文档介绍：Edited by Foxit PDF Editor
Copyright (c) by Foxit pany, 2004 - 2007
For Evaluation Only.
Statistical parisons Pat Hammett
Comparing Quality
Statistical Inference Tests
t-tests
F-tests
Analysis of Variance (ANOVA)
1
Statistical Tests for Quality
n We use statistical tests in quality engineering to:
pare the quality of different products and processes.
n identify the critical sources of variation.
n Among the mon statistical tests are:
n t-tests: used to test means or location effects
n example: if a plant has two parallel assembly lines, t-tests
may be used pare the mean of Line 1 versus the
mean of Line 2.
n F-tests: used to test variances or dispersion effects
n example: F-tests may be used pare the variance of
Line 1 versus the variance of Line 2.
2
1
TQM - University of Michigan
Edited by Foxit PDF Editor
Copyright (c) by Foxit pany, 2004 - 2007
For Evaluation Only.
Statistical parisons Pat Hammett
Case Example: Car Body
Check Point #1
Height of Dash HOOD
Check Point #2 Left Fender
Underbody: Front partment
3
Example: Parallel Assembly Line
n panies use parallel assembly lines (more than
one assembly line performing identical functions).
n Consider the following case of a NA automotive plant
which uses parallel assembly lines for their underbody.
North Framing Line (set #1) North Respot Line (set #1)
Station Station
Station Station
1 2
3 4
Measure
Framing Queue Respot Underbody
Station Station Station Station
1 2 3 4
South Framing Line (set #2) South Respot Line (set #2)
UNDERBODY ASSEMBLY LINE
4
2
TQM - University of Michigan
Edited by Foxit PDF Editor
Copyright (c) by Foxit pany, 2004 - 2007
For Evaluation Only.
Statistical parisons Pat Hammett
parisons
n Question: Are the two assembly lines
producing similar outputs?
n Output Measurement - height of Dash Panel
Suppose we want to test if: Line AVG StDEV
- North -