Quantitative 2

Question 1

Correlation matrix

Answer 1

Select variables with significant assocation outcome to use in analysis

Question 2

Stepwise backward selection

Answer 2

Stepwise removing variable that is the least significant

Question 3

Selection bias

Answer 3

Selection technique doesn’t assume a sample that is representable for the population

Question 4

Stratification

Answer 4

Sorting data into groups or layers
Analysing each stratum seperatly

Question 5

Mediation analysis

Answer 5

Does X have an effect on Y through M?
Adjusting for mediator leads to partial effect

Question 6

Not adjusting for mediator leads to

Answer 6

full effect

Question 7

NHST

Answer 7

Null-hypothesis significance testing

Question 8

Significant findings are more likely to be true if

Answer 8

Lower significant lever (fewer significant results)
Higher power (change of finding a significant result in sample if effect is real)
Higher prior probability

Question 9

Higher power through

Answer 9

Larger sample size

Question 10

Problems with significance

Answer 10

Publication bias
People don’t look at size of effect
Likely to be overestimates

Question 11

95% CI

Answer 11

If the experiment was repeated multiple times, 95% of the CI’s would contain the true value

Question 12

Reducing uncertainty by

Answer 12

Having larger samples > smaller standard error > narrower CI