Page 331 - Six Sigma Demystified

Part 3 S i x S i g m a To o l S 311

Improve Stage
• Since the number of errors tends to be quite small (for even very large
samples), the use of attribute charts is limited in the improve stage.

Methodology

Samples are collected from the process at specific points each time. Each sample
(at each point in time) may have a unique number of units, each of which must
either match or not match some criterion of interest (such as being defective).
Plotted statistic: The percent of items in the sample meeting a criterion of
interest.

(count )
p = j
j n j

where n is the sample size (number of units) of group j.
j
Centerline:
∑ m (count )
p = j =1 j
∑ m j =1 n j

where n is the sample size (number of units) of group j, and m is the number
j
of groups included in the analysis.
UCL, LCL (upper and lower control limit):

p 1( − p)
UCL = p + 3
n j
 p 1( − p) 
LCL = max0, p − 3 
  n j  

where n is the sample size (number of units) of group j, and p is the average
j
percent.
For example, we observe a varying number of telemarketing calls each
week, and we record the total number of calls with one or more errors. Ten
months of data are observed, with the following number of calls with errors
in each sample: 7 of 350, 14 of 312, 12 of 125, 18 of 170, 9 of 165, 11 of 264,
12 of 254, 8 of 404, 11 of 137, and 21 of 312. The control chart is show in
Figure F.23.