I have seen many science teachers argue furiously over accuracy, validity, reliability and precision. Does anyone have some definitions that are universally accepted? I thought this might be a good topic for discussion.
Cheers,
Jeff
Posted by: Jeff Stanger | April 9, 2009 06:57 PM
Comments
Hey Jeff,
When I am explaining this to students (and arguing in the staffroom), I use extreme arguments to highlight the differences.
For example, I say that to measure the temperature of a room, it would not be valid. It may be reliable, but not valid.
Here are my rough definitions for discussion:
Accuracy - how close a result comes to an accepted value (e.g. measuring g).
Validity - relates to the method of an experiment; does it actually test what it claims/aims to?
Reliability - essentially the same as repeatability. Do you get the same results every time? An experiment can be inaccurate but reliable due to a systematic error (e.g. a wrongly calibrated piece of equipment).
Precision - consistent results (e.g. in archery or darts, hitting the same mark every time). It's very hard to distinguish this from reliability - precision refers to a certain degree of consistency i.e. having an answer to 6 decimal places is more precise than 2.
The other comment is that I think that accuracy and precision are best paired/spoken about together, and validity/reliability as another topic.
Other thoughts?
Andrew
Posted by: Andrew | June 30, 2009 04:09 PM