(Not from a paper) Self-report data should be labelled as such

[Dolphin]

http://www.intropsych.com/ch01_psychology_and_science/self-report_measures.html

Self-Report Measures: Notoriously Unreliable

[..]

What should one always do, if using self-report data?

If one must use self-report data, they should always be labeled as such. (The word "they" is used here because the word "data" is plural; "datum" is the singular form.) For example, if you ask people whether they are happy, you are gathering self-report data. The data might not be very accurate, so the researcher should caution readers by labeling this data as self-reported happiness. Indeed, this is a typical approach to the problem of measuring happiness, mentioned earlier. Researchers commonly use self-report data labeled as such

I thought this was interesting.
For example, one will often see in reviews that GET or CBT has been shown to improve "physical functioning". By the nature of the words, one might think some sort of objective test of physical functioning was used. However, what is being referred to in reviews (it may be different in some individual studies) is the results of a questionnaire, the SF-36 physical functioning subscale (which might be subject to response biases). According to the guidance I quoted, they should be making clear it is "self-report" data.

Sometimes the physical functioning finding is extrapolated further e.g. GET or CBT has been shown to restore the ability to work. Again, often/usually what is being referred to is the results of the SF-36 physical functioning questionnaire. But again the self-report nature of the data is often not made explicit.