Page 215 - Six Sigma Demystified

Part 3 s i x s i g m a to o l s 195

Interpretation

The low p value indicates that the null hypothesis of equal means should be
rejected: One or more of the product averages are significantly different.
The confidence intervals are shown in the Minitab results for each of the
products. The asterisk denotes the mean value; the parenthesis indicates the
95% confidence interval about the mean. The results indicate that the mean
of product A falls within the expected range of averages for product C, and
vice versa. Similarly, the mean of product B falls within the expected range
of the product D average but not within the expected means of products A
or C.

Autocorrelation Charts

The autocorrelation function (ACF) is a tool for identifying dependence of
current data on previous data points. It tests for correlation (in this case, au-
tocorrelation) between observations of a given characteristic in the data set.
You may notice a similarity between the formulas used for the ACF and the
correlation index calculated in the scatter diagram. The scatter diagram is
used to test for correlation between observations of different characteristics,
whereas the ACF tests for correlation between observations of the same
characteristic.

When to Use

Measure Stage

• To investigate process autocorrelation and its effect on baseline data

Analyze Stage

• To analyze regression residuals for violation of independence assumption

Control Stage
• To develop a control strategy that considers the serial dependence of the
process