The problem of citation impact assessments for recent publication years in institutional evaluations

•Recent publication years cannot (should not) be included in a bibliometric analysis measuring citation impact.•Using example data, the study presents various options for dealing with this problem in statistical analysis.•One option is to show a line for citation impact regressed on ‘distant’ years showing the trend for ‘recent’ years.•Another way of dealing with the problem is to work with the concept of samples and populations.•The third option is the application of the counterfactual concept of causality.

Bibliometrics has become an indispensable tool in the evaluation of institutions (in the natural and life sciences). An evaluation report without bibliometric data has become a rarity. However, evaluations are often required to measure the citation impact of publications in very recent years in particular. As a citation analysis is only meaningful for publications for which a citation window of at least three years is guaranteed, very recent years cannot (should not) be included in the analysis. This study presents various options for dealing with this problem in statistical analysis. The publications from two universities from 2000 to 2011 are used as a sample dataset (n = 2652, univ 1 = 1484 and univ 2 = 1168). One option is to show the citation impact data (percentiles) in a graphic and to use a line for percentiles regressed on ‘distant’ publication years (with confidence interval) showing the trend for the ‘very recent’ publication years. Another way of dealing with the problem is to work with the concept of samples and populations. The third option (very related to the second) is the application of the counterfactual concept of causality.