some bias factors - in the sample selection could possibly.

My problem with your whole point of view and PoV is not that you brought up ways in which the sample could possibly be flawed, but that you prevented your concerns as definitive. You treated and (and, to some extent, continue to treat) the study as certainly flawed, and you don't know that it is so.

To be more specific, your link on Instapundit leads hundreds of thousands of people to willfully and without cause blow the concept of bias off completely. A study showing 7-1 bias? Oh, wait. Here's some speculation as to some things that might be wrong. Ergo, the authors are quacks and the study is baseless. Ergo, nothing to see here.

The Lancet is a picture-perfect example. Anyone with half a brain knows that the official Iraqi government counts were way understated, but trashing the cluster selection method of a survey suggesting 600K deaths lets people pretend that the Iraqi government's pulled-out-of-thin-air numbers are the real thing, when they're really off by orders of magnitude, and unscientific in more wildly obvious way than the point of comparison.

I don't know beyond certainty that the study 's methodology isn't flawed. But you don't know that it is flawed. And you act like you do. We both think it's very likely that there is relevant, correct information from the study as follows - in one form or another, that the Bush Admin is probably investigating more local democrats than republicans. No one here thinks any biases sort of outright falsification could bring 7:1 down to something close to even. The most important finding of the study - and I'm paraphrasing their specific statement - was that the Admin investigates more Dem than Repub local officials, to a statistically significant extent. 7 to 1 is a more specific accounting of that fundamental point.

You lead people away from that fundamental point. Findings that extreme are very unlikely to be entirely the result of accidental bias. That's my belief. I'm sticking to it.

Glasnost, my final point, responding to yours.

As to misleading people, Paul Krugman's column is read by many more people than my little blog. For that matter, Obsidian Wings and this post by Hilzoy have probably been read by far more people than read my post. Both Krugman and Hilzoy used this study as conclusive proof of a massive bias against Democrats. I felt compelled to shed a little light on that subject, and I didn't notice anybody who found the report supportive of their own political positions being overly worried about misleading the public with shoddy "scholarship." Krguman and others seemed only to happy to try to bolster their claims against the Bush Administration with a little academic authority.

As to the study itself, I do not by any stretch concede what you call its "fundamental point," which is a much lesser claim than that made by its authors. You are wrong when you say "no one here things any biases sort of outright falsification could bring 7:1 down to something close to even." I do think that. I said that I felt that 375 was in the right order of magnitude, meaning that I agree it is more than 100 and less than 1000. There's plenty of room in that less than 1000 to either find a couple hundred overlooked Republicans (the study found no "investigations" of elected officials in 8 states, something I find exceedingly unlikely). In another hour of work yesterday, I found 3 or 4 more Republican elected officials who were investigated but not included in the study, bringing the total up to 11, with just 2 hours work. I think there's also a reasonable possibility that scrutinizing the 375 cases listed would remove some of the Democrats from the list, too. I don't think it's accurate to say that the entire Baltimore City Council was "under investigation" simply because they all received a subpoena for some of their records.

As for the definitive nature of my criticism, it is warranted. My suspicion about their "sampling" method was confirmed by the Philadelphia Inquirer reporter who e-mailed Shields. Beyond that, I will reiterate that the burden of establishing methodology is on the persons conducting it. Had they disclosed the methodology, then it might be my burden to show how any flaws affected it. But they've put slanderous accusations out there, claiming to be scientific about it, with no supporting basis at all. The report proves absolutely nothing, and it's entirely appropriate of me to say so.

Sorry about that. Forgot to close a tag. Should've used preview.

It bears repeating: Both Paul Krugman and Hilzoy should be ashamed to have relied on such a shoddy study. Neither one would cite this kind of unsubstantiated and non-scientific baloney in their own academic work (at least I hope not). Why is it good enough to use here?

I don't see anything in that link to remotely substantiate the claim that it's a "shoddy" study. All I see is a lot of rhetorical questions which don't come close to explaining away the huge mathematical disparity. In fact, most of the column simply serves to suggest that the authors of the study could have examined a different question instead.

Shorter Michael Smerconish: "But-but-but they're Democrats. Did you hear me? DEMOCRATS!!1 'Nuff said."

(Earnestly quoting low-rent O'Reilly wannabes counts as a Fonzie moment in my books. Hey, look - Dan Rather dry-humping Jamil Hussein!)

Mattt,

While I am not going to dispute that characterization of Mr. Smerconish in general, is there anything in particular you have in mind?

Did I cite it as conclusive proof? No. Did I rely on it? No. Do I think it's shoddy? Not enough to call its conclusions into question.

To repeat myself: if whoever was guest posting at Instapundit chose to characterize my post as being "about" the study, that's his business. But the post was not about the study. It did not rely on the study. The study is not the point. It is ancillary to an ancillary point.

To clarify, my 'shorter...' referred to the portion of the linked column Smerconish inquires about the political leanings of one of the study's authors; the parenthetical 'Fonzie' aside was me cryptically stating my opinion (via hypertext) that the thread had jumped the shark; the Hussein/Rather coital encounter was (again) a cryptic (if not crass) reference to Rathergate, Jamilgate, and other examples of instances where some on the right believe that dissecting a red herring counts as debunking.

I'm sorry you (apparently) feel my post not sufficiently wonkish, DTM. Next time I'll try to include more footnotes (and perhaps a Foucault quote).

And now, I bid adieu.

All I see is a lot of rhetorical questions which don't come close to explaining away the huge mathematical disparity.

Sigh. The authors apparently went out and did some Google searches and then threw together a chart of all the relevant results (or a cherry-picked selection of the results? Who knows). This is an absolutely shoddy methodology. Which means that there may NOT BE any "huge mathematical disparity" in the first place.

Why is it a shoddy methodology? Well, basics first:

1. Not all newspapers put their stories online in the first place. Particularly small-town newspapers.

2. Even when newspapers do put their stories online, it's very rare to find news stories from 2001, 2002, 2003, etc., that are still available. By far, most news stories expire after a limited period of time.

3. Both 1 and 2 mean that the authors would have missed many news stories in any search of Google.

4. Beyond that, the authors' reported searches turn up 126,000 results ("public corruption" and "elected") and 223,000 results ("federal grand jury" and elected) respectively. How the heck did the authors winnow through these results, most of which are irrelevant? More importantly, how in the world would you come up with a non-bogus random sample from these search results?

5. Without a random sample (and so far no one has come up with ANY reason to believe that the sample is random), the study is meaningless. If you don't understand this point, please return to Statistics 101.

Mattt,

The link is to his biography.

*Sigh*

John Doe alreadylinked to the Smerconish column I was referring to.

The bio was included because I assumed a lot of non-Philly folks (like myself) wouldn't know Smerconish from Adam (eg, that he was an O'Reilly fill-in and Bernie Goldberg-approved, thus putting his objectivity into question.)

Apologies for the confusion.

Do I think it's shoddy? Not enough to call its conclusions into question.

This is unbelievable. Cite for me any legitimate social science study that:

1) does not discuss its methodology at all;

2) does not even claim that its sample was random (outside of ethnographies that are admittedly focused on a single institution and don't make claims about the nationwide proportion of something);

3) is apparently based on Google results that, as everyone knows, don't give full coverage and cannot be used to create a random sample; and

4) doesn't even bother to mention the possibility of controlling for other factors (which here would include greater Democratic dominance of large urban cities where the incentives for corruption are greater and the frequency of news coverage is also greater).

I'm pretty sure that you'll never find any such thing. Compared to any serious social science study, this looks like the equivalent of the Sokal hoax.

But the post was not about the study. It did not rely on the study. The study is not the point. It is ancillary to an ancillary point.

Well, this double ancillary point struck me as quite a bit more serious than anything that has been said about the 8 fired USAs. I guess I have to repeat myself:

1. 8 fired USAs -- troubling, at least in a few of the cases, but not obviously troubling in a few other cases, and in any event, very limited in scope.

2. 6 years of politicized prosecutions on a nationwide scope, either targetting innocent Democrats or failing to prosecute guilty Republicans -- a much more significant problem, IF it's true.

So Mattt, do I have this right:

1. Michael Smerconish has filled in for Bill O'Reilly on his radio show.

2. Therefore, when Smerconish reports that the study's authors told him the specifics of their Google searches, he's lying. And with amazing brazenness, too, given that the authors could easily point out that they did a different Google search or that they didn't rely on Google at all.

3. Therefore . . . what? What are you trying to suggest? That the methodology is fine and peachy?

Reply to: Anarch's comment| March 13, 2007 at 12:51 PM

You are right on about Mayor Xochilt Ruvalcaba not being a target of Bush's DOJ investigation

1. The investigation was started by Cooley, LA District Atty prompted by the DEMOCRATIC LATINO CACUS as political retaliation for her opposition to their money cow "The Neava Azela Power Plant" - check the facts LA Weekly Sarah Catania exposes how Martha Escutia, Marco Firebaugh and the Dem Latino Cacus received large campaign contributions from The Power Plant's owners SunLaw. In addition Martha's husband Leo Briones was the project manager of the proposed power plant who btw made in excess of $150,000 for one year of service!

2. Xochilt Ruvalcaba was investigated throughly and came out crystal clean!

3. The checks that were signed before leaving office were for fees already incurred by attorneys and were due. Her council was being sued left and right and they had a responsibility to defend the city from frivolous lawsuits!

4. The employees that were laid off were student workers and community services department (established community clean ups, free tutoring and assistance in reporting potholes and sidewalk repairs, etc.)established by the Mayor Ruvalcaba. therfore the new council saw those employees as former allies of the outgoing council.

5. Has anyone driven by that city?!?! It is trashy! Mayor Ruvalcaba whas been out since 2003. you want to tell me that in 4 years the new council cannot manage to clean and repair public property!

You are so right it wasn't Bush - BULLSHIT is definately bipartisian!!!!

F%$k the Dems as well!!!

The comments to this entry are closed.