Predicting scholars' scientific impact

Abstract

We tested the underlying assumption that citation counts are reliable predictors of future success, analyzing complete citation data on the careers of ~150,000 scientists. Our results show that i) among all citation indicators, the annual citations at the time of prediction is the best predictor of future citations, ii) future citations of a scientist's published papers can be predicted accurately (r(2) = 0.80 for a 1-year prediction, P<0.001) but iii) future citations of future work are hardly predictable.

Figure 1. Histograms of a) number of papers per decade and b) number of starting careers per decade.

The publication year of a scientist's first paper is considered as the starting year of her career.

Figure 2. Effect of removing authors with ambiguous names.

a) Cumulative distribution of the number of occurrences of family names in the author lists of distinct papers before the removal. b) Cumulative distribution of number of papers per scientists after removing ambiguous names.

Figure 3. A schematic career for a scientist with

papers . We consider her career from her first paper . At prediction point , we estimate the citations received in of both past papers ( and ), and of future papers published in (). Paper is a future paper which is not published in time-window , and therefore excluded for the time-windows as defined by and .

Figure 4. Explained variance of future citations.

Future citations of published papers (bottom) and of future papers in , , and subsequent years (marked with paper selection time-windows in top ) for to years after the time of prediction were estimated. Explained variance by annual citations () in black; Extra explained variance by including the remaining indicators in red.