Integrity matters

Excellence in research must also extend to standards of conduct, maintains Richard Dyer

Scientists, science funders and the general public all expect those undertaking research to possess great integrity. They expect them to make every effort to design and conduct experiments to the highest standards, to analyse data rigorously and to disseminate results accurately. This modus vivendi is often correctly described as "the pursuit of excellence".

This excellence has to be real: there is no place for any element of dishonesty. Given that integrity of a very high order is a key component of a scientist's professional life as a research worker, one can reasonably ask how it is nurtured in training and in our research institutions. The answer, I fear, is not very well.

Young scientists are taught how to use complicated equipment, make up solutions, analyse and present data. But it is largely assumed that scientific integrity can be acquired along the way and that, in any case, there is no problem of any significance.

Well, there is a problem. I shall give one example to substantiate my view that our collective level of integrity can be improved. When seeking new positions, applicants submit a CV and a list of their publications. This usually contains papers "in press" (accepted by a journal for publication), "submitted" (sent to a journal but the outcome of the review process is unknown) and "in preparation" (the applicant has yet to submit the paper).

"In press" has the highest standing and "in preparation" the lowest. This leads to a temptation to move the status of papers up the chain. It is not uncommon for a scientist to hope that by the time of an interview a paper "in preparation" will have been "submitted" and to list it as such in the application.

Furthermore, in scores of interviews with young scientists, I have never been told directly that a paper listed as "submitted" in the application has been rejected by the journal at the time of interview. Everyone knows about this murky situation. The problem is usually resolved by ignoring everything listed that has not been published. More important than the fate of a particular paper is what this widespread behaviour tells us about scientific integrity when it comes to job-hunting. When there are pressures to lower standards, standards may be lowered.

There is an excellent publication titled On Being a Scientist - Responsible Conduct in Research (National Academy Press) that we give to all new entrants to the Babraham Institute. The booklet addresses many key issues concerning scientific integrity. One case study concerns fabrication in a grant application in which a publication listed as "submitted" was in fact only in an early state of preparation. The individual is dismissed by the faculty. How many have been similarly treated - or even reprimanded - in the UK for a similar transgression?

The culture that "everything is OK: don't worry" is pervasive. Some years ago, a graduate student at Babraham made a serious written accusation about scientific fraud. An inquiry exonerated the accused. But the act of holding an inquiry brought strong external pressure from scientific collaborators "not to be silly". On one occasion when I dismissed someone for improper conduct, the scientist expressed amazement and said I was grossly overreacting.

Most scientists are cautious and careful research workers. I know very many who publish only when they are absolutely confident of the data - even at the risk of being beaten by a rival. Nevertheless, I hope that debate can address how we can better foster integrity in research institutions.

Last year, the US National Academy of Sciences reported that no established measures for assessing integrity in the research environment existed and that although the promulgation of and adherence to policies and procedures were necessary, they were not sufficient to ensure responsible conduct.

They also concluded that institutional self-assessment was a promising approach.

How effective "self-assessment" is to be undertaken is unclear. But at a time when quality assurance is high on the national science agenda, it is important to remember that any such improvements may be undermined without the highest standards of integrity in research conduct. This is a good time to discuss how to monitor and improve our standards.

Richard Dyer is director of the Babraham Institute.