Comparing Usnwr Reputation Versus h-Index Rankings of Social Work Doctoral Programs

Background/Purpose:  The h-index is a popular measure of scholarly productivity. h-index scores have been identified to be highly correlated with other ratings of research output, and with perceptions of scholarly prestige. The US News and World Report (USNWR) Ranking, in contrast, is the most revered system for ranking the reputations of graduate schools in the US. USNWR reputational ranking system for graduate programs in the social sciences relies on peer ratings.  Thus, the assessment of deans and directors are key in USNWR ratings. The purpose of this study is to assess how different school level factors the reputation of schools as shown by the USNWR ranking system.

Methods:  Information on the USNWR rankings, h -index, years of experience, academic rank, and gender of faculty were collected to better understand differences in rankings. A multiple regression analysis was used to explain variance in USNWR scores associated with tenure track faculty size, college age (i.e., years since establishment), region in the United States (i.e., Northeast, Southeast, Midwest, and West), auspice (i.e., publicly funded or privately funded), and each program’s mean h-index score. Following the results of the regression, we compared the rankings of social work doctoral programs using h -index versus the USNWR rankings. After removing 13 programs, due to USNWR rankings that were too low for direct comparison, this comparison assessed the match between rankings on the scores of 62 programs The accuracy of predicting scholarly productivity from USNWR rankings was determined by joint membership in the same quantile block (e.g., 1-10, 11-20, 21-30). 

Results:  The regression model including college age, school size, auspice, and region explained 53% of the variance in USNWR rankings. Adding h-index to the model led to an R2 change of .28, suggesting that 81% of reputation as indicated by USNWR rankings is explained by the variables in the model. h-Index was the strongest predictor in the final model. Of the 62 programs where rankings were compared, we found only 17 of the 62 programs had converging quantiles for the two types of ranking. 31 schools were found to have reputation based rankings in higher quantiles compared to their h-index based ranking. The remaining 14 schools were in quantiles with higher rankings than their reputation based ranking would suggest.

Conclusions:  Although research productivity explained more variance than any other variable in the regression model, our quantile analysis indicated that only 17 out of 62 programs were found to have a reasonable match between rankings based on h-index and those based on USNWR. Other programs have markedly different results between the two ranking systems. The impact of faculty size, college age, and geographic reason all explain variance in reputation that cannot be explained by the dissemination of studies between schools.