Evaluating the Impact of Social Work Scholarship Through H-Index and Related Bibliometric Indicators

Background: Bibliometric indicators are increasingly used as measures of research productivity and may be included as measurement criteria for promotion and tenure and/or funding decisions. The h-index, a bibliometric indicator used in many scientific fields, has only recently been used to measure the productivity of social work scholars. H-index values represent both the number of articles published and the rate of citation; a scholar with h-index=7 has published 7 articles that are cited 7 or more times per article. Norms for social work scholars are needed so that the h-index values of social work academics are not gauged using norms from other fields. Previous work has only reported h-index values for faculty at Top-10 institutions or members of elite editorial boards.

Purpose: To estimate h-index values and related bibliometric statistics for a heterogenous group of social work faculty.

Method: A probability cluster sample of Top-80 schools of social work according to the U.S. News and World Report reputational rankings was drawn, stratified by quartile. Resampling was performed until there were >30 professors at each rank within each quartile. Bibliometric data for 575 professors representing 26 different schools of social work were analyzed. A time-intensive coding methodology was used, with at least 2 raters separately (and blindly) coding each professor. Estimated h-index values were created using decision rules.

Results: Due to the skewed distribution of h-index values, the median represents the best measure of central tendency for these data. Median values for assistant professors were, in the 1st quartile (n=46), 4.0; 2nd and 3rd quartile (n=92), 3.0; and in the 4th quartile (n=34), 2.0. Associate professors in the 1st quartile (n=46) had a median value of 9.0; in the 2nd quartile (n=41), 6.5; in the 3rd quartile (n=62), 4.75; and in the 4th quartile (n=43), 5.0. Full professors in the 1st quartile (n=53) clearly had the highest median value, at 13.5; in the 2nd quartile (n=59), 8.0; 3rd quartile (n=37), 10.0; and, 4th quartile (n=33), 9.0. At each academic rank, there was a statistically significant relationship between quartile and h-index (e.g., for full professors, Kendall's Tau=-.281, p<.001). Beyond h-index, other bibliometric variables, such as g-index, h_c-index, and h_i-index, showed the same general pattern of results.

Conclusion and Implications: When grouped by reputational quartile, h-index values generally fell with reduced ranking. However, there were important exceptions, congruent with the existing literature that suggests reputational and bibliometric rankings measure somewhat different concepts. While statistically significant, the real world importance of differences between quartiles may be questioned. The difference between estimated h-index values among assistant professors is likely not meaningful; median values of 2.0 and 4.0 are very close and could reflect the difference of only a handful of citations. Most importantly, h-index values across academic social work are more modest than the limited previous research has suggested. These results may provide important context for evaluative efforts which use the h-index as a gauge of research productivity and impact.