Pages

Thursday, June 24, 2010

Abortion mortality and the GIGO principle

I found this in an old email I'd sent out when venting, and thought it's not a bad idea to share it.

GIGO is a fairly simple concept: Garbage In, Garbage Out. If you have bad data, you can analyze it 'til the cows come home and you're still gonna have bad data. If my data on, say, what the average person spends on coffee a year comes from me looking at my own coffee budget, I'm not going to be giving Juan Valdez much to go on. I don't drink coffee. I'd rather suck an aspirin tablet. We can take the information on my coffee spending -- which would be maybe one cup a year purchased for a friend -- and we can extrapolate it to the entire country. We can see in which month I spent the most on coffee. We can look at my peak coffee expenditures, my low coffee expenditures, my average daily coffee expenditure, how much I spent on regular versus decaf, and it's not going to give you any useful information about how to market coffee, is it? Because it's not an accurate accounting of how much an average American spends on coffee. No amount of analyzing it is going to change that. You could hire ten thousand of the best statisticians and marketing experts, but if all the data they have is Christina Dunigan's expenditures on coffee, they're not going to know squat about the average American's coffee consumption.

And the same goes for abortion mortality data. We have to know where the data come from and how accurate it is or isn't before we spend any time analyzing it and drawing conclusions from it.

I spent six months of my life -- SIX MONTHS -- finding out exactly how the CDC gets their information about abortion deaths. I contacted every health department and vital records office in the United States -- the 50 states, New York City, DC, and the Virgin Islands (just to find out if territories worked differently from states). I spoke to the coding clerks that processed the death certificates, who analyzed them to see that they were filled out properly who abstracted data and sent it to the National Center for Health Statistics. I ordered database runs of state death records. I spoke to the people at the CDC who actually produce the annual Abortion Surveillance Reports. I got death certificates, autopsy reports, medical records, health inspection reports, medical board disciplinary documents. I talked to the guys at the NCHS. I spent SIX MONTHS finding out where the CDC gets their abortion mortality numbers.

What they do is very much akin to sticking a bushel basket under an apple tree, checking to see how many apples landed in it, and concluding that this is equal to the total number of apples that fell in the entire orchard. Note that it is NOT akin to counting those apples and then extrapolating the number of apples that fell in the entire orchard based on the size of the basket and the size of the orchard. No. They count what happens to land in their basket, then publish the numbers as if they really represent the total. And they defend the total on the grounds that sometimes people pick up apples off the ground and toss them into the basket, so they're not counting just the ones that fall straight down. This is a very clear and appropriate analogy to how the CDC collects their abortion mortality data. I know this because I spent SIX MONTHS LOOKING INTO IT. And the six months wasn't an arbitrary number. It's just how long it took of making phone calls and writing letters and waiting for data runs to come back so I could ask more questions. It's how long it took to get the information. It's how long it took to find out where the numbers come from. I can't stress this enough.

My troll keeps insisting that "there is an entire industry devoted to analyzing abortion deaths" and that therefore the numbers MUST be accurate. I agree that there is an entire industry devoted to analyzing abortion deaths.Where do they get the numbers they get tenure for analyzing?

FROM THE CDC. The guys with the bushel basket under a tree.

GIGO.

15 comments:

  1. On average, I have three cups of coffee each morning. I don't often have any beyond that as I cannot handle too much caffeine.
    Sometimes I have one in the afternoon from the nice man who has a stall at the fruit and vegetable market that we often go to on a Saturday afternoon.
    We have a coffee machine at home so we do drink very nice cups of coffee.
    The first one I make in the morning is usually from supermarket bought beans because my partner often gets up too late for that round. Plus the old taste buds aren't exactly fully tuned in so it really doesn't matter if it's not the best.
    The second and third rounds are made from a variety of flavored beans we buy direct from a coffee farm up on the tablelands about an hours drive from our tropical paradise.
    These can be flavors such as vanilla cream, irish cream, hazelnut cream, macadamia cream and more. We usually get a free cup of coffee when we go because we buy a reasonable amount of beans. We also buy some to send to friends interstate.
    Actually not far from the coffee farm there is a mango farm which produces mango wine, mango champagne and mango liqueurs. These are also very nice and we usually walk out with at least two bottles.

    So you have no real idea of the number of women whose cause of death is directly attributable to having an abortion?

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
  3. Oh Operation Counterstrike, you really need to get a life...

    Christina, Blessings 2 U for giving so much of your life for the least of these...You R AWESOME! Have a great weekend!

    ReplyDelete
  4. Oh, and Mary, this site seems to draw slightly more attention than a number of very similar sites I have observed. So the whole effort seems a little less than a positive return on investment don't you think?

    ReplyDelete
  5. Six months isn't a very long time to spend on a question like how to count abortions. The people at CDC who actually count them have PhDs. That's four-to-six YEARS of research with one-on-one expert guidance.

    ReplyDelete
  6. OC, they don't spend the entire time they're getting their PhDs doing abortion morbidity and mortality research. They get their PhDs in public health or whatever, THEN they go do their data massage on abortion numbers.

    And I've seen plenty of stupid research done by PhDs. One was about coffee use. They interviewed a bunch of old folks in nursing homes about their coffee use, and found that the coffee drinkers had no more or fewer health problems than the non coffee drinkers, and concluded that coffee consumption has no effect on health.

    Well, DUH! They weeded out of their sample anybody whose coffee use had caused health problems that had shortened their lives! This is like interviewing everybody who had just completed a 10K run about their history of gunshot wounds, and discovered that there was no statistically significant difference in race times between the people with a history of gunshot wounds versus those who had never suffered a gunshot wound. Well, DUH! You've weeded out from your sample everybody who suffered fatal or debilitating injury, purely by your study design!

    So just because research is done by a person with "PhD" after their name, or published in a peer review journal, doesn't mean it's solid research. A lot of unmitigated crap makes it through. That's why there is the saying: BS, MS, PhD means Bull Shit, More of the Same, Piled Higher and Deeper.

    I spent six months looking at how those PhDs COLLECTED the data that they then analyzed. The data collection methods were severely flawed. Therefore the analysis of the data needed to adjust for those flaws in data collection. They did not. In as obvious a way as the hypothetical gunshot wound study I just described. In as obvious a way as the very real coffee drinking study I described.

    You can get a PhD and still be an idiot. Richard Feynman vividly described this phenomenon in one of the chapters of "Surely You're Joking, Mr. Feynman!" And that's a guy with a PhD -- a VERY SMART guy with a PhD -- concurring with my assessment that you have to still look at the quality of the work, not just what initials somebody has after their name.

    ReplyDelete
  7. And while we're on the subject of what pure geniuses guys with PhDs are, remember it was guys with PhDs that decided to go ahead and launch the Challenger. Ka-BOOM! A decision that any shade tree mechanic who had dropped out of high school would have recognized as stupid.

    ReplyDelete
  8. No, the problems with the Challenger were not that simple, not that easy to detect. There was miscommunication between the executives and the engineers. It was a language issue.

    Yeah, I met Feynman once. He was very nice to me, didn't behave with any of his usual loonieness. (He was a real loonie.)

    Anyway, my point about the PhDs was just that six months isn't very long to spend on a question like this. You write as if SIX MONTHS makes you an authority. You're not an authority. The likelyhood is you didn't understand most of what you read and were told.

    ReplyDelete
  9. So Rupert if this site is so insignificant then why do you even bother responding?
    Me thinks you doth protest too much...

    ReplyDelete
  10. Well Counterstrike there are enough corpses to go around on this issue, women and babies...Start poking!

    And Rupert it's funny how many times in just your last comment you are telling folks how to live their lives...

    ReplyDelete
  11. I think you'll find I spend a lot more time trying to convince people to stop telling others how to live their lives Mary.

    Don't like abortion? Don't have one.

    Don't agree with sex before marriage? Don't do it.

    Don't agree with contraception? Don't use it.

    But you have no right to try to prevent others from doing so.

    ReplyDelete
  12. Rupert,

    Don't like people disagreeing with abortion? Don't visit these sites!!!

    ReplyDelete
  13. This biggest research flaw here is bias going in. You could have spent 16 years compiling data with a team of 100 of your twins and I would be skeptic of the results.

    It's nice that you care so much about a topic (I'm pregnant and happy about it, BTW) but your "science" is not legit. It is anecdotally interesting though!

    ReplyDelete
  14. OC, the problem with Challenger was VERY simple. People with an agenda overrode people with common sense. It happens all the time.

    Education is no guarantee that you're immune to ordinary human stupidity or evil.

    The reason you defend those clowns at the CDC is because you AGREE WITH THEM. You all worship the same god -- abortion. You have the same blind faith in them that Catholics have in the Pope. I looked at how they collected their data and it's GARBAGE. Their PhDs just give them the tools to take garbage and churn out PR that suits their political agenda. The fact that they're very educated and very good at manipulating bad data doesn't change the fact that the data are bad. And their PhDs don't make them less evil. They think KILLING BABIES is a good thing and they devote their lives to promoting it. And it's my job to expose this evil, and the lies behind it.

    ReplyDelete
  15. Velvet, which science is not legit? It would be useful if you were more specific.

    ReplyDelete