Array I can't find this reported before, and I am pretty sure it's NOT my data.
Array
Above, the relevant bits. Age is an integer variable taking on values in [0, 120].
At the extreme end of the distribution, there are 8 cases aged 111 or older.
NOW. if I plot a histogram of this:
hist age if age > 110, w(1) freq (i.e with unit bins), I get the following, where the 4 elements at 120 are plotted as if they are 119.
Array
SAY WHAT?
And this happens with other variables in the same data set: the final category gets (arbitrarily) combined with the PENULTIMATE category, even if one specifies w(1). Unfortunately data are sensitive, but here is a largely meaningful demo:
The below is a tabulate of frequencies of deaths by date of death in recent weeks for some population. Note the 111 at 24 and 25 jul; the 110 at 26 jul; and the most recent numbers of 94 and 62.
Array
Plotting a histogram of the same, and suddenly the data for 27 July reflect 156 (being 94 + 62)
Array
0 Response to Curious histogram glitch?
Post a Comment