[HPFGU-OTChatter] Re: Moore

[Shaun Hately]

On 4 Aug 2003 at 12:56, bluesqueak wrote:

> Shaun, I am really *not* getting your argument here. When you said 
> in an earlier post that Mr Moore was inaccurate about the number of 
> gun homicides in Australia, I was imagining that you meant he'd said 
> 165, when the real figure was 65, or some similar huge discrepancy. 
> The variation of uh, 1, or maybe, uh, 2 (which works out as an error 
> of from + or - 1.56% to + or - 3% ) would be reasonably acceptable 
> accuracy for me.

Two issues here - the first is that he is *not* aiming for reasonable 
accuracy. If he was just trying to be approximately accurate, he 
wouldn't be using number 11,127. He'd be using numbers like 11,000 or 
10,000. He's the one who has chosen to be precise. And he's inaccurate.

Secondly the figure of 64 gun deaths in Australia was for 1993. Bowling 
for Columbine's research is c1999. The figure for 1999 was 50. For 1998 
was 57, for 1997 79, for 1996 104, for 1995 67.

Basically he'd have to have gone back from 1999 to 1995 to even get a 
single year figure that could reasonably be rounded to 65. 
 
> > If he had said 'Around 11,000' cases, or 'Around 10,0000' cases or 
> > something similar to that, there wouldn't be an issue.
> > 
> > But he chose to use the number 11,127.
> > 
> > And that number cannot be substantiated.
> 
> And as you have argued above, the figures from two institutions 
> (which I assume are both official U.S. institutions - I wouldn't 
> know, being a Brit) vary by 1637 deaths. [Which I hope means that 
> the deaths of 1637 people turned out to be imaginary, but I rather 
> doubt it. I suspect their families are still grieving, but have been 
> told it was 'accident', rather than homicide.]

The reasons for the difference in figures is not exactly straightforward 
- but it has a lot to do with certain assumptions made about whether 
certain things are crimes or not. The BoJ figures are interested in 
trends - not in absolute numbers - and provided the measure used is 
consistent from year to year, the trends can be studied. The UCR figures 
are concerned with specific accuracy about the crimes committed.
 
> So, official figures vary by around + or - 16%. Mr Moore's figures 
> vary from the U.S. BoJ figures by again, about 1%. (Assuming it was 
> 1998 he was working from)
> 
> What you mean is that you can't find the data table he used. Or find 
> the official he got the figures from. The figures he quotes are 
> within the acceptable variation of the statistics; which do vary 
> according to how crimes are reported, whether a death is later re-
> classified as accidental rather than homicide, and so on.

No, what we mean is that NOBODY has been able to find the particular 
data file, Michael Moore used. Not just us - nobody out of quite a few 
people looking. And when Mike Moore has been asked for his source he has 
given only a vague reference to the UCR - which does not back his 
numbers at all.

So what we have are figures from a source nobody except Michael Moore 
seems to know about, and which he will not disclose. If this only *one* 
piece of data, that might not be a major issue - but he's made the same 
mistake with at least three different countries. That does not inspire 
confidence in his ability, or his desire to be factually accurate.
 
> For Australia, have you considered that he may have asked someone at 
> the Australian embassy, who told him 'it averages 65'? Which is 
> reasonable from the figures you quote. And would justify Mr Moore 
> refusing to reveal a source. 

If Mike Moore gathered his data by asking somebody at the Australian 
embassy for a figure, then that simply illustrates how careless he is 
with facts. That would not be a standard of research worthy of calling 
his final product a documentary.

It's certainly not good enough when contacting either the Australian 
Institute of Criminology, or the Australian Bureau of Statistics would 
have given him a fully accurate answer within 5 minutes.
 
> Or would you prefer to say that you can't find the table he used? 
> That 'his figures are inaccurate'? Because to say 'his figures are 
> inaccurate' when you're actually talking about a variation of 1 to 
> 3% is a bit of a statistical trick. It implies a large variation; 
> you are actually talking about a small one. It implies there is such 
> a thing as a definite, final figure, when in fact the figures could 
> easily vary by one or two depending on the way they were collected.

Mr Moore is the person who chose to use a figure of 11,127. Not 
approximately 11,000 - but a precise number. There is no data set as yet 
identified - by me, or a lot of other people - that supports that 
precise number. And Michael Moore refuses to say where he got it from 
except a vague reference to the UCR which has totally different numbers.

A 1 to 3% difference may not seem like a lot to you - but frankly, I 
expect a documentary worthy of being proclaimed the best documentary 
film in the world for a given year to be better than 97-99% accurate on 
basic census-type data.

> > And for a documentary maker who chooses to publically attack 
> > governments for presenting inaccurate information, while receiving 
> > an award for a film in which he does just that is, IMHO.          
> > Unacceptable.
> 
> When he's attacking a government whose figures between departments 
> vary by 16%, he could have a point. :-)

Does he have a point? Probably - but it's not one that is enhanced by 
sloppy use of basic statistics.

> Oh, yes they would, Shaun. You're doing exactly what you say 
> wouldn't happen. 'His figures are inaccurate!' you cry. 'Look; he 
> said 65 people were killed in Australia in one year! And you know 
> what? It WASN'T 65!'
> 
> No, it was 64. Gee whiz. That would be an acceptable variation in a 
> lot of scientific experiments.

No, it was 50 - or 57 - or 59 - or 104. To arrive at a number close to 
65, Mike Moore had to *ignore* the four years of more recent data easily 
available to him before he would get close to that number. Including 
ignoring the year of the Port Arthur Massacre - the worst case of murder 
by a single gunman in world history. 35 people killed in one afternoon 
(that incident is the single reason why the number for 1996 is so much 
higher than normal). He'd have to ignore those types of numbers before 
he'd even get close to a 65.
 
> But the headlines would be "Mike Moore admits figures were false!"  
> And how many people would read beyond the headlines?

So we should let him get away with fuzzy figures, just so he can avoid 
bad headlines? 

I don't know, I just have the crazy idea that we should expect 
documentary films to meet a basic standard of factual accuracy. If I 
used figures as sloppily as he does in my first year university course, 
I'd fail the assignment.

> But you seem to be missing the point. Variation of 1 to 2% or no 
> variation, there were approximately 10,000 to 11,000 deaths in the 
> U.S. by gun homicide. In Australia there were approximately 65. The 
> population of the U.S. is 250 million. The population of Australia 
> is 17 million (rounded up to nearest million, figures taken from 
> Collins World Atlas).

No, I think you're missing my point.

I happen to agree with Mike Moore basic contentions. I routinely argue 
in favour of gun control using Australia as a counterexample to America, 
using population comparisons etc. If you wanted, I could probably mail 
you a dozen posts I've made on the subject to a particular list.

I don't disagree with Michael Moore's contentions at all. I'm *not* 
missing his point. I just don't think that his point is enhanced by 
basic errors of fact.

I just disagree with the idea that Bowling for Columbine was a 
particularly good documentary film. I expect a documentary to get basic, 
easily verifiable, facts correct. Not just nearly correct - but correct 
and accurate. And I expect the maker of such a film to be prepared to 
cite his sources when people identify potential discrepancies.

I expect a documentary film to accurately quote people - not to splice 
in unrelated statements from different speeches made at different 
locations months apart in a manner that leads the viewer to believe they 
are seeing a single speech.

I expect a documentary film to accurately depict real life events - not 
to modify what really happened in order to create a more exciting story.

I agree with what Michael Moore is trying to do. I just disagree with 
his methods. He's careless with the truth - and that is putting the best 
spin on what he does that I can. 

> So if Australia had the same gun homicide rate as Australia, there 
> should be 680 gun deaths each year. If the U.S. had the same gun 
> homicide rate as Australia, they should have 956 deaths per year.
> 
> That is the point I think Mr Moore is trying to make. The point is 
> not arguing about whether his figures are accurate to within 0.0%, 
> or whether the table he took his figures from is accurate in itself. 

My point is that if Mr Moore wishes people to take his views seriously, 
he shouldn't be making basic mistakes that make it incredibly easy for 
his critics to attack his fundamental position.

I stand up and say the US has 48 states - are you likely to listen to 
anything else I have to say about America? If you can't get basic facts 
right, people will not listen to you.
 
> The point is that the variation between cultures is huge. The point 
> is that the question should be asked whether the gun is so important 
> to the U.S. sense of self that they decide they must live and die 
> with that homicide rate. 

Yes, and I agree that that is a question that is worth asking. My point 
is, however, that I'd rather that the question be asked in a context of 
reasonable accuracy - not an acceptance of inaccuracy.
 
> It's not my question - my culture has other questions to ask itself. 
> And the U.S. would be well within its rights and constitution to 
> decide 'yes, guns are that important to us'. 
> 
> But quibbling about the difference between 65 and 64 (or 67) does 
> not actually help to answer the question. The important difference 
> is the one that says ten times more people (per head of population) 
> die from gun crime in the U.S. than in Australia. The variations in 
> figures that you are arguing about don't actually change that at all.

Which I have said several times:

"Would it have harmed the 'documentary' for Mike Moore to have used
accurate figures"

"If these inaccuracies had been necessary to make his point, I could
understand them - but they weren't. They are simply clumsy."

"Thirdly, I do not believe that Bowling for Columbine is a film whose
value has been negated. Quite the contrary."

"But I personally think it would have been a better film if Mr Moore had
stuck to the facts, without embellishment. He could have still made his
point extremely successfully, and it would be a lot harder for people to
dismiss his work."

I have no doubt whatsoever about the basic value of 'Bowling for 
Columbine' as a film. I just have this whacky idea that a documentary 
film - especially ones that wins an Oscar - should actually be factually 
accurate.

If BFC hadn't won an Oscar, I wouldn't be worried about it at all - but 
I am *frankly* stunned at how low the standards must be if work that is 
this shoddy is acclaimed as the world's best for a given year. And I 
think it says a lot more about politics than respect for basic truth 
that that happened. 

Yours Without Wax, Dreadnought
Shaun Hately | www.alphalink.com.au/~drednort/thelab.html
(ISTJ)       | drednort at alphalink.com.au | ICQ: 6898200 
"You know the very powerful and the very stupid have one
thing in common. They don't alter their views to fit the 
facts. They alter the facts to fit the views. Which can be 
uncomfortable if you happen to be one of the facts that 
need altering." The Doctor - Doctor Who: The Face of Evil
Where am I: Frankston, Victoria, Australia