It is only because
People don’t UNDERSTAND statistics that is taken to be lies.
It is possible to use mathematics, and statistics is a branch
of mathematics, to lie, e.g., 1+1=3, but the lie is very obvious. It is only because people don’t understand
the terms in mathematics or statistics that they think they are lies. In the routine “Why a Duck”, it is only because
Chico Marx doesn’t understand the terms ford or a viaduct, etc. that it is a comedy. When Johnathan Swift called for the boiling
of Irish children, he was intending it as satire, not as a cookbook. When Archie Bunker sings “Those Were the Days”,
we are supposed to laugh at him. There is a
difference between tactics and strategy, between a battle and a war, and between
weather and climate. Just because they are similar does not mean they are the
same. If you think that there is no difference
between a standard deviation and standard error, then like Lou Costello, you
can expect someone like Bud Abbott to take advantage of you.
https://www.youtube.com/watch?v=9udNrOh5DyA.
Standard deviation is σ.
Standard error is σ divided by the square root of n. Yes, σ is a Greek letter. This is a because much of our mathematics was
developed by Greek philosophers. In fact,
Science used to be called Natural Philosophy to distinguish it from Spiritual Philosophy. If you know the terms in philosophy, then you
should also know the terms in science, mathematics, AND statistics.
N in the equation of standard error is the size, number of
observations, in a distribution. A distribution
is just another name for a group. The standard
deviation multiplied by itself is the variance.
A standard normal distribution, such as those in Nature, has a standard
deviation of, which also means that the variance is, 1. During the COVID
pandemic, the phrase “flatten the curve” became popular. To a statistician that phrase is identical to
“increase the variance”. Its opposite would
have been “steepen the curve”, “decrease the variance”. Decreasing the variance has problems of its
own. https://dbeagan.blogspot.com/2022/01/resiliency.html.
If the goal is to eliminate error, the response should be
to INCREASE the size of the distribution, group. It should not be to set the deviation to
zero. If the deviation is zero, then any
value of n, including decreasing the size of the group, is possible. Thus, claiming that to reduce error, deviation
must be zero is absolutely compatible with exclusion. However, IT IS NOT NORMAL. Not only aren’t statistics lies, but we would
also be better off if our politicians understood statistics.
No comments:
Post a Comment