When generating a control chart for continuous individual variables, whats different with Rule Violations for Run and Rule Violations for MR?

(The dialogue makes two graphs, one of the run and one of a moving range?)

------------------------------
Andrew Liu
------------------------------

#SPSSStatistics

Good morning, Andrew. It's been awhile since I've looked at these so I'm rusty, but here is an example:

DATA LIST FREE /rainfall.
BEGIN DATA.
1.56 0.52 1.01
1.23 3.14 1.97
5.37 2.56 2.59
0.76 2.54 3.31
0.81 1.21 1.21
4.37 5.32 3.02
5.92 3.10 5.58
1.37 1.42 0.54
END DATA.

COMPUTE lag_rain=abs(rainfall-lag(rainfall)).
DESCRIPTIVES variables rainfall lag_rain.
GRAPH /LINE(SIMPLE)=VALUE(rainfall).
GRAPH /LINE(SIMPLE)=VALUE(lag_rain).

SPCHART /IR=rainfall /SPAN=2 /SIGMAS=3.​

As you can see, the charts measure different things. The first is a simple line chart of the observations, and SPCHART overlays the average of those as a reference. The second (moving range) uses uses a lag function to calculate the difference from an observation ( say "j") from the previous one ("j-1"), and overlays a line of the average of those "ranges". The upper and lower confidence limits are determined by standard formulas available in the Algorithms manual. Of course, violations of the rules depend on which chart they refer to.

------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------

Hi thanks! 

I think I understand how the I-MR control chart code works. 

I'm a little confused by how to interpret the MR section of the chart though. I understand it signals based on preset rules, but how do you interpret an out of control MR chart? (Assume my I chart and MR charts both show points out of control. It's my understanding that a MR chart with out of control points makes it difficult to interpret a I chart.)

Thanks so much,
Andrew

------------------------------
Andrew Liu
------------------------------

Original Message:
Sent: Tue February 08, 2022 12:03 PM
From: Rick Marcantonio
Subject: Control Charts

Good morning, Andrew. It's been awhile since I've looked at these so I'm rusty, but here is an example:

DATA LIST FREE /rainfall.BEGIN DATA.1.56 0.52 1.011.23 3.14 1.975.37 2.56 2.590.76 2.54 3.310.81 1.21 1.214.37 5.32 3.025.92 3.10 5.581.37 1.42 0.54END DATA.COMPUTE lag_rain=abs(rainfall-lag(rainfall)).DESCRIPTIVES variables rainfall lag_rain.GRAPH /LINE(SIMPLE)=VALUE(rainfall).GRAPH /LINE(SIMPLE)=VALUE(lag_rain).SPCHART /IR=rainfall /SPAN=2 /SIGMAS=3.​

As you can see, the charts measure different things. The first is a simple line chart of the observations, and SPCHART overlays the average of those as a reference. The second (moving range) uses uses a lag function to calculate the difference from an observation ( say "j") from the previous one ("j-1"), and overlays a line of the average of those "ranges". The upper and lower confidence limits are determined by standard formulas available in the Algorithms manual. Of course, violations of the rules depend on which chart they refer to.

------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------

Hi, again, Andrew. It seems to me that the charts are telling you different things (the first the individuals, the second the process, which shows change over some number of measurements aggregated). They are possibly related, but that would depend on how the points fall on both. There is an example here that shows each chart with some point(s) out of control. Perhaps it will help.

------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------

Original Message:
Sent: Sun February 13, 2022 02:27 PM
From: Andrew Liu
Subject: Control Charts

Hi thanks! 

I think I understand how the I-MR control chart code works. 

I'm a little confused by how to interpret the MR section of the chart though. I understand it signals based on preset rules, but how do you interpret an out of control MR chart? (Assume my I chart and MR charts both show points out of control. It's my understanding that a MR chart with out of control points makes it difficult to interpret a I chart.)

Thanks so much,
Andrew

------------------------------
Andrew Liu
------------------------------


Original Message:
Sent: Tue February 08, 2022 12:03 PM
From: Rick Marcantonio
Subject: Control Charts

Good morning, Andrew. It's been awhile since I've looked at these so I'm rusty, but here is an example:

DATA LIST FREE /rainfall.BEGIN DATA.1.56 0.52 1.011.23 3.14 1.975.37 2.56 2.590.76 2.54 3.310.81 1.21 1.214.37 5.32 3.025.92 3.10 5.581.37 1.42 0.54END DATA.COMPUTE lag_rain=abs(rainfall-lag(rainfall)).DESCRIPTIVES variables rainfall lag_rain.GRAPH /LINE(SIMPLE)=VALUE(rainfall).GRAPH /LINE(SIMPLE)=VALUE(lag_rain).SPCHART /IR=rainfall /SPAN=2 /SIGMAS=3.​

As you can see, the charts measure different things. The first is a simple line chart of the observations, and SPCHART overlays the average of those as a reference. The second (moving range) uses uses a lag function to calculate the difference from an observation ( say "j") from the previous one ("j-1"), and overlays a line of the average of those "ranges". The upper and lower confidence limits are determined by standard formulas available in the Algorithms manual. Of course, violations of the rules depend on which chart they refer to.

------------------------------
Rick Marcantonio
Quality Assurance
IBM