De Granda-Orive et al.1 argue that self-citations should not be counted when measuring the impact of publications. Garcia-Pachon et al.2 argue that self-citations should be counted in journals with limited distributions–particularly those that are not published in English. I want to propose a third alternative. This is that the in-text citations should be counted rather than those in the reference lists–and more than once if they occur more than once in the text.
I have published my reasons for suggesting this elsewhere.3 They are briefly summarised as:
- -
The number of in-text citations indicates the weight attached to these references in an article (Hou et al.4) This weight is reduced if there are many in-text citations but only one or two in the reference list at the end. This reasoning applies both to self-citations as well as conventional ones.
- -
Not all self-citations are there just for self-enhancement: they can tell the reader where to find a key publication; provide evidence for the writer's claims; draw the reader's attention to little-known (and unknown) work; indicate to the reader the scholarship and experience of the writer; align the author with a particular school of thought; and show the development of thought.
Thus the best way, I submit, to count the references in an article is to count their frequency in the text (not the list) according to their purposes. In my article3 I provide three examples of how this procedure might work.
Please cite this article as: Hartley J. Las autocitas deberían considerarse relevantes. Arch Bronconeumol. 2015;51:102.
