Lecture #19 - Bias Pt 1 Flashcards
Sampling error is a form of what error?
Random error
What is the definition of bias?
Following on from this, what is the difference between random error and systematic error?
“…..any systematic error in an epidemiological study that results in an incorrect estimate of the association between exposure and risk of disease.”
Random error varies - larger sample or more time to take the results means decrease random error. It has no pattern
Systematic error has a pattern to it - won’t change no matter how many times you do study or no matter sample size - bias to do with systematic
Why does bias matter?
-
Findings can be what? (3)
Important when? (4)
-
What can we do about bias?
- What two bias can ONLY be controlled during the design and data collection phases of study?
- So at the planning stages stages of a study, investigators must:
- identify what two things?
-
What are the three types of bias?
-
Selection bias
- Definition
- So when thinking about who’s in the study; they need to be _____/______ then have to _____ to participate and also _____ in study if it goes overtime
What are three main ways of recruiting people? What are their respective explanations?
Does everyone agree to participate - why would you try to maximise response rate?
Does everyone remain in the study (3) Why would you try to minimise loss to follow up?
-
How would you minimise loss to follow up? (3)
-
Selection Bias - Cross Sectional studies
- What is a cross-sectional study and what does it measure?
- Considering selection bias: what might we wanna ask ourselves? (3) We’re interested in the people included in the study
- “Is the sample representative of the population” - what 4 kinds of samples and high……(is there a cut off for high/low and consider what if low)?
- If the sample wasn’t representative of the source population - is the prevalence likely to be what three things?
-
Selection bias: Case control studies
- Explain what a case-control study is and give 4 advantages and 4 disadvantages
- How does selection bias relate to here? Like what part of the definition is relevant? I.e. what aspect of case-control studies makes them susceptible to selection bias?
- Participants are selected on the basis of their outcome status (exposure has already occurred). So if selection is in some way…..so the selection of both MUST be independent of what?
-
Selection bias: Case control studies (potential biases in selection of cases)
- Identification or participation of cases if……..
- How to minimise? (3)
-
Selection bias: Case control studies (potential biases in selection of controls)
- Controls not _____ of the which population?
- ensures controls are from the…… - Another reason bias can be introduced - how to avoid?
- Another reason bias can be introduced - how to avoid?
Go check out the example she gave
-
What are the three main general approaches to effects of bias?
Like three questions to ask
-
Selection bias: cohort studies
- How does this study work and what are the four strengths and 5 limitations?
- Selection bias - loss to follow up
- study can go and people might not……
- If loss to follow-up is related to the _____ and the _____ then it can lead to bias? Why do we try to minimise it?
Go have a look at example
What is healthy worker effect? Form of what?
-
Selection bias: historical cohort studies
- Exposure and outcome have already occurred so if classification of participants based on exposure status is dependent on the _____ then potential source of selection bias
-
