Help with Statistics

Companies in the life science sector must comply with standards and regulations that require the use of statistical techniques appropriate for the type of analysis to be done, and they must provide a rationale for the sample size selected based on the associated risk.


Clients find it beneficial to seek support with statistical queries used for data collection, data analysis, demonstrating process capability & demonstrating product capability.  There are a number of sampling plans to choose from, and the selection of the correct sampling plan can be challenging.  I cover several statistical equations in my training course on the estimation of measurement uncertainty  SQT Training Ltd. Many of these equations appear daunting at first. My advice is to break the equation into bit-size pieces and determine what each symbol of the equation means. 


Laboratory Proficiency Testing  

Participation in proficiency testing is an essential independent verification of the lab's performance.   I can explain the calculations behind the various scores clients receive from their proficiency test providers, including;

·     Z score

·     Zeta score

·     En score

·     Z prime

Clients find it helpful to understand that the result (score) is not just down to the input from your lab but will also be influenced by the other participating test labs and the proficiency test provider.  

I demonstrate how labs can check their results using appropriate statistical calculations and share helpful ways to trend them.  I believe it is essential for labs to verify these scores and to plug the data into the formula to double-check the result. 

Did you know that data from proficiency testing can also be used as part of the estimation of measurement uncertainty?

Trending of Results

A simple control chart is just a chart of the results over time.  Control charts are valuable tools for monitoring and analysing laboratory or production floor results. However, some statistics knowledge is required to determine trends within the data.   

Various control charts are available, and I help companies select the appropriate charts for their requirements and develop control limits.   Quality Control charts are also used as part of test method validation and method comparison, estimation of measurement uncertainty, detection limits, checking the drift of equipment, comparison or qualification of personnel, and evaluation of proficiency tests. 

As with anything else, it's essential that you select the correct chart for your requirements and that you justify its use; otherwise, daily stacks of paper or computer-generated charts are churned out and possibly ignored.

The most common questions/statements I encounter are;

·     How do I pick a control chart?

·     Where do the limits come from?

·     How do I understand the trends, and how far do I go with corrective action?

·     We have always had control charts, but we're uncomfortable showing them to customers/auditors. 

Let me know if I can help demystify control charts.