Cue the scary music – statistics and data analysis…a scary endeavor for most. But with proper training, you can avoid the pitfalls before they happen, create a quality product, and have pretty charts to prove it. Again I state that this is not a portal for teaching details, rather a gentle reminder of what Project Managers (should) do all day. Eat (or is it crunch? I never can remember…) Numbers!
“Statistics describe physical conditions and situations by using two indicators: measures of location and measures of dispersion.”
–Lesson 9 :: KSU PM Applications :: Prof Swaim
Measures of Location
Huh? You want me to measure my location? There area three ways to measure this; by median, mode and mean.
Normal distributions are, well, normal. Remember the bell curve? 
The highest point is the mean which is an average (don’t use this word among statisticians!) of all of the data which isn’t always the mode or the median for that matter.
Lots of data can be categorized using this curve.
Measures of Dispersion
Data cannot be purely measured by viewing, rather by doing. ‘Doing’ means you must put some work into it.
LEGEND:
X = value
N = number of values
Range = largest X – smallest X
Mean = Sum(all X) / N = xBar
Mean Absolute Deviation (MAD) = Sum(X – xBar) / N
Variance = Sum((MAD)Squared) / (N-1)
Standard Deviation = Square Root of Variance = SD
Take another look at the Normal Distribution Curve; the SD is the next point on the horizontal line where the next distribution occurs. So if you look at the percentages 34.13 % are in the SD+1 —– so if the mean is 50 (random even number I picked from the air) and 11 (also taken from the air) is the SD, then 34.13% of the distribution are between 50 and 61. If we go to the left of mean at SD -1, then 34.13 % are (50-11) at 39.
It is just saying that the greater the SD, the greater the spread (or distance from mean).
Variation and Process Improvement
Instead of just being hit or miss, one theory we can follow is the Taguchi loss function to help with quality control or product variance. It identified improvement opportunities.
Statistical Process Control
To reduce variation and boost quality. This requires a Project Managers dedication to measure results regularly and systematically.
Ask yourself:
- who – are the stakeholders
- what – do the stakeholders want and what do I want to achieve
- how – does the process get carried out and what is its output
- how – can it be better
- what – are the barriers
I’m really good at #5 – my husband calls it worrying; I call it thinking ahead and using risk assessment principles.
Filed under: Project Management Certification
