September 5, 2011

Stat of the Week Competition: September 3-9 2011

Each week, we would like to invite readers of Stats Chat to submit nominations for our Stat of the Week competition and be in with the chance to win an iTunes voucher.

Here’s how it works:

  • Anyone may add a comment on this post to nominate their Stat of the Week candidate before midday Friday September 9 2011.
  • Statistics can be bad, exemplary or fascinating.
  • The statistic must be in the NZ media during the period of September 3-9 2011 inclusive.
  • Quote the statistic, when and where it was published and tell us why it should be our Stat of the Week.

Next Monday at midday we’ll announce the winner of this week’s Stat of the Week competition, and start a new one.

The fine print:

  • Judging will be conducted by the blog moderator in liaison with staff at the Department of Statistics, The University of Auckland.
  • The judges’ decision will be final.
  • The judges can decide not to award a prize if they do not believe a suitable statistic has been posted in the preceeding week.
  • Only the first nomination of any individual example of a statistic used in the NZ media will qualify for the competition.
  • Employees (other than student employees) of the Statistics department at the University of Auckland are not eligible to win.
  • The person posting the winning entry will receive a $20 iTunes voucher.
  • The blog moderator will contact the winner via their notified email address and advise the details of the $20 iTunes voucher to that same email address.
  • The competition will commence Monday 8 August 2011 and continue until cancellation is notified on the blog.

Nominations

  • avatar

    Statistic: “Horror Pacific Island education stats… Foreign Minister Murray McCully opened a Pacific conference today with some shocking statistics about the education failings of the Pacific region…. A million school-aged children around the Pacific do not go to school at all.

    He said around 40 per cent of school children in Pacific Island countries do not complete a basic primary education, and only 20 per cent graduate from secondary school.”
    Source: NZ Herald
    Date: 5 September

    Whenever I read something like this I immediately want to know the context for these “shocking” and “horror” statistics before making conclusions, i.e.:

    * How many school age children are there around the Pacific?
    * What were these figures historically?
    * Are the figures similar in all countries around the Pacific, or are there major differences?

    The problem of context is in the original speech but one would hope that the reporters would provide this for their readers.

    3 years ago

  • avatar

    Statistic: This time, I like the statistic but I don’t like the domestic commentary. So I’m nominating Jennie Connor’s trashing of the PLOS Medicine piece on alcohol consumption and health.
    Source: NZ Herald
    Date: 9 September 2011

    The PLoS Medicine article is here: http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.1001090

    The authors look at a set of British nurses in a longitudinal study. Correcting for age, smoking history, current smoking status, BMI, physical activity, education, husband’s education, marital status, postmenopausal hormone use, family history of various ailments, personal medical history, and dietary habits, they find that women consuming up to three standard drinks per day had better health outcomes than nondrinkers.

    Here’s the key part: the odds ratio for moderate drinkers (higher odds = good outcomes) was 1.43 before adjusting for all the health and other covariates and 1.35 afterwards, both statistically significant. There are a rather large set of health covariates in the full model, and correcting for all of them only knocked back the odds ratio by .08.

    Ok. So what does O’Connor say to trash the study?

    First, she cited her 2005 Commentary piece in the Lancet with Rod Jackson where they noted the potential for that associations between health outcomes and drinking are driven by unobserved health differences and that if the reference category of non-drinkers is contaminated by former drinkers, then results are overstated. Fair enough. But she doesn’t mention that those concerns were completely answered in the subsequent literature: in particular, di Castelnuovo and Donati (2006) and Rimm & Moats (2007). The first is a metastudy that carefully sorts those studies that conflate former and never drinkers; it finds the “J-curve” remains strong if we use the proper reference group, it’s just not as pronounced.

    Ok. Next, Connor says that unobserved lifestyle differences could be driving things. Sure. Possible. But if correcting for smoking, exercise, and diet only reduces the odds ratio from 1.43 to 1.35, there’s little chance that unobserved health correlates would drive that down to 1.0. AND, because the study’s authors were worried about residual confounding, they also ran things on within-strata subsamples like never-smokers. The results held up but statistical significance was lost with reduction in sample size. AND they excluded entirely anyone who could have been a “sick quitter”.

    So, Connor trashes the study for a bunch of reasons that the study’s authors did a decent job in controlling. But, do read it for yourself. I’ve given the link.

    For other background lit on the “J-Curve” in alcohol and health, I summarized things here a while back:
    http://offsettingbehaviour.blogspot.com/2010/03/moderate-drinking-and-health.html

    A fuller discussion is here: http://offsettingbehaviour.blogspot.com/2011/09/this-weeks-nomination.html

    3 years ago