Error and Bias in Health Science Flashcards
Error is what?
a comparison
- error compares what we thought or believed or did with what we ought to have thought/believed/done
Errors Have size… explain
We can’t possibly eliminate all error - therefore we have to try to reason how BIG the error is to know how accurate a measurement will be
is it possible to know how much error any individual measurement contains?
no - need multiple measurements
every measurement has two components.. what are they?
measured value = true value + error
Do error occur according to a mathmatical principle?
no, errors occur at random.
that is they plus/minus errors are equally likely to occur
big errors are less likely to occur than small errors
therefore: Errors cancel out in the long run - the more data= more precision
how can precision be increased?
- more data
- reducing measurement error
how are error/precision linked?
error is part of every individual measurment
However, we can become more precise when each individual measurement is put together
under what circumstances are errors not random?
ex) people are more likely to underestimate their alcohol consumption - this is true with most social behaviours
when errors develop a patter, what is it called?
bias
what is the difference between precision and accuracy?
precision = depends on amount of error in measurement - more results increase precision
accuracy = depends on amount of bias in the measurement - more results will not increase accuracy
what is selection bias?
when the research participants are not representative of the study population
what is ‘observor bias’?
the prior knowledge of the person recording the data could influence what they record
what is social acceptability bias?
people try to present themselves to the researcher in a good like - underestimating bad behaviours and overestimating good ones
what is courtesy bias?
people tell you what they think you would like to hear
how do we reduce bias?
with research design and analysis
