
08/17/2009, 11:32 AM
#196
Originally Posted by TheMarco
Sigma  is working! However there seems to be an issue with entering single items rather than data pairs. Probably because of some stack lifting going on that shouldn't happen in these cases. I'm not completely sure what the behavior should be like when using single entries. Can you give me an example? Also, how can it be 2.5? From what I know the display after Sigma+/ always shows the amount of data points gathered which means it should be 1 after deleting one in your example rather than 2.5. Am I wrong?
I have only ever used the stat registers for single values, not data pairs; I'm aware that they can do data pairs, never found the functionality useful/productive. I'd shift to a stat package on my workstation when I needed something more involved like that.
In the example I cited upthread, the 2.5 value would be the expected result of the xbar function (average, mean, etc) if the sigma function had canceled out the "4" entry, leaving stats only for values 2, 3. Immediately after hitting sigma, I would expect x to show "2", reflecting how many values had been entered into the stat registers.
Sigma should remove the value in x (and I guess, y) appropriately from each stat register, then decrement the count so that if there were n values in the stats registers, after hitting sigma minus, the count should be n1. As near as I can tell, sigma is mainly an undo function for data entry; don't know if there are other uses of this key or not; that's the only way I've ever used it.
I'm also finding that the standard deviation function isn't matching my 15C:
clear sigma, 2, sigma+, 3, sigma+, 4, sigma+, s > 0.81649.... on 1.0b21. On my 15C and in Excel, I get a value of 1.000000.
My last stat class was a while ago *cough*moreyearsagothantheageofsomeforummembers*cough* but I vaguely remember that the formula for s is different when n is small. I wonder if you're using the formula for large n even for small datasets.


