Top 5 Fridays! 5 Types of Bias Involved with Research

5 Types of Bias Involved with Research

Bias can be defined as "a difference in a particular direction between the results of a study and what happens in real life."

Research cannot fully control all types of bias, but a well-designed study will always examine the effects certain biases had on their results.

Types of Bias by Cameron Faller

👉Selection Bias - This type of bias involves the selection of participants, or groups, based off from who will likely respond well in which proper randomization will never be achieved (i.e. measuring the effect exercise has on blood pressure for a group of individuals who love to exercise and will also eat healthier options)

👉Performance Bias - Participants are more likely to perform better when they know that the task being observed is expected to have an outcome. This ultimately undermines the capacity to make valid inferences about the effectiveness of an intervention.

👉Detection Bias - Detection bias occurs when a therapist or assessors has knowledge of the treatment a participant receives and may subconsciously interpret outcomes that favor the intervention group. Studies are able to control for this by blinding the assessor or therapist when measuring outcomes

👉Attrition Bias - No matter how well a study is run, there will always be participants who drop out. This skews the data especially when a greater number of participants dropped out of one group compared to another. Having a follow-up percentage of at least 85%, or obtaining an intention-to-treat analysis are ways studies can help decrease this bias.

👉Confirmation Bias - Confirmation bias occurs by interpreting the evidence or running analyses until the findings support current beliefs. This can occur both when conducting research, or reading the results of specific research.

Being able to assess the risk of bias helps give more or less weight to the value of the evidence.

Taking evidence from a study without assessing any risk of bias leads to the potential of providing inaccurate advice or expectations.

👬Know someone who might benefit from this information? Tag or Share!

“The Devil is in the Details”

Thanks to Dr. Cam Faller - reposted with permission

Learn more online - new online discussion group included!

Want an approach that enhances your existing evaluation and treatment? No commercial model gives you THE answer. You need an approach that blends the modern with the old school.