The Hidden Biases in Scientific Research and Publication

Biases can be found anywhere around us, and are introduced to us early on in our lives. Right from childhood, biases are introduced at home and school, later from the media, and even in non-fiction news we acquire from credible sources. As such, scholarly articles published by the means of thorough research, editing, and review may also consist of biases. Nonetheless, publishing scientific literature is crucial to reproduce basic research and apply it to clinical situations (Hausmann et al., 2018). Thus, biases in research and publishing can create potential flaws in this transfer of credible information. 

Biases within scientific literature can arise anywhere along the process, right from the moment the author begins his research, to the editorial process, and even during publication. When an author, or a group of authors, synthesizes a scientific article, they usually include aspects, such as study design, results, data analysis, and the presentation of evidence (Hausmann et al., 2018). Within these aspects, authors will attempt to avoid any personal biases that they may carry as a result of prior research, learning, or experiences, but to what extent is this actually possible, is a question of interest even today. Such classifications of bias were grouped into the following stages of the research process by Sackett and Choi: reading papers for knowledge, selecting a study sample, conducting the experiment, measuring the variables, analyzing data and results, and the publication process (Delgado-Rodriguez & Llorca, 2004). 

When authors conduct research in the desired field of study, there are factors that researchers may ignore due to personal preconceptions, beliefs, or ideas, and this can interfere with the objective nature of the research process.  

Moreover, the process of data collection holds the risk of being exposed to biases. When researchers aim to collect data for research purposes, oftentimes they collect information from a sample of a population. This form of survey data collection can result in biases which can be grouped into selection bias, observation bias, and recall bias (Medical Research Council, n.d). When the chosen study population does not accurately represent the intended audience, selection bias occurs (Delgado-Rodriguez & Llorca, 2004). However, such errors can be limited by choosing the sample population randomly, or on the basis of chance (Medical Research Council, n.d). Self-reporting bias can interfere with observational research methods in medical or clinical studies, where data is acquired through questionnaires, surveys, or interviews (Althubaiti, 2016). To fit in with society’s expected norms or to seek approval from society, participants may feel inclined to give inaccurate answers, which is often referred to as social desirability bias (Althubaiti, 2016). Recall bias can occur when participants provide answers based on their ability to remember certain events, and this interferes with the accuracy of their responses (Althubaiti, 2016). 

Confirmation bias can emerge when researchers make decisions about their topic of interest based on preconceptions or beliefs (Althubaiti, 2016). It can also be a result of factors that researchers may ignore to include in their own work due to the lack of importance given by other researchers in the field. For instance, the research of genetic causes of lung cancer persisted for decades whilst a behavioural cause, tobacco smoking, was ignored from the research process (Pinto, 2019). This shows that prior research can steer present research in a direction where researchers become ignorant to other considerations. Specifically in medical research, confirmation bias is one of the main reasons for diagnosis errors and improper methods of treatment (Althubaiti, 2016).

Publication bias can occur when the publication of research is impacted by the strength of the results (Ayorinde et al., 2020; Franco et al., 2014). This can indicate that statistically significant data have a higher chance of being published than nonsignificant data, which results in key findings being left out in the publication process (Franco et al., 2014). This form of bias holds great importance to medical care, because the available evidence can influence professionals to make decisions that can impact the health of patients or the allocation of resources (Ayorinde et al., 2020).

It is important to keep in mind that although the resources we use for our own research purposes are well researched, and come from credible sources, they can include biases which arise at any stage of the research and publication process. Being diligent researchers means taking such biases into account and considering how to minimize them when conducting our own studies.