Bias

Question 1

3 components of bias

Answer 1

source/type

magnitude/strength

direction (bias can under or over estimate the true measure of association)

Question 2

Informational bias

Answer 2

any aspect of the way the researcher collects information which creases a systemic difference between groups and in the quality/accuracy of their information

errors in measurement of classification (i..e misclassification)

Question 3

What would misclassification be considered?

Answer 3

informational bias

Question 4

Selection bias

Answer 4

any aspect in the way researchers select subjects which create a difference

failure to pick subjects that do not reflect the population or when subjects dont come from the same group

Question 5

Healthy worker bias

Answer 5

selection bias

Question 6

Self-selection/Responder bias

Answer 6

selection bias

“only people who are educated or uneducated filled in survey…” etc

Question 7

Recall bias

Answer 7

Information/Observation/Measurement bias

differential level of accuracy in detail in recalling events.

people who were traumatized by an event have a better time recalling it

memory gets in the way of things

Question 8

Contamination bias

Answer 8

Information/Observation/Measurement bias

members of the control group accidentally receive treatment or are exposed to the intervention being studied

Question 9

Compliance bias

Answer 9

Information/Observation/Measurement bias

groups being interventionally studied have different compliances

Question 10

Lost to follow up bias

Answer 10

Information/Observation/Measurement bias

people who pull out of the study early will change the impact of the findings

Question 11

Observer related variation

“Proficiency bias”

Answer 11

Information/Observation/Measurement bias

systematic difference in soliciting, recording, interpreting on the part of the researcher

interviewers knowledge may influence the structure by alterting tone of questions, follow up questions

or the skill of interviewers changes during the study because of personell change or lack of consistency

Question 12

Observer related bias

Diagnosis/Surveillance bias

Answer 12

Information/Observation/Measurement bias

Different evaluations/classification/diagnosis/observation between study groups

Question 13

Hawthorn-like effect from the researchers perspective

Answer 13

Diagnosis/Surveillance bias
Information/Observation/Measurement bias

different evaluations classifications between study groups

observers have preconceived expectations of what they should find in examination

Question 14

Misclassification bias

Answer 14

error in classification either disease or exposure

Question 15

“non differential” means there’s

Answer 15

error in both groups equally

misclassification of one (exposure or disease) unrelated to other (exposure or disease)

misclassification of this type can move measure of association towards null hypothesis

Question 16

A study with a RR of .3 ends up with an RR of .7. What kind of bias might this be

Answer 16

misclassification - nondifferential

moves toward null hypothesis

Question 17

Differential error

Answer 17

misclassification error

bias can more in either direction from null hypothesis

Question 18

A false positive would do what do what to the data in relation to the null hypothesis

Answer 18

move away from it

type I error

better chance of rejecting the null hypothesis when in fact it was true

Question 19

a false negative would do what to the data in relation to the null hypothesis

Answer 19

move toward it

type II error

Question 20

if RR or OR is .1, and misclassification causes them to adjust to .5 what effect does this have?

Answer 20

you attenuate towards null hypothesis

Question 21

if RR or OR is 1.8 and misclassification causes it to adjust to 5.6, what effect does this have?

Answer 21

you inflate away from the null hypothesis

Question 22

What is the fundamental difference between differential and non-differential error?

Answer 22

non differential error always moves data closer to the null hypothesis, creating a type II error

differential can move toward or away from the null hyp