Statisticians are full part of the research team and contribute to the trustworthyness of your work. The are often doing more than blending figures: :quantitative researchers or qualitative researchers who validate, analyse and interprete the data collected, in the form of numbers or any other form of documentary, field observation, interviews, artefacts should certainly be included in the list of authors.

I do believe that statisticians of medical papers must be co-authors. Statistics is part of the science presented in the article.

Statisticians who contribute meaningfully to a project should certainly be authors. The basis of the acceptance of a quantitative article is in part the quality of the statistical analysis and the demonstrated statistical significance. The knowledge of that statistician - which test to use, what statistical process to implement, and what that all means - is an integral part of a journal article. I always tell other new statisticians in the field to be an active member of their research team, and provide thorough interpretations of the results to underscore the fact that they are valuable research minds, not proxies for computer programs.

Statisticians are full part of the research team and contribute to the trustworthyness of your work. The are often doing more than blending figures: :quantitative researchers or qualitative researchers who validate, analyse and interprete the data collected, in the form of numbers or any other form of documentary, field observation, interviews, artefacts should certainly be included in the list of authors.

Definitely, they are integral part of the research team and thus qualifies as Co-authors.

Absolutely. Without the statistical analysis the manuscript is unpublishable.

Yes, of course. If they are contributing to the statistical analysis presented in the data that is going in the manuscript, they should be given authorship. Anyone who contributes to the paper significantly should be given authorship. However, in few cases where they work as fee-for-service core facilities, they are just acknowledged. Again, this depends on how much work contributes to the paper can also decide if they should be included in authors list. In most cases, they should.

Dear Dr. Ahed Alkhatib,

Definitely, I am not talking about the people who do work in offices by telling 300$ for statistical analysis of the research paper, 600$ for statistical analysis of thesis. I am talking about the statistician who is working in a medical college as a Professor or faculty and analyzing the data and interpreting freely for the personal research papers of each faculty. Rather than wasting time like this for other people work, he can analyse and write his on papers, where he will get first author rather that not getting any authorship. In my paper anybody do free statistical analysis definitely i will give them authorship. Every discipline has its own qualified people. Non statistician is analyzing the data will be always like a non medical person is doing a Cardiothoracic Surgery. So, rather than taking the credit of the statistician who really done the work, keep him as a co author.

Yes, Dr. Alex Russell,

What ever you have told is correct. Study design part also researcher should take the suggestion of a statistician otherwise there will be several problems related to randomization, sample size, confounding, etc.

Statistician is full part of the research teamI if he is working as a contributor and analyzing the data and interpreting freely then he is co-author.But if the statistician receives money to do the statistical analysis he is not an co-author

Anyone doing statistical work to help the paper should be listed as coauthor. Any meaningful work should be given coauthorship.

@Alex - good points. I would just add one thing - and that is about 'negotiation' and 'ground-rules'. Statisticians, as a rule of thumb in my mind, should always establish the parameters and, if possible, obtain some form of formal contract early on. This will include their inclusion status in any publications (i.e. their place in the author order) - depending on their 'share' of input. I would also advise statisticians, as much as possible, to be included in wider aspects of the manuscript development - such as volunteering to proof-check etc

Absolutely. It's quite sad, the researchers in the medical field are still unaware of this fact, especially in India.

Of course, they should be included as a part of the research team. Without them the interpretation of the work cann't be done.

Without a doubt!

Thank you for prompting this very useful discussion. I think there is a timely and important, but neglected, discussion to be had here. It is much easier to progress with a researcher hat on than a statistician hat, while in reality statistician work is an integral part of research, especially medical research. At this day and age, statisticians are not just providing services to others following their guidance; they are often in charge of adopting complex methodologies to address pitfalls in data and sometimes in the original design advising on future methodological issues when repeating certain research. It is very shameful then that the suggestion of exclusion from publication is on the table in the first place.

A knowledgeable and experienced statistician has a skill set that is vital to the development of a peer-reviewed paper. If brought on board during the funding phase (as should be done), he or she addresses issues of study design, analytical methodology, and power estimation. During data acquisition, the statistician can monitor and identify data quality issues. In the analysis phase, the statistician conducts exploratory analysis to verify that initial distributional assumptions are still reasonable, recommends modifications to the analytical approach when appropriate, and conducts the final analysis -- with tables and graphics. Then, he/she writes the statatistical methodology section of the paper, and either interprets the results for the P.I., or writes the results section. And lastly, the statistician reviews the manuscript for appropriate integration of the findings, with respect to what one can and cannot conclude from them.

Even if a statistician does only HALF of the tasks above, he or she will have contributed substantially to a manuscript's development, and is, by rights, an author of that manuscript.

Failure to acknowledge such contribution with authorship is simple wrong.

Further: I find it very disheartening when researchers treat statisticians as wind-up toys, that you take out of the box, play with for a while, and then put aside. It's frustrating on so many levels that I don't know where to start.

It is so obvious that the statistician (or in any case the person that perform the data analysis MUST be in the author list, dear Brijesh that your question puzzles me very much. This puzzling does not stem from the fear this could happen here in Italy (otherwise I must consider my career ended when in fact I'surey overvalued by my engagement in statistics that makes me to appear in decidedly too many papers...)) ) but on how it is possible that in your country (the home of some of the most great geniuses in statistical methodologies like dr. Rao and many others) the data analysis work has so low a prestige..could you please explain us how it is possible ? Is it a generalized use, or it happens only in few places ? I'm sorry for my curiosity....

The question of authorship of academic papers is not without contention. In response to the original question, authorship SHOULD be related to a clear and distinct contribution towards the article - in the case of the statistician, this would be in the analysis and interpretation of results. The roles played by others may not be so clear cut, however everyone whose name appears as an author should at least have made some contribution towards the design or conduct of the study and been part of the process by which the findings are interpreted. Everyone should have commented upon the draft text of the paper, and indeed most journals now require this to be formally declared, together with an outline of their relative individual contributions. The extent to which all publications comply with these guidelines is less certain. Many academics will be aware of cases where authorship of papers may be gifted rather than earned, or indeed where authorship is omitted where it should have been given. Suggestions as to how we might continue to address these issues would be welcomed!

Absolutely. Your research is based on your statistical analysis. People who just contribute even one idea should be included. I am on statistics papers with 10 c0-authors. The issue is who is the first author. Then, the leader of the research group and the group decide on whose contributions should be first, second or last. Once your research self-efficacy increases you will not be so concerned with sharing the light that the research brings. My research professor is famous for statistics. He always insists on being last and includes anyone who contributed. This is Dr. Jeff Kromrey, USF research and measurement. He is the most humble man I have ever met. I have met others who have lower self-efficacy and have to always be first author even over their own Doctoral students. When you look at your life, research and impact on others; then you will see the important piece is the incremental advance of your research. Once you get over worrying about who is better, etc. then you will be able to be a part of a lot more research. Remember, the laws of science and Newton's law of what you put out you get back. There is an equal and opposite reaction for every action (Newton's third law of Physics). Finally, there are other science laws that also support the biblical law that "Give and it shall be given unto you." Whether through science or religion., forget if you are first or second and focus on your value in the research you are sharing with your field. Human beings all live a short while relatively, however your research will have a potential 100 lives impact. Share your work for altruistic reasons and let all good things come to you. I guarantee you will feel happier when you give someone else a chance to be first author. The first time it is understandable, after that realize your job and impact will be far greater as a helper than as a first author.

Best:)

Patrice Your motivational statistics educator!

Hear, hear, a wee ode

Re the authorship code

For those prepared to analyze

Then there should be no surprise

To find themselves significant

As for any participant!

Whenever a research article is written, it is based on the results and the findings. In any research study the person who does these analysis and prepare the tables or the outputs that we put under the result section of the report or research article has equal right to be one of the coauthors in the publication. Wether it is statistican or any other person he must be included. Statistician plays a vital role from the begining of the study till the end and he/she should be included in the study publication.

To bring in the statistician early in the planning of the research project often adds important views concerning the research procedure, and of course as contributors of both planning and analysis the statistican should be co-author.

It can be complicated...if a statistician is "contracted" and paid from a grant and produces the results that are then organized and interpreted by a research team and brought into a report and research papers, then it may or may not be appropriate to include the statistician as a formal author of the paper. Contracted work that is paid creates a different dynamic in terms of contribution and then authorship.

If the statistician is part of the research team in a more comprehensive way - adding to the design and interpretation and discussion of the work, then it would be appropriate to bring the statistician in as an author (whether or not their position in the team is paid). Sometimes the contribution is contractual, paid or unpaid, and very peripheral to the overall design, implementation, analysis and write up...and sometimes the stats are central to the analysis and a (paid or unpaid) stats report is so complete and developed that it IS the analysis and needs credit as such,,,so it depends of the level of contribution.

Many of the journals my teams are submitting to explicitly ask for the roles and contributions of listed authors; on that reflection, making these decisions about authorship and order of authorship depends on the contribution to the research design/implementation and the execution of the paper. Very good converstation to have here and within our research teams...thanks everyone!

I don't buy the contract-worker argument. We're not talking about peripheral, incidental involvement here. If the statistician has made material, hands-on contribution to the development of the manuscript, then authorship is called for. The method of financial compensation - salaried staff, subcontractor, what have you - is irrelevant.

@Mary, I totally agree about the contract not being the main decisive point here. In the end, almost everybody on the project is on a contract, right? The PhD student and the Post Doc have their contract on this research money, and it is (often) just a different type of contract which covers the data analysis - no matter how comprehensive. Nobody discusses the contribution of PhD students and post docs - and subsequently their authorship - on a paper just because they get paid for their work - where should be the difference to statisticians? If statisticians should not get paid when they are co-authors, so should the PhD students and post docs have to give back money for every paper they are listed on as co-authors - ???

statistics is an important part of the research so is the statistician. Research is a team work so all the key personnel should be mentioned.

if statistician is hired by institution and paid for analysis - then, perhaps, not (should not be included as an author) - it is his/her job, done for the salary. Otherwise - yes, statistician should become a co-author.

Its nothing related to the payment....we all get paid fo our work and working on research paper is not the main work its part of our job which enhances our knwoeldge and provide us a platfor where we can share our work with broader audinece in research field. Statistician who is like one of the main key person of any study or research paper should have full right to get an authorship in this kind of work and if someone thinks that doing analysis is his job and should not get credit of the support that he/she has extended then i would suggest that researcher should also get paid on the basis of research articles he/she publishes in a month or year.

Dr. Alex Russell,

Yes, I am talking about the same. It is applicable for the statistician working in medical college. Researchers will be telling him it is a 10 minutes work do it. Statistician will be spending a lot of time for this work. Finally no money no credit. The actual rule is statistician has to just give the suggestions whether tests or methods used are right or wrong as a guide and resource person for every PG students and faculties.

@Alex

Let me explain. First of all, maybe we do not understand situation from the same positions.

In my country, statisticians are not working in biomedical labs or institutions. So, if there is exception - then this person get his/her salary for making routine tasks - statistical calculations. Salary is paid from the budget. We do not include laboratory workers as co-authors, do you? They also are making their tasks, and getting salary from the budget. Recently I read Requirements for authors in one of the leading journals in my field: person, who presented material is not a co-author; person, who did technical task, such as lab analysis - is not a co-author. then, why statistician, who did routine calculations, is a co-author?

Case No 2 is different: for the new project I plan to employ statistician; he helps to design sampling scheme, he proposes methods of analysis, etc. And, lastly, he maybe uses his/her knowledge to propose new ways of analysis - then of course he is co-author, and honoured co-author of the paper, yielded from the project, irrelevant if he is paid or not.

Case No 3. After the publication you get e-mail, where statistician propose you a new method to analyse your data. In the new publication he may be the first author, irrelevant if he is paid or not.

So, please, do not make generalization about the situation in your countries before we check details. Is statistician in your case scientist? Or he is the technical worker, making routine tasks? I see big difference here, but maybe you not?

Are you going to include all persons, who hjelped you in the field, as co-authors? How about those who made figures for your paper? Is there a difference if you have such service in your institution, or you ask someole fluent in Photoshop, to spend three days doing work with veryy specific requirements?

@Brijesh

are you statistician working in College of Medical Sciences? I myself worked with some good surgeons, having no ideas about statistical methods. I newer pretended to became their co-author.

@Brijesh Satiyan, o.k., I still find it hard to imagine, but the case you describe is really a bit special - a recommendation could indeed be made from the top of the head within 10 min. or less and just a recommendation might not be considered a crucial contribution to a paper for which the decisions regarding data analyses might be made based on a number of such recommendations. However, as soon as you spend hours on the analysis, as you then go on to describe, that could become a very crucial part of the whole endeavour, and in that case, it could go two ways. I agree with Linas Balciauskas that technicians are NOT ALWAYS included on the author list (sometimes they are, maybe only every third paper or so, but still credited with authorship). Usually, technicians contribute standard techniques to this project as well as to others and USUALLY, they are on indefinite positions, regardless of the funding for a particular project. However, if the technician provides and implements some technical advice which is crucial to the success of the project, I would strongly argue for his place on the author list. If then a statistician just applies the standard analysis to yet another data set of the usual kind, then I agree with Linas that you might consider him a data technician doing his routine. However, I would hazard a guess that usually, the statistician has to cope with a new type of data set (if he was not involved in study design which in itself is a new technical contribution), and has to come up with informed knowledgable ideas how to treat these numbers in a specialized way. Maybe one could cut it down to: new routes (new techniques implemented by the technician, new analysis type by the statistician), warrant authorship. What do you think?

I agree, Susanne, as in all roles on a team, the level of contribution determines authorship. We've had statisticians (contracted/paid or simply as part of their faculty roles) who have been added as authors...generally if not it is because they were contracted, completed the discreet area of the contract without contribution to the overall design or deeper levels of interpretation and write up. In all cases the authorship is discussed in advance and re-evaluated as the projects developed. It has not been problematic, but could have been if expectations were there that everyone would get authorship or that those contracted (statisticians, technicians/data organizers, etc) somehow don't automatically get authorship...it is a conversation to have and negotiations to make along the way.

Sonya, statisticians don't do their job in a vacuum, nor is analysis as simple as you seem to suggest --- unless you're referring to the simplest descriptive analyses, involving little more than means, variances, and percentages, with no hypothesis testing. Statistical modeling, on the other hand, can be quite complex. Too often, statisticians don't contribute to the design of the study because they're not consulted at that time. And I dispute the notion that deeper levels of interpretation are possible without a statistician's input in an evidence-based analysis.

Agreed...the conversation has to happen all along a project. I think that is another issue that is emerging here - how much and at what point is the statistician involved - in terms of determining authorship.

If statistician is involved from beginning, sample planning or experiment scheme, and especially if his suggestions are changing plans of biologists/whoever - he of course is co-author of the following papers. If statistician is doing analysis - routine analysis - afterwards, and is paid from institution budget, he is not.

This answer supports those who do not want to give statistician;s their major role. It is an absolute joke if you are able to do 1) 2) and 3) then do not have a statistician. Do it yourself.!!

\

I do, of course.

Dear Dr. Linas Balciauskas,

I am talking for my friends sir. I have 150 published research works and i am editor of more than 10 journals including owner of Nepal Journal of Epidemiology.

Brijesh - commendable, of course, that you've raised this question on behalf of your 'hidden' colleagues; just one more forum to raise this issue and highlight the dilemma

I think there is serious confusion about who is being called a statistician.....I'm hearing people talk about some mythical person that simply conducts the analysis---I guess loads the data, runs the computer software (as directed by the "true authors" ) and then outputs the results in an organized fashion. Presumably this is all done with little thought related to the experimental design or selection of appropriate methods or interpretation of the statistical output .

In my experience I have never seen this person.....Although I have been in some situations where PI's thought this was the role of "their" statistician. Warning signs are words like ..."How long will it take to turn the crank"...."I have the report written, I just need the Stats"....."Jokes about damned liars of course"...... "Can you just take the error bars off those graphs"......

I usually wish these folks good luck and move on to work with people interested in science......

Bravo, John Kern!

John and Alex - very good points. You may have different experiences than mine - and I'm not for a minute suggesting that my experiences represent 'universal' practice. However, my general dealings with clinically-based research, especially let's say medical clinical trials, have the bio-statistician very much placed as a 'number-cruncher' who is not widely consulted or involved in the overall process. They may often be a named author on publications - but usually lower down the 'pecking-order'. Maybe it all depends on the discipline you are working in.

Dean, I think discipline explains a great deal of this issue. When similar designs are implemented and repeated over and over, there is not much need for a lot of statistical forethought...and probably not many interesting contributions. That said, I'm not sure how frequently this is the case, as I'm not doing that type of work, as you noted.

As the editor, I know what is Vancouver Protocol. But, honestly, how many of you always are looking for it? This is for first.

And for the second: If I get conclusions, not foreseen by a person, who ask the help for number analysis, or if mistakes are eliminated, or if co-authorship was proposed for the analysis author cannot make by himself - it's OK. As the statistician, I always do interpretation of the results, and wrote text. But I am biologist, paid for the completely different work. So, if analysis is done by hired person, are you, guys, going to pay twice? Salary and co-authorship? Then it is matter of agreement, perhaps?

Third: if we have a GIS group, who have their task to provide data, help and analysis for biologists, and they are hired to fulfill this, are they necessary co-auhors? Then, if you sent your sample to genetical lab and they did analysis for the payment, are thay also co-authors?

Finally, I am trying to get difference between paid job, and the scientific input. I am deeply convinced, that statistician may become a crucial figure. but let me remind initial question, at least part of it: "But most of the main authors are not giving authorship to the statisticians and sometimes they just give a acknowledgement which is useless. Statisticians are not paid for the personal works, according to rules statisticians has to just check and give the suggestions not to analyse every bodies articles free, but still all will come for their analysis. Is it a good practice?. "

If I understand correctly, "statisticians" mentioned are hired by institution. Then, what should be iunderstood under "are not paid for the personal works, according to rules statisticians has to just check and give the suggestions not to analyse every bodies articles free" - they are also paid by the author of data?

If yes - no authorship; you buy labor.

LInas....I don't think it is that cut and dried.....As an independent contractor, I frequently assemble the team on which I subsequently work as a subcontractor...In these situations I may even formulate the research questions and identify the necessary discipline specific expertise I need to conduct the project...biologist, chemist, environmental engineer...

These discipline experts usually prime the work as principles through their Institutions and I'm ultimately paid as a contractor. Yet clearly in this situation I'm integral to beginning, middle and end of research process...

The Vancouver Protocol is the right way to assign authorship, and there is nothing in the protocol about contracted support....or that anyone is getting paid for that matter. Conversely, contractual position is clearly stated in the protocol as not be a criterion for or against authorship.

But living by the protocol requires trust among colleagues that contributions will be fairly and reasonably recognized.

This questions suggests that some are experiencing problems with this recognition and some statisticians believe they are getting the short end of the stick.

I think that when a person has made a valuable contribution to enabling a paper reach publication status they should be included amongst the authors. This should not stop at contribution regarding statistics but also other aspects such as chemical analysis etc. It goes without saying that the contribution should be substantial and necessary to the paper being accepted.

Is a statistician qualified to be a co author in a research paper which he has done the data analysis? Up to now I did not know about either the Conference or about this particular question . If you ask me now, the question seems very strange indeed. I maintain that any man or woman who contributed as a specialist in this or that method relevant for implementation of a particular research to its success (statistician included) is a co-author by definition - at least, of a primary publication.

If I started to exclude "stepwise" those of my colleagues whose input in a study I'd estimate as rather technical than creative I'd be a single author of a number of my publications and ... very soon nobody would be willing to work with me. What would I be able to publish next?

Only in some summarizing or self-overviewing papers I may list as authors but 2-3 names (never only my own!) mentioning all others in a footnote as important contributors to the summarized studies, independent on their being biochemists, epidemiologists or statisticians.

@John, you are right about the protocol. My situation is different: having offer about co-authorship in written form, I just discovered, that paper is already in print. What you would suggest me to do? If I send all letters to the journal Editor, some persons could suffer. But such situation is not the single or the forst one. In biology, when one runs new project and needs data/material/samples from many countries, good offers and promises are given, later, ok, depends on the conscience.

I think it depends. If some one want to test the significance of collaborator' s contribution, simply he can remove their contribution and weight his research if no value missed and his research still strong he may discard their authorship and substitute them in acknowledgement but if his work will be unacceptable so he must mention them in author's bar.  

I think that it depends on the level of contribution by the statistician. I have published a paper where the person who was largely responsible for directing the path taken by the chemical analysis was included, quite rightly, as an author.

I stay on myself:

1. If statistician is hired for the job by institution, acknowledgement could be enough.

2. If you pay statistician for his work, then co-authorship is matter of agreement.

3. If this work was done without payment - co-authorship is really worth the job done.

By the way, if you think otherwise - why you do not do statistic yourself?

Hello, Toni

so we are on the same position. If one cannot do job by himself, he need co-author or should buy results, isn't it?