Median age difference of references as indicator of information update of research groups: A case study in Spanish food research

Median age difference (D) is obtained by subtracting median value of the age distribution of references of a scientific paper from citing half life of the journal that published it. Such an indicator can be related to the state of knowledge of research groups and can show some interesting properties: 1) it must be related with the incorporation of information pieces in an informal way, say the rate of self-citations; 2) it must follow the natural tendency of the groups towards a progressively updated state of knowledge, and 3) more productive groups will tend to use more recent information. These natural hypotheses have been investigated using a medium sized Spanish institution devoted to Food Research as a case study. Scientific output comprised 439 papers published in SCI journals between 1999 and 2004 by 16 research teams. Their 14,617 references were analyzed. Variables studied were number of published papers by every team, number of authors per paper, number of references per paper, type of documents cited, self citation rate and chronological range of reference lists.
Number of authors per paper ranged between 1 and 15. The most frequent value (N = 128) was 3 authors. Average number of authors per paper is 4.03 (SD = 1.74). Mean number of references per paper (including review papers) is 33.3 (SD= 17.39) with slight differences between the groups. Mean self-citation rate was 13.72 % (SD = 11.7). The greatest chronological range was 119 years; half of all ranges was 30 years and the general mean for this variable was 33.34 years (SD = 16.34). D values were associated with self-citation rate and a negative relationship between D and chronological range of references was also found. Nevertheless, correlation figures were too small to reach sound conclusions about the effect of these variables. Number of references per paper, number of contributing authors and number of papers published by each team were not associated with D.
D values can discriminate between groups managing updated information and delayed research teams. Publication delay affects D figures. Discontinuity of research lines, heterogeneity of research fields and the short time lapse studied could have some influence on the results of the study. It is suggested that a great coverage is needed to evaluate properly D figures as indicators of information update of research groups