Introduction to Biases in Research
– and steps for avoiding them
A bias, in general, is a predictable error that sways someone’s judgment in a particular direction, favouring a certain outcome over others. In research, bias means a deviation from the truth in data collection, analysis, interpretation or publication.
Why research is prone to bias
You might introduce bias into your work consciously or subconsciously, just to prove a ‘pet theory’ or to publish exciting results. Researchers want to publish (the presence of) something rather than the absence of a finding. Moreover, conflicts of interest (COIs) – for example, research sponsored by a pharmaceutical company – could lead to biases in reporting. For instance, if a drug does not work in treating a specific disease, this finding may go unreported. However, you must remember that the scientific record is meant to represent everything that has been found or not found, and not just something exciting all the time!
How biases can affect your research
- Biases can lead to false conclusions, which might be misleading or even harmful. The use of biased results to inform further research or guide policies may have damaging consequences.
- Biased studies are not reproducible and will affect the credibility and validity of your work.
- Basing your own study on biased research will perpetuate the misrepresentation of invalid findings.
Stages of research at which biases can occur
Every stage of research, from planning to publishing, is prone to bias.
a. Reading the literature
Choosing to exclude certain literature could lead to omission bias.
b. Planning the study and collecting the data
Irrespective of whether the study is quantitative or qualitative, a flawed study design results in room for various biases to emerge, e.g. sampling and selection biases may happen as a result of non-random sampling.
c. Analysing and interpreting the data
Selective mismeasurement arising from prior preferences (confirmation bias) may lead to overestimation, skewing, etc.
d. Publishing the findings
Results not supporting a researcher’s hypothesis are often filed away, never to be published. Such selective reporting or withholding of information leads to publication bias. Then, ‘spin’ is a type of bias wherein findings are exaggerated.
Practical steps for avoiding bias in your research
Some degree of bias is inevitable. However, as a responsible researcher, you must identify potential biases and the ways to prevent them. Here are some practical measures to follow:
- When searching the literature, cast your net wide enough to include studies with divergent viewpoints, variables and results and not just the ones you think might support your proposed hypothesis.
- When planning your research, be mindful of the potential for bias at every stage of your work, particularly sampling and effect size. Have someone outside your study review your research plan and data. (For biomedical studies, it helps to have a biostatistician look at your research at various stages. Learn more here: How having a Biostatistician can help you at every stage of your research)
- Have multiple people independently analyse and evaluate your data. When interpreting your findings, make efforts to avoid overgeneralisation.
- Avoid post-hoc rationalisation, overfitting data and hypothesising after obtaining results (HARKing). Note that both positive and negative findings are valid and should be reported.
- Always keep records of all research material and raw data.
- Be transparent about any potential COIs related to your study.
Biases have ethical and practical implications in research. It is crucial to be objective and avoid bias before, during and after your research.
Maximise your publication success with Charlesworth Author Services.
Charlesworth Author Services, a trusted brand supporting the world’s leading academic publishers, institutions and authors since 1928.
To know more about our services, visit: Our Services