Scribd
Upload
101 views

Introduction To Control Charts: XMR Chart: Farrokh Alemi, PH.D

This document introduces the XmR control chart, which is used to visually analyze changes in a key measure over time and determine whether those changes represent common cause variation or statistically significant special cause variation. The document explains how to calculate the moving range of observations, determine the average moving range, and establish the upper and lower control limits. An example demonstrates how to plot observations on an XmR chart, interpret any points that fall outside the control limits, and share the results with others.

Uploaded by

suresh2711991
Copyright
© © All Rights Reserved
Available Formats
Download as PPS, PDF, TXT or read online on Scribd
101 views

Introduction To Control Charts: XMR Chart: Farrokh Alemi, PH.D

This document introduces the XmR control chart, which is used to visually analyze changes in a key measure over time and determine whether those changes represent common cause variation or statistically significant special cause variation. The document explains how to calculate the moving range of observations, determine the average moving range, and establish the upper and lower control limits. An example demonstrates how to plot observations on an XmR chart, interpret any points that fall outside the control limits, and share the results with others.

Uploaded by

suresh2711991
Copyright
© © All Rights Reserved
Available Formats
Download as PPS, PDF, TXT or read online on Scribd
You are on page 1/ 17

Introduction to Control

Charts: XmR Chart


Farrokh Alemi, Ph.D.

Purpose of Control Chart


To judge whether change has led to

improvement.
To visually tell a story of changes in a key
measure over time.

Farrokh Alemi, Ph.D.

Index

Elements of Control Chart


X-axis shows time periods

UCL
50

Y-axis shows the observation

values
UCL line shows the upper
control limit
LCL line shows the lower
control limit
95% of data should fall
within UCL and LCL
Values outside the control
limits are likely to be
statistically significant

45
40
35

Observations

30

LCL

25

Farrokh Alemi, Ph.D.

1st
Qtr

2nd
Qtr

3rd
Qtr

Index

4th
Qtr

Definition of Statistical Control


Variation occurs in any outcome of interest over

time.

In a stable situation, some variation will occur just by


chance, but it will be predictable over time. Statisticians
call this common cause variation or within control
limits.
If there is a significant change, data points will show up
outside the range expected for chance variation alone.
Statisticians call this special cause variation or
outside control limits.

A control chart allows us to detect statistically

important changes.

Farrokh Alemi, Ph.D.

Index

Which Chart is Right?


For different outcomes different control

charts are appropriate.


Click here to see which chart is appropriate
for the outcome you have in mind.
This presentation focuses on one type of
chart named XmR chart.
Assumptions of XmR chart
Farrokh Alemi, Ph.D.

Index

Assumptions of XmR charts


There is one observation per time period.
Patients case mix or risk factors do not change in

important ways over the time periods.


Observations are measured in an interval scale,
i.e. the observation values can be meaningfully
added or divided.
Observations are independent of each other,
meaning that knowledge of one observation does
not tell much about what the next value will be.
Farrokh Alemi, Ph.D.

Index

Moving Range
An XmR chart is based on the absolute

differences between consecutive values,


displayed as a Moving Range
Even when observations come from nonnormal distributions, differences in
consecutive values form a normal
distribution as the number of observations
increases
Farrokh Alemi, Ph.D.

Index

Calculating the Moving Range


The number of observations is n.
The absolute value of the difference

between consecutive values is the moving


range, R
An example follows

Farrokh Alemi, Ph.D.

Index

Example Data
Time period
Observation

1
90

2
85

3
92

4
67

Farrokh Alemi, Ph.D.

5
98

6
83

7
94

Index

8
90

Calculating the Average


Moving Range
Time period
Observation
Moving range

1 2 3 4 5 6 7 8
90 85 92 67 98 83 94 91
5 7 25 31 15 11 3

Mean
87.50
13.86

Add the differences and divide by n minus


one to get the average moving range.
Mean R = |(Xt - Xt-1)| / (n-1)

Farrokh Alemi, Ph.D.

Index

Calculating Upper and


Lower Control Limits
If E is a correction constant, then:
Upper Control Limit = Average of observations
+ E * Average of moving range
Lower Control Limit = Average of observations
- E * Average of moving range
Farrokh Alemi, Ph.D.

Index

Correction Factor Depends on


Number of Time Periods
Number of
Number of
time
time
periods E values
periods
11
2
2.660
12
3
1.772
13
4
1.457
14
5
1.290
15
6
1.184
16
7
1.109
17
8
1.054
18
9
1.010
19
10
0.975
20

E values
0.945
0.921
0.899
0.881
0.864
0.849
0.836
0.824
0.813
0.803

d2 values

Based on Wheeler DJ. Advanced topics in


statisical process control, 1995 SPC Press
Inc, Knoxville TN 37919

Farrokh Alemi, Ph.D.

Index

Calculating Upper and Lower


Control Limits
Time period 1 2
Observations 90 85
Moving range
5
Upper control limit
Lower control limit

3
4
5
6
7
8
Average
92 67 98 83 94 91
87.50
7 25 31 15 11 3
13.86
= 87.50 + 1.054 * 13.86
= 87.50 - 1.054 * 13.86

E for 8 time periods is 1.054


Farrokh Alemi, Ph.D.

Index

Plot the Control Chart


Plot the x and y axis
Plot the observations
Plot the upper control

limit
Plot the lower control
limit
Variation among
observations that fall
between control limits
is likely due to chance

110
100

Observations

90
80

UCL
LCL

70
60
1

Farrokh Alemi, Ph.D.

Time periods

Index

Interpret the Control Chart


Points outside the limits

represent real changes in


the outcome of interest
The observation at time
period 4 falls below the
LCL; it is unlikely that this
is due to random chance
events
The next step is to
determine the possible
causes of this significantly
different observation

110
100

Observations

90

UCL

80

LCL

70
60
1

Farrokh Alemi, Ph.D.

Time periods

Index

Share the Results With Others


Control charts are effective ways to

visually tell a story


Distribute the chart by electronic media, as
part of a newsletter, or as an element of a
story board display
Show that you have verified any
assumptions, check that your chart is
accurately labeled, and include your
interpretation of the finding
Farrokh Alemi, Ph.D.

Index

Index of Contents
Purpose of Control Ch

Example Data

art
Elements of Control C
hart
Definition of Statistica
l Control
Which Chart is Right?
Assumptions
Moving Range
Calculating Moving
Range

Calculating Average

Moving Range
Calculating Control
Limits
Plot the Control Chart
Interpret the chart
Share the Result

Farrokh Alemi, Ph.D.

Index

You might also like