Home > Uncategorized > Pell Pill Punditry, cont.

Pell Pill Punditry, cont.

September 25, 2010 Leave a comment Go to comments

Cardinal George Pell sits somewhere between “irresistible” and “too easy” on the ridiculometer. Like many gay men and/or virgins of his advanced years, he is a silly and fussy and anachronistic person. When you consider his outlandish costumes and fetish for arcane rituals, Pell, like most of his priestly colleagues, is not an actual person at all: he is an elaborately manufactured persona. He is an avatar in a carefully crafted other-world: Vaticanland. Like Hogwarts or Narnia, Vaticanland is an brilliant fictional construction, with its own rules and shared history, its own myths and shibboleths, its own secret rituals. By grasping this, we can begin to understand how Vaticanland fanboys, in glorious regalia, can debate for days — with straight faces if not inclinations — how many angels can fit on the head of a pin, or how long a dead, unbaptised newborn must linger in purgatory.

Mocking Pell, therefore, requires all the imagination and effort of teasing Dungeons and Dragons fanatics or making fun of Trekkie conventions: tempting, for sure, but hardly satisfying.

But Pell will insist on engaging in the “public discourse”, as if he feels bound to speak out on behalf of the “Australian Catholics” he “represents”. This tips his delusion status to extreme.

Australian Catholics, like their counterparts in other country’s where literacy prevails, are no more aligned with Pell on matters of philosophy, morality and politics than I am. Even on “deal-breaker” issues like abortion and pre-marital sex, Catholics in the US, Western Europe and Australia are slightly less inclined to support Catholic teaching than non-Catholics. They are more politically liberal, less prone to fundamentalism, and more respectful of science and reason than the general population. A survey reported in July that 57% of of Latino Catholics in the US support same-sex marriage compared to just 22% of of Latino Protestants. This suggests that “mainstream” Catholics are quite capable of adjusting to the real world when the lights come up on the once-weekly screening of Vaticanland. Pell himself acknowledges the daylight between Catholic teaching and mainstream Catholic opinion, but does nothing whatsoever with the insight.

Pell has a new ghost-writer — I am guessing male, quite young and easy on the eyes — and he has found in Murdoch’s Australian a newspaper willing to publish his little fictional forays. Pell should thank god that god does not exist. After his craven little effort on Saturday, Pell would surely be destined for an eternity in that corner of hell reserved for the rhetorically dishonest.

I will leave it to well-credentialed feminists to demolish the central argument — that the contraceptive pill has actually deprived women of economic and political power, based on a single article in a Vaticanland fanzine of no apparent credibility — and will comment instead on an area where I am feel amply qualified: Pell’s fraudulent presentation of polling data.  Here is the key sentence:

A May CBS News poll of 591 adult Americans found that 59 per cent of men and 54 per cent of women believed the pill had made women’s lives better.

This poll result is clearly presented to support Pell’s argument that, despite what you may think, the pill has been bad for women — because, he uses this finding to assert, women themselves acknowledge this in greater numbers than men (thus the 5 percent gap between the sexes). This is brazen bullshit:

  1. The poll itself concedes a margin of error of 6 percent for the sample of men, and 5 percent for women. With this taken into account, the percentage of men who believe the pill has been good for women is anywhere between 53 and 65 percent and, for women, between 49 and 60 percent. This is the dictionary definition of a statistical tie. It is sad and revealing that Pell’s ghost-writer would cling so gamely to such insignificant findings.
  2. Pell’s ghostwriter has mysteriously reported only the percentage who support the Pill — odd, given his overall argument. When you examine the poll, it is soon clear why: only 6-8% of women, and 9 percent of men, believe that the pill has made women’s lives worse; the rest say it has made no difference.  By failing to account for the “no difference” crowd, this was a misleading attempt to present the findings as if the public are quite evenly split on support for the pill, when nothing could be further from the truth. The CBS poll (in full here) is exhibit A in the case against the Pell critique, and his audacious and selective use of the findings is staggering. This is the intellectual equivalent of very loose morals indeed.
  3. Other findings of the CBS Survey neglected by Pell’s ghostwriter include: 99 percent of women believe it has either improved or made no difference to their career prospects, and 84 percent believe it has either improved or made no difference to their family lives. Both these data points are utterly devastating to the Pell-ghostwriter thesis, and go unmentioned in favour of one statistically insignificant half-finding.

It is pointless to ask the shameless to hang their heads in shame — and this goes for Pell, his twinky ghost-writer and the the opinion editor at the Weekend Oz who made the decision to publish this nonsense without questioning it even slightly.

Categories: Uncategorized Tags: , ,
  1. Michael
    September 25, 2010 at 16:58

    It’s a good post, but that really isn’t what “margin of error” means. And referring to a “statistical tie” is never a good sign either. Statistical accuracy is important both ways.

    • PQ
      September 25, 2010 at 17:05

      Fair enough, although the point that the gap between men and women is statistically meaningless remains correct, right? The term “statistical tie” is a turn of phrase, not a statement of science. I agree but statistical accuracy is important both ways, so if you think my interpretation of the poll data from CBS is incorrect, feel free to clarify.

      • Michael
        September 25, 2010 at 17:51

        The poll definitely doesn’t support the hypothesis that the two rates are different with any reasonable degree of confidence, because the confidence interval of difference includes zero. You can’t figure that out just by adding on the claimed margins of error, though, and it’s misleading to put the figures that way.

        A statement like “the percentage of men who believe the pill has been good for women is anywhere between 53 and 65 percent” is just wrong. It could be anywhere outside that range as well, it’s just an unlikely event, but it’s also unlikely to be near either tail. The actual range in question is not the one you get with ±m.o.e. either. And the poll does suggest that the population rate for men is likely to be higher, but it isn’t able to state that with confidence. There is no such thing as a statistical tie.

      • PQ
        September 25, 2010 at 17:55

        Again, fair enough. I would argue until my dying breath the the phrase “statistical tie” exists in the English language, where it means an effective dead-heat. I think you are making the point that a statistical tie doesn’t exist in mathematics, but that isn’t my field. My interest was highlighting the rhetorical tricks Pell was using, which just happened to be in the area of misrepresenting poll data.

      • Michael
        September 25, 2010 at 18:11

        It isn’t an effective dead heat, that’s the point. It is still far more likely that the rate for men is higher than for women. It just isn’t 90%-confidence-interval sure. The misrepresentation is really in the flagrant lying-by-omission of point #2.

        But yes, statistical accuracy in the media is a personal bugbear, so I’ll leave it now. CBS’s account of the poll doesn’t make any sense either.

      • PQ
        September 25, 2010 at 18:22

        Michael, you seem determined to judge me by standards to which I would never dream to aspire. I am making the point that the difference between men and women is not as significant as it is presented partly because of the lack of certainty in the polling and partly because neither men nor women support Pell’s idea that the pill has been bad for women, so his use of this one data point at the exclusion of everything else is dishonest and sneaky. This is a rhetorical response to a rhetorical device. Applying a burden of statistical veracity on me seems a little churlish on your part since I never attempted to pretend the CBS poll proved anything more than what they said it did: that the vast majority of men AND women think the pill has had either a netural or positive effect on women’s careers, family life and general economic wellbeing. I think it is a little unfair to say that my argument “doesn’t make any sense”.

      • Michael
        September 25, 2010 at 18:31

        I wasn’t saying your account didn’t make sense, although I see that I did say that; I apologise. “…, either” was not meant as applying the preceding statement to what you’d written. CBS’s account conflates all kinds of things in ways that don’t make sense but probably make for good stories.

        My only real complaint is that the spread of that usage of “margin of error” is insidious and needs to be stamped out, because it’s misleading. And it’s particularly important when picking somebody else up for statistical flimflammery. Certainly your broader argument is correct, including all the other points you’ve raised.

    • PQ
      September 25, 2010 at 17:27

      I thought I knew what I margin of error was, and I have just revisited the subject in light of your comment. While I have not given a statistician’s overview of “margin of error”, I don’t think my discussion of how it applies in this case was inaccurate in any substantive way. I also happen to live with a professor of psychology who specialises in methodology and he is unperturbed by the way I put it. We are now both curious as to what you mean.

    • September 25, 2010 at 18:47

      I’m following the exchange with interest. Michael, it sounds like you are faulting PQ for not stating that the actual range of possible percentages is between 0 and 100, which while conceptually accurate is also a given and is not relevant to PQ’s point. Yes, the range of margin of error is conditional on some level of stated confidence. One can fault pollsters, I suppose, for rarely being explicit at what level of confidence.

      My quick calculations bear out Michael’s point that statistically, the two percentages are not different from each other (pretty confident at 95%), which ultimately supports the broader point that the op-ed author blatantly ignored error and inaccuracies of polling data.

      • Michael
        September 25, 2010 at 19:07

        No, I’m not suggesting that you need to mention the 0-100 range explicitly. The difference that the truncation makes is so vanishingly small at the distance from 54% that it doesn’t matter. For the error calculation to be meaningful it does need to be calculated from the actual figure in use, though, yes, and the degree of confidence is important but doesn’t really need to be mentioned given the link to the original polling data.

        What I’m saying is please not to reinforce the idea that you can just add and subtract “margins of error” willy-nilly and get something meaningful at the end. Pell has ignored the error but PQ’s calculation just doesn’t show that. I don’t want to detract from a good post though, and I think this really is at this point.

      • PQ
        September 25, 2010 at 23:47

        I appreciate your comments a lot. It’s great to get feedback, especially when it is so constructive. I will watch myself on methodology in future. Thanks a lot.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: