Lecture 11

Question 1

When does selection bias occur? (3 things)

Answer 1

1. When individuals in a study differ systematically from the population of interest leading to a systematic error in an association or outcome.
2. When participants differ in ways other than the interventions or exposures under investigation. If so, the results are biased by confounding.
3. When the observed effect may be exaggerated or when it is not possible to assume an effect in those not included in the study.

Question 2

What does a sampling bias involve?

Answer 2

It involves sytematic error due to a non-random sample of a population.

Question 3

What is poor cross-validation?

Answer 3

1. Naïve split of data set into 70% training data and 30% test data without randomizing the order of the data first.

Question 4

What is data fishing/p-hacking?

Answer 4

Presenting the most significant patterns/results of a data dredge as if they were produced by a single, well-designed experiment.

Question 5

What is Publication bias?

Answer 5

A bias that occurs once only studies with significant results are published. Especially occurs when a follow-up study combines the results of a dozen of studies.

Question 6

What is Attrition bias?

Answer 6

Bias caused by attrition (loss of participants), discounting trial subjects/tests that did not fully run.

Question 7

What does Lost to follow-up mean?

Answer 7

This is a form of attrition bias. Often happens in medicinal studies over a length period of time, when participants stop with the experiment, don’t respond anymore, get ill, etc. Also called Non-response or retention bias.

Question 8

What is information bias?

Answer 8

Any systematic difference from the truth that arrives in the collection, recall, recording and handling of information in a study, including how missing data is being dealt with.

Question 9

When is information Bias more likely to happen?

Answer 9

In observational studies (particulary retrospective ones), but is not impossible in experimental studies.

Question 10

What do Non differential misclassification of measures lead to?

Answer 10

To an underestimation of effect.

Question 11

What is a Non differential misclassification of measures?

Answer 11

When errors in measurements occur equally in all comparison groups.

Question 12

What is a Differential information bias?

Answer 12

There are different levels of inaccuracy between comparison groups. It could work in either directions, resulting in an overestimate or underestimate of the true effect.

Question 13

What is an Observer bias?

Answer 13

Different observers may use or evaluate certain criteria or procedures differently. It can also occur when the subject knows they are being examined (Hawthorne effect). The subject may act differently.

Question 14

What is a Recall and reporting bias?

Answer 14

Refers to selective revealing or suppression of information. For example, participants may not be able to remember their medical history of behavior in the past, and do not adequately report it.

Question 15

What are strategies to avoid information bias?

Answer 15

1. Choosing an appropriate study design.
2. Following well designed protocols for data collection and handling and the appropriate definition of exposures and outcomes.