Bias and Confounding

Question 1

Difference between Random Error (RE) and Systematic Error (SE)

Answer 1

Random Error = Chance

Systematic Error = moves value away from true value

Question 2

Random Error (RE) (3)

Answer 2

1) any variability in data that cant be explained
2) influenced by sample size
3) Influences the presence of our measures

Question 3

Systematic Error (SE) (3)

Answer 3

1) moves measure away from true value
2) influences accuracy of the measure
3) more likely to conclude an incorrect inference about what we’ve observed

Question 4

How does RE effect epidemiological research?

Answer 4

The effect of random error (Chance) may result in either an Underestimation or Overestimation of the true value

Question 5

How does SE effect epidemiological research?

Answer 5

A type of sampling error where you can make wrong conclusion about observations:

1) Type I error - Rejecting the null hypothesis when it is true
2) Type II error - Accepting the null hypothesis when it is false

Question 6

Difference between Internal and External Validity

Answer 6

Internal validity looks at the approach used is a population of a study whereas External validity (directly proportional to internal validity) determines whether it is valid in a general population unrelated to the study.

Question 7

Examples of Confounding

Answer 7

1) Heard size, wearing an apron, and leptospirosis
2) Age, Living at the Gold Coast, and High Mortality rate
3) Smoking, Carrying matches, and Lung Cancer

Question 8

External Validity

Answer 8

Directly linked with Internal Validity
Describes how appropriate it is to apply results to a population other than study population
- if the internal validity of a study is poor, the external validity will also be poor

Question 9

Confounding Factor

Answer 9

“Causal” relationship to Outcome and a “non-causal relationship” to Exposure

Example) relationship between smoking and laryngeal cancer
Confounding factor: Alcohol
People who smoke may also drink

Question 10

Controlling confounding (3)

Answer 10

1) Restriction - Study large numbers
2) Matching or stratification - match by herd size
3) Analytical control - multivariate approach

Question 11

Bias (definition)

Answer 11

Results in “observed effect estimates” which “differ” from those which truly exist in the target population

Question 12

Features of Misclassification Bias

Answer 12

refers to the measurements of the outcome or exposure AFTER units were selected

• differential
• non-differential

Question 13

Examples of selection bias

Answer 13

Objective: Estimate females in population
Method: Sample from Rugby game
Problem: Disproportionate population at these games

Question 14

Examples of misclassification bias
Definition
3 subgroups

Answer 14

Errors in the information that is recorded once participants have been selected for inclusion

1) Recall bias
2) Interviewer Bias
3) Obsequiousness bias

Question 15

Features of Selection Bias

Answer 15

procedures used to select units that are included in a study

Question 16

Non-response bias

Answer 16

Selection bias

owners non-response or refusal to participate in study

Question 17

Lost Follow-up bias

Answer 17

Selection bias

Owners withdraw animals from study

Question 18

Exclusion bias

Answer 18

Selection bias

Animals develop health problems unrelated to the study and have to be excluded

Question 19

Survival Bias*

Answer 19

Selection bias
Ex) introduction of insulin has increased the lifespan of diabetic patients, producing an apparent increase in the prevalence of the disease

Question 20

Survival Bias*

Answer 20

Selection bias
Ex) introduction of insulin has increased the lifespan of diabetic patients, producing an apparent increase in the prevalence of the disease

Question 21

Example of Misclassification Bias

Answer 21

Objective: Ask people if they know what newcastle disease is
"Yes" answers may include
- had experience with disease
- read about the disease
- don't want to sound ignorant

Question 22

Change variation

Answer 22

Internal validity
Reflects the variability in the data that can’t be explained
Increase in sample size will result in reduced chance variation

Question 23

Change variation

Answer 23

Internal validity
Reflects the variability in the data that can’t be explained
Increase in sample size will result in reduced chance variation

Question 24

Sampling Error (2)

Answer 24

A Type I error - Rejecting the null hypothesis when it is true
A Type II error - Accepting the null hypothesis when it is false