By Fred Schenkelberg
A common question about reliability testing is “what is the sample size?” It is also a difficult question to answer well. The right sample size is a balance between cost, accuracy, and variability. In some cases, we also consider the time to results.
When planning reliability testing an important task is to determine the necessary sample size. Too few samples and the results will be indistinguishable from statistical noise. Too many samples and the cost of testing becomes more than necessary.
Understanding the basics of sample size calculations for a range of different situations and approaches permits you to quickly estimate the right sample size given the information you have available. The first estimate often leads to discussions concerning the constraints, goals, and striking the right balance of samples and outcomes.
Let’s discuss the many challenges and array of solutions available to you for sample size determination. We’ll examine a few different situations, formulas, and best practices to settle on the right sample size for each situation. Sample size starts as a statistical exercise then becomes a business/project conversation.
Learning to determine sample size is a critical skill for reliability professionals. Let’s talk sample size determination, getting the fundamentals right.