Types of bias, continued

 

Below is a quick chart defining each type of bias. Don’t worry, we’ll go into more detail later in the course.

 

Type of bias What it means Why this bias might alter a study’s results
Selection bias
The bias that occurs when the process used to select the study groups results in groups that are different from each other.   If the study groups are different, for example, if one group is older and sicker than the other, then it’s difficult to tell if it was the treatment or those personal characteristics that affected the study results.
Performance bias
The bias that occurs when study groups receive different types of care, or are exposed to other factors, beyond the treatment that the researchers are trying to study. To truly compare a group of people who receive a treatment and a similar group of people who don’t, researchers need to make sure that those groups aren’t exposed to other factors, such as other treatments, that might make that comparison more complicated.
Detection bias
The bias that occurs when outcomes for the treatment and the control groups are measured or collected differently. Different measurement techniques make it difficult to tell if the outcomes for the treatment groups were actually different or just different because of the way the data were collected.
Attrition bias
The bias that occurs when there’s a meaningful difference between the characteristics of people who drop out of a study (or have missing data), versus those who stay in the study (or don’t have missing data. If the people who dropped out of the study are different from the people who remained, a study might over or underestimate the result of the treatment because the people who had terrible results, or no results at all, stopped participating.    
Reporting bias
The bias that occurs when study authors pick and choose the information they report. If study researchers only report the information they want to share and not the full set of information, they’re shaping the results of the study and not presenting the full picture of what they found.
Other biases
These are biases that don’t fit into the previous  categories. Some are specific to certain study techniques. Others describe bias that can happen when you combine all of your studies into a systematic review. Depends on the specific type of bias.

 

Now let’s talk about how to look for these types of bias in your studies.