Excel is dangerous when people think they know what percentages are; or they think they can do "statistics".
I was involved in some work around English health care and "dignity and respect".
The process was horribly flawed. Volunteers asked patients some questions off a questionnaire. The answers were sent back, and reports were written.
Obviously, the sample sizes were not random; they were tiny; etc etc.
And the people writing the reports didn't get any help from statisticians. So you got comments like "28.69% of people said they felt [...]". I pointed out this obvious glaring error. They were grateful, and said of course they percentages should have been rounded to the nearest 0.25%. (This was with a sample size of 16 people.)
Anyway: Panko has some nice stuff about (scary!) spreadsheet errors.
1 comment
[ 3.8 ms ] story [ 9.6 ms ] threadI was involved in some work around English health care and "dignity and respect".
The process was horribly flawed. Volunteers asked patients some questions off a questionnaire. The answers were sent back, and reports were written.
Obviously, the sample sizes were not random; they were tiny; etc etc.
And the people writing the reports didn't get any help from statisticians. So you got comments like "28.69% of people said they felt [...]". I pointed out this obvious glaring error. They were grateful, and said of course they percentages should have been rounded to the nearest 0.25%. (This was with a sample size of 16 people.)
Anyway: Panko has some nice stuff about (scary!) spreadsheet errors.
(http://panko.shidler.hawaii.edu/My%20Publications/Whatknow.h...)
And there's even a European Spreadsheet Risks Interest Group.
(http://www.eusprig.org/stories.htm)