A real life case: a firm desing, produces and sells 52 vehicles. I collect failure data of the components of the vehicles. after a time, -for example- for the oil filter I have 52 data. Some of them are failed and the rest are still operational. I have durations info for all 52 filters. These filters are bought and installed from a supplier.

My question is, this set of 52 filters is my population or a sample?

My guess: population

My rationale: if i would be able to collect only 20 of the failure data and knowing there are 52 vehicles, then this set of 20 filters would be 'a sample'. Also note that, My vehicles operates understanding specific environmental stress. So the population is NOT the thousands of filters that the supplier produces and sells all over the world. I only interested in the failure times of MY VEHICLES.

Am i thinking in the correct way, can you please enlight me. Thanks, Best regards.

  • $\begingroup$ I think this is a very helpful real-world question. One of the hardest things to know is "does the specific environmental stress" make the 52 vehicles significantly different from another population of filters. One could have either a physics of failure rationale or a statistical rationale for deciding this though the former may well be hidden from view. $\endgroup$
    – Puffin
    Apr 9, 2019 at 21:38

1 Answer 1


You are thinking correctly. When you can analyze all the subjects with the characteristic you want to investigate you can say you are analyzing the entire population. When you can't physically analyze all the subjects you can take a representative "sample", study its characteristics and make inference on the entire population.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.