Project Quality Management: Control Limit vs Specification Limit for PMP Exam

Project Quality Management: Control Limit vs Specification Limit for PMP Exam

Control charts are one of the Seven Basic Quality Tools used in Control Quality as described in PMBOK® Guide to give a visual depiction of the changes in the outcome of a process. Though in the ideal situation, a process will give the same results for each and every time it runs. However, in reality, the results will be different from each other owing to a number of factors (common causes and special causes). The Control Limits and Specification Limits are a threshold for evaluating when the process is under control or not.

This post will expound on the similarities and differences of Control Limit vs Specification Limit and what Aspirants would need to know for the exam.

Control Limit vs Specification Limit

  • Control Limit: the limit established for the control chart based on statistical analysis or from historical records
    • There are 2 Control Limits: Upper Control Limit (ucl) and Lower Control Limit (lcl) indicating the maximum and mininium allowable values respectively
    • By convention, the Control Limits would usually be±2 or ±3 standard deviations (σ) from the target value, though this will vary from process to process
    • If there are results falling outside the Control Limits, the process is said to be unstable (i.e. out of control) ; root analysis is needed to investigate the special cause and rework/scrap would be needed
    • In addition, there are still some other pattern that will indicate the process is out of control, e.g. 7 consecutive data points on either side of mean (Rule of Seven)
  • Specification Limit: the allowable deviations requested based on customer expectations
    • There are 2 Specification Limits: Upper Specification Limit (usl) and Lower Specification Limit (lsl) indicating the maximum and mininium allowable values respectively
    • If there are results falling outside the Specification Limits (but within Control Limits), improvement works would be needed to adjust the process to give greater precision; nevertheless the process is still considered stable
    • The Specification Limits would be documented in agreements and penalties would be imposed if the results fall outside these Specification Limits
    • Specification Limits may be within or beyond the Control Limits — that would depend on the precision that is required for the process

Control Limit vs Specification Limit Illustrated

Let’s take the project of PMP® study and preparation as an example to illustrate the concept of  Control Limit and Specification Limit.

When working on some mock exam papers, your score for 5 different mock question sets were:

Scenario 1

  • 72%
  • 70%
  • 68%
  • 69%
  • 73%

The mean for the above 5 mock exam paper results is 70.4% which is above the minimum recommended score (70%) for mock exam results and all these scores were within the control limits of 2σ (74.5% and 66.3%). So the the Aspirant is said to have a stable performance and is expected to be able to pass the exam in first try.

Scenario 2

  • 62%
  • 60%
  • 58%
  • 59%
  • 63%

Though all the scores are within the control limits (64.5% and 55.3%), the mean of 60.4% is below the recommended minimum score. Though the PMP® Aspirant has a stable performance, he/she will likely to fall the real exam.

Scenario 3

  • 65%
  • 65%
  • 66%
  • 65%
  • 66%
  • 65%
  • 100%

The mean here is 70.3%, which is above the minimum recommended score (70%) for mock exam results, however, the exam results with 100% is well beyond the control limits (96.4% and 44%). Action would be needed to find out the root cause for getting the 100% results (e.g. is it just too easy?).

Scenario 4

The PMP® Aspirant would like to place a Lower Specification Limit of 70% for all the mock exam results as a threshold for writing the real exam. From the results of the above 3 scenarios, there are quite a number of results that are lower than this Specification Limit and the Aspirant will need to study harder and bridge his/her knowledge gap before attempting the real exam. If he/she has booked the exam, he/she will need to postpone the exam date then.

Summary: Control Limit vs Specification Limit

Aspirants would need to remember Control Limits are there to indicate whether the process/system is under control or not. Results falling outside the Control Limits would mean the process is unstable and root cause analysis is needed. Specification Limits are imposed by agreement with customers on strictly quality requirements, i.e. Specification Limits must be within the Control Limits.

Hope this article can illustrate the differences between Control Limit vs Specification Limit well.

recommended PMP resourcesAdditional FREE PMP® resources: 47+ Commonly Confused Term Pairs with detailed explanations. If you found this article useful, you may wish to reference other Commonly Confused Term articles.

Most Popular PMP Certification Exam Articles

GreyCampus PMP online training course bundle for US$149 only

Support website running for FREE, thanks!

If you find this post helpful and if you are thinking of buying from Amazon, please support the running cost of this website at no extra cost to you by searching and buying through the search box below. Thank you very much for your help!

Edward Chung

Edward Chung aspires to become a full-stack web developer and project manager. In the quest to become a more competent professional, Edward studied for and passed the PMP Certification, ITIL v3 Foundation Certification, PMI-ACP Certification and Zend PHP Certification. Edward shares his certification experience and resources here in the hope of helping others who are pursuing these certification exams to achieve exam success.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *

8 Responses

  1. Chinmaya Das says:

    The article states “Specification Limits must be within the Control Limits.” where as I believe “Control Limits must be within the Specification Limits.” Could you please clarify?

    • Edward Chung says:

      Yes, “may be” is a better term here than “must be” though in general practice, Specification Limits usually fall within the Control Limits in my experience.

      Maybe you have a different perspective?

      Wish you PMP success!

  2. Stefan says:

    Hi Edward,

    First of all, congratulations to your webiste. I have scheduled my PMP exam in less than 2 weeks and I am going through your 47 Commonly Confused Term Pairs. Some of these terms were confusing indeed, but your explanations and examples helped a lot.
    I have trouble following your calculation for ucl and lcl of 2σ (74.5% and 66.3%) in scenario 1. Can you please explain how you calculate ucl and lcl?

  3. Joe says:

    RE:”Specification Limits must be within the Control Limits”
    I think above statement of yours should state “may be” instead of “must be” since PMBOK 5th, current, edition states “This area (referring the Specification Limits) may be greater than or less than the area defined by the control limits.” on page 563. Also other sources I have visited express/show that the Control Limits within the Specification limits.

    • Edward Chung says:

      Hi Joe,

      Thanks a lot for pointing this out. I have amended the article above accordingly.

      Yes, “may be” is a better term than “must be” though in general practice, Specification Limits usually fall within the Control Limits in my experience.

      Wish you PMP success!

  4. rakesh says:

    Does Control limit less then the Specification limit help prevent the defects? As such it will trigger/flag and prevent the product/deliverable go out of specification.

    • Edward Chung says:

      On the surface it seems yes.

      But the very aim of control limits is to be used as a telltale as whether the system is stable and under control. Defining control limits is more a statistically process — i.e. we need to run the process and record the results for a number of time and determine the level of “common cause variation” of the process. From the data, we can then calculate the control limits according to our requirement for precision (commonly ±2σ or ±3 σ).

      There may be some common cause that prevent the process to be as “precise” as we want that would deem the process to be out of control if the control limits are set to be so tight. We may need to rework the process if more stability is required. A lot of time and cost would be involved.

      In theory, a 100% precision without fluctuation in the deliverable is the ideal process. However, this may not be possible owing to other limitations that are beyond our control.

      The specification limits are the allowable deviations requested based on customer expectations. If the process is stable (deliverables within control limits) but not precise enough (deliverables outside specification limits), we can then fine tune the process to make the output more precise. The cost would be much lower.

      After all, quality control is more an art of balancing cost and quality.

      Hope this helps to clarify.