Donate IconDonate

Research transparency and reproducibility in research

Research transparency is a term used to describe the sharing and dissemination of research and research outcomes. It is considered good ethical practice to share research in an open and transparent way, both within the scientific community and with the wider public.  

Why is research transparency important? 

Research transparency requires that the methods, experimental design, data, analysis, evidence and conclusions from research are comprehensively reported and disseminated openly (and ideally free of charge). In scientific research, progress is often made in incremental steps building on work that has gone before. The sharing of research is therefore important not only within the scientific community but also within the public domain.  

Transparency is considered a key aspect of ethical research practice and it is important for many reasons: 

  • It informs future research questions and methodologies.  
  • It allows experimental protocols to be scrutinised (to identify misleading results due to poor experimental design or analysis).
  • It helps prevent unnecessary research studies and protects participants (and animals) from the associated harm or suffering. (By preventing the repetition of valueless research, or research that is likely to have no useful outcomes.) 
  • It helps maximise research funding and prevent wastage of time and resources – including human tissue, equipment and animals’ lives.  
  • It shares findings that can be used to inform future policy. 
  • It builds the public research evidence base. 

How can we increase research transparency? 

In the current system research is shared through the publication of papers in scientific journals. These can be published open-access (available to everyone, for free), or in closed-access journals where only those with a subscription can view the papers. Whilst these papers cannot then be published elsewhere due to copyright issues, a short abstract of the paper can be shared and searched for, and a payment is made to the publisher to receive the full article. Cost can sometimes impact transparency, as publishing a paper open access often costs an author significantly more, whereas in the traditional, paywalled publication model, the cost is charged to the reader. 

Barriers to research transparency: 

  • Publishing poorly conducted or poorly reported research. Where experimental design, methods, or analysis techniques are inadequate or have not been clearly or completely shared the methodology cannot be scrutinised. Therefore, even if the research is good quality, it cannot be replicated. Transparency in reporting is particularly key when evaluating animal research. Read more about the importance of good experimental design. 
  • Publication of negative results. Null results are those where the expected change does not happen in an experiment. Often these results are perceived as less interesting, and therefore are often not accepted, or even submitted for publication. This is a scientific flaw and directly impacts transparency as there is something to be learnt from null results. Sharing of null results is important in preventing duplicate research, evaluating reproducibility, identifying lines of research not worth further investigation and allowing resources to be directed into other lines of enquiry. This in turn reduces the likelihood of further resource (or animal) use looking at the same research question. There is recognition in the field that this is the case, with the existence of journals specifically to publish null results and a move for all journals to embrace their publication.  
  • Publication bias. Just as with null results, journals can be swayed to publish or not publish certain papers. This may be bias towards an author, an area of research, or particular methodologies, for example, animal or non-animal approaches. Whatever this bias might be if it creeps in it can affect what research gets published and has a knock effect on transparency. 
  • Publication pressure. In scientific research, there can be pressure from funders and universities on academics to get work published, and in specific, respected journals. This can lead to bias during analysis or poor reporting in a bid to get papers published. Hopefully, these cases are few and far between, but the pressure to publish is rife. Learn more about animal bias in publishing and this 2014 article ‘Publish or perish: Where are we heading?’.
Read more about the case for null and negative result publication:   
Filling in the Scientific Record: The importance of Negative and Null Results 
In praise of replication studies and null results  
Dealing with positive publication bias: Why you should really publish your negative results 

Transparency in clinical and preclinical research aids scientific progress 

Clinical trials are studies that require people to be used to assess the outcomes of a health-related intervention, such as assessing the safety or efficacy of a potential new drug. Clinical trials are now required to be added to a public register before they are carried out. This requirement was introduced to counteract the distortion of scientific evidence that happens as a result of unsuccessful trials (those where there were no clinical benefits or negative/null results) not being published or therefore shared. As mentioned earlier, knowledge can be gained from negative results as well as positive – it is critical to funding and policy decisions to know what doesn’t work as well as what does. Although the current system for registering clinical trials may not always be fast or efficient, it helps improve transparency in medical research. 

Animals are used in many preclinical trials (early drug development studies to evaluate efficacy and safety) and there is no requirement currently to publicly register or share outcomes of unpublished animal research. As with clinical trials, this lack of transparency inhibits scientific progress by allowing a range of poor practices to potentially continue unseen and means that valid negative outcomes often remain invisible. It is also a significant ethical issue with animal lives being wasted repeating research that has already been conducted. 

Current systems for increasing transparency in animal research

When research is not published, that research cannot easily be searched for, seen, scrutinised, or used to inform other studies.  

There are some systems in place to address this to increase transparency, although they are not compulsory, or widely used. 

  • Registration of research projects and protocols at the start of the study. This is currently a legal requirement for all clinical research using volunteers. There are calls for a similar mandatory system for preclinical research or animal research such as this BMJ Open Science article Publication rate in preclinical research: a plea for preregistration. There are some voluntary global databases available currently for this including Preclinical Trials and https://www.animalstudyregistry.org/ but uptake has been slow.  
  • Sharing unpublished papers. Some universities allow the uploading and storing of unpublished work that can be shared via a link. There are also online archives where ‘preprints’ (completed papers that have not yet been published) can be shared such as bioRxiv. Whilst these papers have not been ‘peer-reviewed’ as published work has; they are at least in a public forum for scientists to access. 
  • Sharing details of planned animal research projects. Legislation (ASPA) requires animal research to have other systems in place to increase transparency including the annual publication by the Home Office of statistics of procedures on animals, and the annual publication of NTS’s (Non-Technical Summaries). The NTS is a section of a project license that explains aspects of a programme of animal research written specifically to be shared publicly. The NTS describes the project, explains how the 3Rs have been implemented and justifies the need for animal use in the research.  
  • Some animal research projects also require a retrospective assessment to be carried out on completion which must be submitted to the AWERB along with an updated NTS. This is reviewed and then reported back to the Home Office. Although the retrospective analysis is not currently available publicly, if it were, this would be another step to increase transparency. 
  • The Concordat on Openness is an initiative launched by the organisation Understanding Animal Research that UK life science organisations, including universities, can sign up to. It consists of a set of commitments that the signatories agree to follow to share information on when, how and why animals are used in their research.  

Ultimately encouraging transparency in animal research will help build the evidence base to understand and challenge the usefulness of animal studies in different areas and replace it with approaches that offer more scientifically and ethically robust solutions. To this we need to strengthen and improve current systems that increase transparency in animal research and give the researchers the tools and confidence to use these systems and without abuse. 

Read more FRAME articles on transparency: Encouraging transparency in animal research | FRAME 

Reproducibility in research 

What is reproducible research? 

Research findings are “reproducible” if a study can be independently carried out using the same methods to obtain the same data and results. If research is reproducible, it is more likely that the results are valid and therefore correct.  

Can we tell if published research is reproducible? 

If research is transparent and contains all the necessary detail in the publication about how the study was conducted, it is easier to establish whether it is reproducible. Sadly, there are examples of published research, accepted by peer-reviewed journals that are not reproducible. Common causes of problems with reproducibility are the implementation of poor experimental design in a research study and poor reporting of studies both of which make it difficult to accurately repeat the study and obtain the same results.  

Why is poor reproducibility in animal research studies a concern? 

Reproducibility is a particular concern in studies that use animals. Scientific research that is not reproducible is generally considered to be of little or no value as it cannot be carried out again or ‘reproduced’. There is an ethical issue here where animals have been used as the lack of reproducibility may have been caused by an avoidable failure in study design or reporting. There is then also a risk of repeated studies in the same area to correct design flaws and attempt to identify valid conclusions, or if the lack of reproducibility is not picked up, to ask slightly different research questions based on the initial flawed results. 

The ARRIVE Guidelines provide a good practice checklist for researchers for reporting in vivo (animal) experiments to help improve transparency and reproducibility. Home | ARRIVE Guidelines 

Read more FRAME articles on reproducibility Updated ARRIVE guidelines: reproducibility and animal research | FRAME 

resources

Additional Resources

To receive updates from FRAME, please enter your details.