Averages of ratios compared to ratios of averages: Mathematical results

The recently awakened discussion on the usability of averages of ratios (AoR) compared to ratios of averages (RoA) has led to the mathematical results in this paper. Based on the empirical results in Lariviere and Gingras (2011) we prove, under reasonable conditions, the following relations between AoR and RoA for a set of points:(i)The regression line of RoA in function of AoR is the first bisectrix.(ii)(AoR−RoA)/AoR in function of the number N of papers is a cloud of points comprised between a multiple of 1/N and −1/N.(iii)(AoR−RoA)/AoR versus RoA has a decreasing regression line.

[1]  Thed N. van Leeuwen,et al.  Towards a new crown indicator: Some theoretical considerations , 2010, J. Informetrics.

[2]  Loet Leydesdorff,et al.  Simple arithmetic versus intuitive understanding: The case of the impact factor , 2011 .

[3]  Gunnar Blom,et al.  Probability and Statistics: Theory and Applications , 1990 .

[4]  Guido Van Hooydonk,et al.  Methods for accrediting publications to authors or countries: consequences for evaluation studies , 2000 .

[5]  J. Martin van Zyl,et al.  A simulation study to investigate the accuracy of approximating averages of ratios using ratios of averages , 2013, J. Informetrics.

[6]  Leo Egghe,et al.  Average and global impact of a set of journals , 1996, Scientometrics.

[7]  Leo Egghe,et al.  Elementary statistics for effective library and information management , 2001 .

[8]  Leo Egghe,et al.  A General Framework for Relative Impact Indicators / Cadre général pour les facteurs d’impact relatifs , 2003 .

[9]  Thed N. van Leeuwen,et al.  Towards a new crown indicator: an empirical analysis , 2010, Scientometrics.

[10]  Leo Egghe,et al.  Averaging and globalising quotients of informetric and scientometric data , 1996, J. Inf. Sci..

[11]  Vincent Larivière,et al.  Averages of ratios vs. ratios of averages: An empirical analysis of four levels of aggregation , 2011, J. Informetrics.

[12]  Loet Leydesdorff,et al.  Caveats for the journal and field normalizations in the CWTS ("Leiden") evaluations of research performance , 2010, J. Informetrics.