 |
| Thought Monkey |
As a
medical graduate stepping into the medical work space I am no professional on
this topic but I think it would be helpful for me to share my knowledge on
research bias to spread the awareness and be a catalyst for deeper
understanding of the topic to our fellow readers. Shout-out to @medlifecrisis
on YouTube for inspiring this article.
What is
bias?
Bias
happens when a disproportionate weight is given to or against something or
someone. It can happen by intention or be unintentional. Without proper
knowledge and intervention to minimize biases, the results of a study might be
flawed and inaccurate. Even so, we might not be able to truly avoid all the
known biases depending on the type of study and the subjects involved.
Types of
Bias
1.
Confirmation
bias
Confirmation bias happens when an individual search for or favour
information that supports their belief or hypothesis with disregard of
contradictory evidence. This is not only prevalent in research but in our
everyday lives as well. For example, I believe the cough I have been having
recently that won’t go away is a sign of me getting lung cancer. Without
knowing better I go online and search “does cough cause cancer”. Sure enough, I
found out that cough can be a sign of cancer and I start to panic even though
there are dozens of other reasons why a person coughs and cough was never a
strong indication of cancer in the first place. While this is an oversimplified
example of confirmation bias, it is not uncommon to see unethical use of confirmation
bias to ignore the opposing facts and justify an intended outcome in research
often for financial gain.
2.
Selection
bias
Selection bias occurs when a sample population selected for a study is
not truly random. An example would be a study looking at the number of smokers
in Malaysia and the volunteers for the study recruited online. Despite having a
large sample size, the result at the end does not reflect the whole Malaysian population
because only people living near urban areas have access to the internet and
participated. This leads to a study population that mostly stays in cities and
cannot be generalized to the population of the whole of Malaysia.
3.
Confounding
Confounding occurs when other factors that are not part of the study
affects the result. This can happen when potential affecting factors are not considered
or properly eliminated when designing the study. It is most common in
observational studies where the effects of a particular intervention are only
observed. While it may seem as though the intervention is causing a change in
the study population, the real cause might be other confounding factors not
taken into consideration and is not actively controlled or eliminated. A good
thing to remember is that correlation does not always imply causation.
4.
Recall
bias
Recall bias is as what it says, a bias that arises when participants are
asked to recall information from the past. It is often used in retrospective
studies where a particular incident has already occurred and researchers are
looking backwards to look for the source.
A good but cheeky example given by @medlifecrisis is that when we ourselves
are not able to clearly remember our lunch from yesterday, how are we able to
accurately recall information often months or years in the past for a study?
5.
Blind
spot bias
Blind spot bias is a type of cognitive bias where a person recognises
the impact of biases on other’s judgement but not own their own. Whether it’s
because of increased self-esteem or a lack of introspection, we must remind
ourselves that we are just as easy to fall into the trap of biases because bias
often occurs without conscious intent. Only when we actively seek or avoid them
do they become obvious.
6.
Publication
bias
Publication bias work in a situation where studies with positive
findings (ie the result supports the hypothesis) are more likely to be
accepted, published or receive news coverage. This might lead researchers to
either disregard their negative results or lose interest in the study as a
whole because of the negative results. But, both positive and negative findings
hold equal weight and are equally as important for the advancement of knowledge
in all fields. New knowledge must be tested and old outdated incorrect information
needs to be disproved to avoid misconceptions.
This is far
from an exhaustive list of research bias but I hope this article might have
sparked an interest in the topic for you.
