Join Chris and Fred as this thing we call incomplete data. This was prompted from a question we received from an Accendo listener. A piece of equipment has been in service for 8 years … but our listener only had data for the last 4 years. So what can we do?
- Data is often not ‘complete.’ This refers to scenarios where we might test a product, and only observe that it was not working at the end of the test duration. So we know if failed before (for example) 1000 hours of testing … but we don’t know when. Or it was still working at 1 000 hours. So the failure time is greater than 1 000 hours.
- Then there is ‘expert judgment.’ Many of us don’t like expert judgment. It feels subjective and too prone to biases and ‘nefarious’ types who want to ruin everything. But when you think about it … virtually every decision we make on a daily basis uses expert judgment. So get used to it!
- So how do we deal with ‘incomplete data?’ You need to modify your statistical inference. You can use the same model to analyze complete versus incomplete data. Data (incomplete or otherwise) is evidence you can use. You just need to modify how your model incorporates it.
Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques to field data analysis approaches.