random pattern of points and non-random patterns

In a control chart what is meant by random pattern of points and non-random patterns 

 

 

Random Points : The points on the control chart, within Control Limits (UCL & LCL) AND NOT CONSECUTIVE SEVEN POINTS.  Means the process is under control, no need to take any action


Non Random Points : The points on the control chart, within Control Limits (UCL & LCL) and CONSECUTIVE SEVEN POINTS above or below Mean (Actual Required Value). Means the process is out of control, need to take any action

Sai, As per the PMBOK glossary random points happens during normal cause and non-random during special cause. So I m not sure whether your explanation regarding Random points is correct.

Common Cause.

A source of variation that is inherent in the system and predictable. On a control chart, it appears as part of
the random process variation (i.e., variation from a process that would be considered normal or not unusual),and is indicated by a random pattern of points within the control limits. Also referred to as random cause.
 
Special Cause.
A source of variation that is not inherent in the system, is not predictable, and is intermittent. It can be
assigned to a defect in the system. On a control chart, points beyond the control limits, or non-random
patterns within the control limits, indicate it. Also referred to as assignable cause.

 Hi Sai, Pkukilla,

 

 “Control Charts” is a Statistical Process Control (SPC) technique for observing and controlling process behavior. Although you both have given partial facts, the statements contain few ambiguities as well as few wrong interpretations.

 

Let me take this opportunity to first explain control charts and then show how it must be interpreted. I hope this is going to clear a lot of doubts but this is going to be a long drive so please fasten up your seat belts J

 

SPC assumes that studying each and every measurement from an entire population is not practical most of the times and hence is based on sampling and sampling distributions. (Assume that a factory produces 10,000 fuses every minute. It will be highly impractical, if not impossible, to test each and every fuse for its quality standards). SPC requires “random” samples to be taken from the population periodically to estimate the population parameters, in other words, the process behavior. To estimate the population parameters based on samples, the samples must be taken out randomly from the population so that the each and every measurement in the sample is independent of all the other measurements in the sample and also all the measurements in the sample are normally distributed. In simple words, samples must be scattered and not very closely related, otherwise the population will not be accurately estimated.

 

We plot these samples’ measurements in a control chart. If the points in the chart are showing a “pattern”, that means that the sampling was not random and hence we cannot accurately estimate the population (this will be a rare event, a “special-cause”). If the pattern is random, it implies that sampling was random (it includes normal process variation, i.e. “common cause”), and SPC techniques can be applied to study the process behavior.

 

In simple words if the samples are not random they do not represent a ‘normal distribution’ and hence SPC’s techniques cannot be applied to study the process.

 

Having explained that, let me focus on the interpretation of the control charts: If any sample point is outside the control limits, this means that the process is out of control for sure. However usually quality assurance engineers/auditors do not wait for such an event to happen before they declare a process to be out of control. Process behavior are analyzed in a pro-active manner so that issues are caught during their inception. For this we apply “run rules” to a control chart to find out whether such problems are being apparent.

 

There are 8 “run rules” that catch a process if it starts to show unusual variation. Among these 8, there is a rule that states that if nine points in a row are on the same side of center line (above or below), the process is out of control. We conclude this because this unusual event can happen only due to two reasons. Firstly, this can happen due to non-random sampling. Secondly, this can happened due to a rare or special cause event has occurred. In both cases, we have to further study the measurement system or the process to remove this issue. Please note that most of the project management books refer to this rule by “Rule of Seven” with an exception that the rule applies to 7 consecutive points rather than 9 points. This is due to the fact that in project management, usually we don’t have big population and sample sizes, so for better and early control of the process, we tighten the SPC rule from 9 points to 7 points.

 

If I haven’t succeeded in explaining these concepts, please let me know.

 

Regards,

 

Exam Support Team
The PM PrepCast - http://www.premiumcast.com/vp/50398/16780/10389/

 Perfect. Thanks for clearing my doubts

Perfect. Thank you very much.

I'd never seen this explanation regarding the 'Rule of Seven' resulting from the small sample numbers. Perhaps I've learned something new today.

Thanks to Exam Support Team for the elaborate & lucid explanation !!!

Regards,
Ranjit

These explanations are absolutely correct. These are very detailed and explicitly stated vs what is explained in most PMP classes. I commend you for taking the time to clearly state why we measure the 7 points on either side of the mean and out of control points.