Comparing the Research Productivity of Social Work Doctoral Programs Using the h-Index

Background:  The search for meaningful metrics of program excellence has been a long-standing effort on the part of social work schools and colleges. The h-Index is one metric which has been highly correlated with both journal impact factors and subjective perceptions of scholarly prestige. However, there is a current gap in social work research that investigates the relationship between the h-index and school level characteristics. The purpose of the study was to examine the aggregate research productivity of faculty in social work doctoral programs, exploring the relationship between scholarly output and school characteristics.

Methods:  This study examined the scholarly productivity of the full population (N = 1699) of tenure-track faculty in all 76 United States social work doctoral programs accredited by the Group for Advancement of Doctoral Education (GADE). The h-Index scores of each faculty member were determined using the Google Scholar database. Subsequently, individual faculty h-index scores were aggregated to create mean and median school level h-index scores.  Information on funding sources, regional location, year of establishment, and faculty demographics was collected to better understand the differences in h-Index between these diverse social work doctoral programs. A hierarchical regression analysis was conducted. The first model assessed the variance in h-index scores explained by faculty differences (proportion of men to women, proportion of senior faculty (i.e., full professors and associate professors to assistant professors). The second model added school level variables, including faculty size (i.e., number of tenure-track faculty), college age (i.e., years since college or university was established), auspice (i.e., private or public), and region of the United States (i.e., Northeast, Southeast, Midwest, West).

Results:  The mean h-Index of all tenure-track faculty across the 76 schools was 11.10 (SD = 4.99, Mdn = 9.00), the average faculty size was 22.34 (SD = 10.16, Mdn = 21.00), and the average years of experience was 17.34 (SD = 4.69, Mdn = 16.00). The faculty characteristics in the first model of the hierarchical regression explained 47% of the variance in h-Index scores. The final model, adding in school level characteristics, explained a total of 51% of the variance in h-Index scores (ΔR² = .06). Academic rank had the largest affect on school h-Index. Each school’s faculty size, gender proportion, region, college age, and auspice also contributed to the predictive power of the model. The proportion of senior faculty (Associate Professors and Full Professors) and college age were the strongest predictors based on standardized regression coefficients.

Conclusions:  The finding that academic rank contributed the most variance to the regression model provides empirical support to the long argued importance of publication in career advancement. The overall results of the model confirm that institutional factors such as faculty size, region, and auspice do have unique effects on research productivity even after accounting for individual level differences in faculty across diverse social work programs. Although aggregate h-indices are not inflated by sample size, there was a positive relationship between faculty size and h-index. Why Western and Midwestern programs outperform their Northeastern and Southeastern counterparts is unclear.