Statistical Question

[flyingmonkeyofdoom]

I want to use weekly reports to be more pro-active in spotting, and resolving, potential problems for the month
At the moment its done monthly and is purely reactive

The only problem i have is that the sample size is so small that any extreames cause the data to be highly inaccurate

Once compiled for the monthly the extreams aren't as extream

Can anyone help, advise or point me in the direction of something that maybe able to help me work around this problem?

 

[WaterGypsy]

Have you considered a rolling 4 week report ?

Instead of 4 weekly reports you always report the last four weeks and then once a month it is the monthly report.