# What is the difference between tags "validity" and "validation"?

Should and be synonyms, or there is a difference in meaning between them?

So far (07th, July, 2013):
was used 16 times on CV and 69 times.

Edit 1:

This is a first suggestion for tag wiki excerpt given the answers of @gung and @Glen_b.

Validity is the extent to which a concept, conclusion or measurement is well-founded and corresponds accurately to the real world. The validity of a measurement tool is considered to be the degree to which the tool measures what it claims to measure. DO NOT use this tag for "validation" model purposes.

Validation is the process of deciding whether the numerical results quantifying hypothesized relationships between variables, obtained from regression analysis, are in fact acceptable as descriptions of the data. DO NOT use this tag for asking if it is valid ("validity") to use a certain test, measurement or analysis.

References: Wikipedia 1, Wikipedia 2

Edit 2 (status-completed):

The tag wiki excerpts were edited and approved according to modifications suggested by @gung and @Glen_b. See them here and here.
Further improvements can be suggested directly on the tag wiki excerpts pages or with the opening of a new thread.

Thank's.

• -1 / +1; I think the excerpt for [validity] is fine & shouldn't be changed. Your suggestion for the excerpt for [validation] seems fine; I'd say you should go ahead & edit. Jul 18 '13 at 15:28
• @gung. Thank's for replying. And what about the part of "do not use this tag for validation model purposes". Do you think it would be useful to aware askers about "validation" tag? Jul 18 '13 at 15:49
• I think your proposal for the [validation] tag is fine; you could include that part as well. Jul 18 '13 at 15:54
• +1 as it stands, however I think the later part of the final sentence of validation "for asking if it is valid ("validity") to use a certain test, measurement or analysis." doesn't quite relate to the sense of 'validity' being conveyed in the first tag, which is about the relationship between tests or instruments (like a particular IQ test) and the real world (i.e. 'How can we be confident this IQ test actually tells us about something real?' is a question relating to validity). The sentence gives the impression the validity tag is about when it's okay to do t-tests. ... (ctd) Jul 19 '13 at 2:18
• (ctd) ... I suggest that the wording be altered to "for discussing validity of a measurement or instrument -- such as that it measures what it purports to." or something similar. Jul 19 '13 at 2:20

I don't think these should be synonyms:

• Validity is a property of something. For example, it can be a property of a measurement--that it measures what it purports to measure.

This is what the tag wiki excerpt currently indicates, and that seems to be how it's being used (note: I edited one question where validity was being used in a different sense). Nonetheless, there are other possibilities, such as a property of an experiment (e.g., internal & external validity) or a property of a model; if need be, we could create additional tags to denote these and ensure the excerpts do a good job of guiding people in their use. At present, this doesn't seem to be necessary, though.

• Validation is a process. Examples might be cross-validation or sensitivity analyses. (I'm sure there are more.)

However, I do think we need to add a tag wiki excerpt for to guide people in its appropriate use. We simply need to figure out here what that should be.

• I agree with this - they shouldn't be synonyms, and if they're being used as such, they're being misused. Jul 9 '13 at 21:45

Usage of 'validity' tends mostly to refer to the sense in used in psychometrics, education and other social sciences and often refers to specific tests or instruments to measure some particular thing ("IQ-test validity"). 'Validation' is usually used in a different (though not entirely unrelated) sense relating to the properties of particular statistical models - model-validation, predictive-validation, cross-validation.

While occasionally they're used to talk about the same thing (like when a education researcher talks about validation of an instrument or someone who has just done some model validation talks about model-validity), usually the two are used quite differently.

As such I don't think these should be synonyms.