LOCKDOWN LUNACY, page-513

  1. 16,797 Posts.
    lightbulb Created with Sketch. 8267
    Boy, talk about flaky logic.

    I was fully aware that January and February deaths in Sweden were running below trend, but when one is talking about a data set which is deaths - particularly when the deaths in question are occurring at ages at, or in excess, of average life expectancy - then to identify real variations requires analysing the data over longer, not shorter, timeframes.

    This is to eliminate statistical anomalies - such as, in this case, “pent-up” deaths - and
    factors other than that which one is trying specifically observe.

    It stands to reason that, due to the very nature of the data being studied, if there is a period of “low” deaths, then it will inevitably be followed by a period of elevated deaths - the “pent-up data” phenomenon (unless, of course, life expectancy has somehow magically risen meaningfully and we know that isn’t possible).

    So you can’t selectively separate periods of low and high deaths in a cohort which is at, and in excess of life expectancy, because those periods are not mutually exclusive.
    The one informs the other.

    So to point out that the peak in deaths (March/April/May) was preceded buy a period of below average deaths makes my very argument for me, namely that a number of deaths which would normally have occurred over a certain time frame, say 12 months, have been concentrated in a certain sub-period, but when one looks at the overall, statistically more representative time period, the positive area under the curve where deaths exceed the average (March/April/May), is likely to be not materially different to the negative areas where deaths were below average (January/February and July/August).
    Last edited by madamswer: 27/08/20
 
arrow-down-2 Created with Sketch. arrow-down-2 Created with Sketch.