Section 1: Chapter 5:
USMLE Biostatistics: Random Error, Systematic Error, Selection Bias, & Methods to Control for the Selection Bias
The different kinds of research biases are a favorite of the USMLE Step 1 & Step 2 CK. Learn everything you need to know about the main kinds of biases in this concise biostatistics lecture.
Errors in health research are extremely common and often lead clinicians to expose patients to potentially ineffective or even dangerous therapies. For this reason, it is extremely important to recognize when errors in research occur. There two main classes of errors in research; systematic error (also called bias) and chance error (also called random error. Bias can be further broken down into three types or “families” of bias; selection bias, information bias, & confounding bias. Every type of bias falls into one of these three families. Examples of selection biases include the Berkson’s bias, attrition bias, Neyman’s bias, volunteer bias, and non-response bias. Examples of information bias (also known as the measurement bias) include the observer bias, recall bias, reporting bias, surveillance bias, Hawthorne Effect, and lead-time bias. In this chapter we will discuss the following types of selection bias.
- Berskon’s Bias: Occurs when the study sample has higher rates of exposure and outcome compared to the general population. Commonly occurs when the study sample is obtained from hospitalized patients.
- Attrition Bias: Occurs when study participants leave the study or are lost to follow up before the study is finished.
- Neyman’s Bias: or prevalence-incidence bias occurs when the observation or measurement are made at a time when many subjects have already died, recovered or were never symptomatic and as a result were not included in the study sample.
- Volunteer Bias: Occurs when people who volunteer to participate in a study are systematically different from those in the target population.
- Non-Response Bias: Occurs when study participants who do not complete surveys or questionnaires are systematically different than those who do.
In this chapter we will introduce and discuss the different kinds of errors in research; systematic error and random error. We also will also introduce the selection bias and describe some important subtypes of selection bias. We will also introduce the common techniques used to prevent or control for the selection bias.
- Definition of random error (also known as chance error).
- Methods to measure and control random error in research.
- Definition of systematic error (also known as bias); what is bias and why is it important in research?
- Main types of bias: selection bias, information bias (also known as measurement bias), and confounding bias.
- Definition of the selection bias; what is the selection bias?
- Definition and examples of the most important subtypes of the selection bias; Berkson’s bias, attrition bias, Neyman’s bias, volunteer bias, non-response bias.
- Methods to control and prevent the selection bias.
- Video length: 25 min
- Practice exercises: 10 USMLE style multiple choice questions with in-depth explanations
- Estimated time to complete video lecture and practice exercises: 1 hour
- Next chapter: Chapter 6 – Types of measurement/information bias and methods to control it