Scientists are traditionally evaluated from the citation numbers of
their publications (h-index). Although, the h-score has a global aspect due to
readers’ appreciations, it remains limited by the absence of (i) explicit
evaluation criteria, (ii) structured analysis of papers’ contents and (iii)
protection against biasing self-citations. Analysis of papers’ contents in relation
to their citation scores and authors’ h-values provides an integrative evaluation
way. For that, a simulation approach was developed by characterizing
published papers’ contents by different cooperation and production variables
(numbers of co-authors, countries, affiliations, figures, tables, pages) which
were subjected to correlation analyses to highlight different publication
strategies associated with different citation ranges. The simulation approach
was based on simplex constraint making working on ratios between different
papers’ components (cooperation and production variables). After simulation,
multidirectional and multiscale variations between such ratios in different
papers’ classes helped to highlight regulatory effects of production and
cooperation components a priori on citation levels a posteriori. Application was
carried out for biological papers from Tunisian researchers by working on six
classes combining three h-ranges of authors with two papers’ citation levels
(h vs h). Publication strategies and citation improvements were linked to
relative weighting between production and cooperation variables.