Biases at Different Stages of Research
– and ways to avoid them

Bias is the conscious or unconscious influencing of a study and its findings. Bias in research has potentially far-reaching effects in terms of misinforming readers and end users of research results. Every step of research—from conception to publication—is prone to bias. Thus, it is imperative to understand the various forms and flavours of bias so that you can avoid them and produce verifiable, replicable and reliable research.

1. Biases at the planning stage

Bias may arise at the planning phase due to poor study design in the following ways:

• Sampling bias: When researchers select elements for a study that are not representative of the population from which they were drawn
• Response bias: When only certain types of subjects respond to an invitation to enter a study (similar to selection bias)

Ways to avoid biases during planning

• A well-designed research protocol, clearly outlining data collection and analysis
• Random sampling in quantitative studies
• Selective (or intentional) sampling for objective results in qualitative studies

2. Biases while conducting the study

Biases entering the study phase can be divided into participant and researcher biases.

Participant bias

In studies recruiting human participants (such as drug trials and behavioural studies), biases introduced by the participants themselves can affect the findings. Participant bias arises when participants’ responses, for a host of reasons, are deliberately or unintentionally different from their intended answer.

Here are two types of participant-introduced biases:

• Recall bias: This occurs when participants have selective recall about specific experiences or events in the past. This would be a problem in retrospective studies (rather than prospective studies).
• Observation bias: Also known as the Hawthorne effect, this bias occurs when study participants are aware that they are being observed and knowingly alter the way they behave or answer questions.

Researcher bias

As a researcher, you may unknowingly include only data that you think are relevant or interpret data to prove your hypothesis. This bias manifests in the following ways:

• Interviewer bias: When a researcher asks questions in a manner (such as in a certain order or wording) that could affect participants’ responses
• Measurement bias: When researchers measure variables without rigour
• Confirmation bias: When researchers look for patterns in their data that confirm their own ideas

Ways to avoid bias during a study

• Standardised interaction of the interviewer with respondents
• Blinded data collection
• Representation of outliers
• Independent analysis of the results by other researchers (such as a biostatistician)

3. Biases at the reporting stage

At the reporting stage, bias may be introduced by selective disclosure, suppression or distortion of information related to the study. 

• Outcome reporting bias: When only some outcomes are reported from a study
• Publication bias: Choosing what should and should not be published, e.g. when a researcher decides to withhold the reporting of negative findings (also known as the file-drawer effect)
• Spin: When researchers report their findings in a way that misleads readers into viewing the results more favourably, e.g. a title, an abstract and conclusions that exaggerate(s) middling findings

Ways to avoid bias during reporting

• Pre-registering your research in public databases
• Ensuring that the title, the abstract and conclusions present a balanced and representative summary of the results
• Submitting negative findings for publication

End note

Awareness of potential biases will help you carefully plan your study and avoid or reduce their occurrence. It is a good practice to indicate the limitations of your study and the efforts you took to avoid bias while planning, analysing and reporting.

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