home-icon
  • checkmark What are violations of research integrity?
  • checkmark Threats to research integrity
  • checkmark Fabrication of research results
  • checkmark Falsification of research results
  • checkmark Plagiarism in research
  • checkmark Other Unacceptable Practices
  • checkmark How to deal with violations of research integrity
  • checkmark References for module 4 - Violations of research integrity

Fabrication of research results

jumping-icon base

Fabrication of research results

mindthegap

The ALLEA Code defines fabrication as: making up data or results and recording them as if they were real.

Other codes provide some more detail on what may be involved in fabrication or differ according to local culture:

‘Fabrication is the construction and/or addition of data, observations, or characterizations that never occured in the gathering of data or running of experiments. Fabrication can occur when ‘filling out’ the rest of experiment runs, for example. Claims about results need to be made on complete data sets (as is normally assumed), where claims made based on incomplete or assumed results is a form of fabrication’. (Erich W. Schienke, definition copied from websitePennState) 

mindthegap

Bias

Fabrication may be related to research bias in some of the following ways:

  • Fabrication may be (a symptom of) an extreme version of confirmation bias: researchers may become so convinced that their hypothesis is correct that they may fabricate results to support it (or fabricate initial results to enable further research).
  • Hot stuff bias: failure to detect fabrication may result from an uncritical approach to ‘hot’ topics on the part of journal editors, peer reviewers, or funders.
  • Positive results bias: fabrication may be more likely to arise because of pressure to publish positive results.

Impact

As with other forms of misconduct, fabrication damages science because it distorts the scientific record. In particular, by introducing results that have no empirical basis. Fabrication involves fraudulent behaviour (obtaining and misusing financial resources), and costs in terms of waste of efforts of peer reviewers and publishers, of other researchers involved in the project who were not aware of the fabrication, and of researchers who subsequently take up the results in further research.

mindthegap

Some high profiles cases:

Senior Researcher

Senior researchers may be involved by fabricating results themselves, by putting junior researchers under pressure to fabricate results, or by failing to take steps to prevent others from fabricating results. Conversely, senior researchers play a key role in preventing or detecting cases of fabrication, both in their own work and in the work that they are supervising.

Junior Researcher - Phd student

Junior researchers may be involved in fabrication either by fabricating results themselves or by facing pressure to fabricate results. They may also play an important role in preventing fabrication by following correct practices, and may become involved in reports or investigations of cases of fabrication. In the ‘Stapel’ case, Diederik Stapel’s PhD researchers detected his fabrication, and reported it to their university.

Funders - Journals Journals - Publishers

Journal editors and peer reviewers are often on the front line in detecting fabrication. They are among the first people to see publications in their final version, and they bear responsibility for setting up protocols to try and detect fabrication.

Other Researchers

Many cases of fabrication have been detected by researchers carrying out reviews of their colleagues’ work. For example, John Carlisle (2017) analysed demographic factors of (supposedly) randomised clinical trials, using statistical features of the demographics to show that the distribution of the factors could not be random, and that there was a possibility that the results of the trials were fabricated.

mindthegap

Detection doesn’t always stop publishing

The following sometimes happens when fabrication is detected by journals. The authors first submit an article with fabricated data to a higher profile journal. The article is rejected, possibly because the fabrication is detected by the peer reviewers. The authors then submit the article to a less vigilant journal, possibly with changes to hide the fabrication. Because the article is published in a less vigilant journal, there might be less chance that the fabrication will be detected (but the authors still get credit for the article). 

mindthegap

When to think about this?

Fabrication can appear at almost any stage of the research lifecycle, up to the point that the research is actually submitted for publication: 

  • Poor initial planning of data management and unrealistic promises at the proposal stage may increase the temptation to fabricate results at later stages in the project. 
  • Formulation of research questions and goals may influence the temptation to fabricate data: if such questions are poorly or unrealistically formulated, this may increase the temptation and opportunity to fabricate results. Poor project management structures may also increase the temptation and opportunity to fabricate results: practices such as well-developed plans for data collection and clear record keeping can prevent fabrication. 
  • Data gathering and analysis phase: this is obviously the phase at which data fabrication is most likely to occur. 
  • Fabrication may also arise at the write-up phase if researchers are tempted to fabricate data to improve the (apparent) results of their research. 
Collaborating

Collaborations may reduce the risk of fabrication occurring: collaborative working increases openness and scrutiny. Several of the most high-profile cases of fabrication involved senior researchers who appear to have worked alone when fabricating data, and only involved others in the analysis of their results.

Local collaborations

Local collaborations require vigilance to reduce the risk of fabrication of research results occurring. Clear and transparent plans for data storage and management will help reduce the risk. Open data practices are also a positive approach: you can contact your university’s data management service or team for advice on protocols.

Interdisciplinary collaborations

On the one hand, interdisciplinary collaboration may pose additional risks for fabrication, since researchers working together from different fields may not be able to recognise the anomalies that are indicators that research has been fabricated. On the other hand, approaches such as triangulation (using mixed methods to approach a single research question) may help prevent fabrication.

International collaborations

The distance between researchers in international collaborations may make it harder to monitor data production. However, data management and open data practices may help.