Reduce Bias
Ensure the research avoids research bias.
Research bias is where the research itself, rather than the research subject, causes the findings. Since the aim of research is to seek answers from real world data, research bias is like "noise" that clouds an incoming signal. The noise can be an observer effect, an interpretation effect, or an intentional effect (the researcher dishonestly fabricates data). For human subjects bias can involve:
Biosocial effects. Of researcher gender, age, race etc., e. g. male researchers may give different results from females.
Psychosocial effects. Due to researcher personality or attitude, e. g. pleasant researchers may get different results.
Situational effects. Including "good subject" effects, where subjects try to please the researcher.
Expectancy effects. Simply expecting an effect can produce it, as a self-fulfilling prophecy.
A lot of research "good practice" aims to reduce bias:
- Use a research design that controls for it.
- Standardize and/or limit researcher-subject interaction
- Use an expectancy control group
- Encourage subject honesty -ask them to be honest.
- Make the research non-threatening, e.g. let subjects be anonymous
- Keep subjects "blind" to what is expected.
- Keep researchers "blind" to what is expected
- Use more than one researcher
It is not possible to reduce all bias, but knowing how often in the past we have fooled ourselves into believing what is untrue, we should always try to reduce bias in research.
Tags: Method
Example(s)
(Use a descriptive name, e. g. "ITExample". Or click on an existing collection and edit it.)