The most careful researcher may sometime fall into one of the various types of research bias. Readers of a research work should therefore be aware of the different types of research bias so as to be aware where the bias can creep in and where it is present. Bias in a research does not discredit the research as a whole but it is important for the reader to be aware of it.
Given below are the most common types of research bias:
In conducting a scientific study, any bias that creeps in because of error in selection of individuals or groups participating in the study is a selection bias. It causes distortion in the statistics obtained and ends up giving a false conclusion. Also known as ‘selection effect’, this bias is the result of incorrect method of sample collection. Two major causes of these types of research biasare: one, self-selection , that is when sample uses itself, and two, convenience sampling, that is when particular samples are chosen because they are easily accessible. Early termination of a trial when it supports a desired result is also a case of selection bias.
The design of the research should be appropriate for the central question of the research. If there are any errors in the accuracy of the method, these constitute the measurement types of research bias. For example, a given questionnaire that aims at assessing learning objectives of a particular training session but actually measures just the learner’s satisfaction. Measurement bias can also happen because of instrument bias. Example, an unbalanced weighingscale would skew the results of any study that uses the instrument.
Opinions of the interviewer, his/her prejudices, and personal biases may get displayed in the process of interview. This may color or influence the result of the study or may give a skewed picture of the same. These types of research bias seep in in subtle ways and can be difficult to point out if not blatant.
These types of research bias are caused when the subjects shape their responses to please the person interviewing them. The subjects may also believe that they have an idea about what the result would be and hence respond in a way so as to match the supposed results.
These types of research bias occur in the ways a study is published and disseminated. Common types of reporting bias are: publication bias, bias that occurs when a research study suffers more chances of being published because its results are significant in terms of its statistics and less likely when the statistics are not that significant, and the magnitude of the results may become a more important factor for publication than study design, relevance of the study and its quality;the time lag bias, that is when positive results are published much more rapidly; language bias, studies with positive results have more likelihood of getting published in English while negative studies are sent to non-English journals; the multiple publication bias, that is when different variations of the same study orresearch are published in a number of different journals; lastly, citation bias, wherein positive study results are more often cited by others.