Quality Management 2017

A NALYZE

5. Characterization of the data distribution

• Statistical tests to verify the normality of the data

Part 1 :

 Graphical method : Henry line, Quantile-Quantile plot (Q-Q plot)

Introduction to SPC

✓ Aligned points  Hypothesis of a normal distribution ✓ Can be done without any particular sofware.

Part 2 :

Essential concepts to start with SPC

Intervals

✓ Not as powerful as numerical tests.

Cumulative percentages

Part 3 :

 Numerical methods : Shapiro-Wilk test, Anderson-Darling test, Kolmogorov test… ✓ Statistical softwares can easily do those tests ✓ The value of p-value measures how far the data is from a normal distribution. ✓ Choice depends on your data

SPC in practice

Part 4

Conclusion

03/10/17

18