The contribution of each author cannot be unambiguously determined by looking at the order of the authors. Each discipline has its own rules. Some use alphabetic ranking, others contribution-based and others a combination of both. Calculating the impact of a single author goes against the spirit of science, where the impact is generated through a collaboration of different scientists, sometimes disciplines. The only impact that is generated is the impact of the team, i.e. the set of authors and co-authors. We all know that the contribution of each author is not equal and that some of them may have not contributed at all, but there is no satisfactory general rule for attributing impact among the authors in a joint publication. Hence, total impact calculation is the next-best solution.

Some journals now require authors to explicitly state the individual contribution of each author on a publication, but as long as not all journals adopt this practice, the safest way to determine impact is total impact, but explicitly acknowledge that the impact has been generated by the team and not the individual author.

The contribution of each author cannot be unambiguously determined by looking at the order of the authors. Each discipline has its own rules. Some use alphabetic ranking, others contribution-based and others a combination of both. Calculating the impact of a single author goes against the spirit of science, where the impact is generated through a collaboration of different scientists, sometimes disciplines. The only impact that is generated is the impact of the team, i.e. the set of authors and co-authors. We all know that the contribution of each author is not equal and that some of them may have not contributed at all, but there is no satisfactory general rule for attributing impact among the authors in a joint publication. Hence, total impact calculation is the next-best solution.

Some journals now require authors to explicitly state the individual contribution of each author on a publication, but as long as not all journals adopt this practice, the safest way to determine impact is total impact, but explicitly acknowledge that the impact has been generated by the team and not the individual author.

Thank you very much for your informative comment. I do know the importance of the team work but please note that there is no specific border between being a co-author or being someone who is acknowledged in the acknowledgement. Therefore, some of our colleagues such as statisticians/methodological consultants (and sometimes our bosses!) are usually among the authors even when they have no any major contribution. By the way, this policy would make them giants to us!

And how about who is author correspondence in all his papers and final author in authors order?

I agree in part with Matty's answer. However, it does to an extent overlook the gifting of authorships. For example, a chap I used to work was highly suprised when, one day, a cheque turned up in the post for a paper which had won a prize. He didn't seem to know any details of the paper, although he had shared a few beers with his 'co-authors'!

In addition, the contribution of some authors cannot be measured in lines of text written but advice given or even just presence. I seem to remember research that shows that papers with well-known co-authors have a higher citation rank.

Well-known authors probably wouldn't want their name attached to poor-quality research, or research destined for a low rank journal, which may explain higher citation ranks.

Andrew, I assume they typically have higher visibility and readers also make the same assumptions you described.

In the perfect world the order of the co-authorship can reflect contribution of each co-author. Yes, really successful researchers probably will not recklessly put their names in any article but it maybe just what we want to believe. I had bad experience that some program directors want their name in every single artcle from the program so that they can build up accumulative impact. Researchers all face the pressure for productivity issue so some inappropriate behavior will develop. Do not blame the junior co-authors since they are just poor workhorse for the research group and try to keep themself afloat.

As a follow-up with my perevious input. I have but to say some program directors are really honest. My program director, when I was still a fellow in Alabama, refused to let me putting his name in my research paper because he did not provide too much input in my research even though he invested a lot of money for my supplies. He is really my role model.

Dear friends, I will just attach some good threads with so good answers and it will enrich this discussion!

Also, a very good paper From CiteSeer to CiteSeerX: Author rankings based on coauthorship networks is attached!

https://www.researchgate.net/post/Is_a_large_overlap_in_top_co-authorship_a_sign_of_true_collaboration_or_a_mark_of_different_fields_with_different_standards_for_paper_co-authorship

https://www.researchgate.net/post/Does_co-authorship_lead_to_more_scientific_researches

https://www.researchgate.net/post/Who_should_be_among_co-authors_of_the_scientific_paper_TOPIC_CLOSED

https://www.researchgate.net/post/Co-authorship_in_a_research_paper_and_its_effect_on_impact_factor_or_the_overall_weightage_of_the_research_paper

https://www.researchgate.net/post/Can_someone_advise_on_an_opinion_that_co-authorship_is_a_currency

Article From CiteSeer to CiteSeerX: Author rankings based on coautho...

"Out of all the ResearchGate statistics, the total impact points have the highest

correlations with the five academic ranking systems. This is unsurprising because it is based upon a traditional impact metric, the journal Impact Factor. The lower correlations for both ResearchGate views and downloads, despite being arguably more direct indicators of the value of an individual document, suggest that these statistics are a somewhat skewed

reflection of the articles' academic impact. For instance, they may reflect educational impact rather than academic impact, may be too low (particularly for older articles) to give the statistic sufficient statistical power or perhaps many downloads and views are created by students from the same institution out of curiosity. Similarly, the ResearchGate score, which incorporates both impact and activity, had low correlations, suggesting that there was not a strong association between the most active members of ResearchGate and membership of the highest ranking institutions..."

I think new matrices are required to be defined.

agreed