1.6 Practical Significance + Statistical Power

Question 1

if the null is true, then…

Answer 1

the hypothesised distribution is the same as the actual distribution and nothing is happening

Question 2

type 1 error (alpha)

Answer 2

we reject the null when it’s actually true (we’re too liberal and think something is happening but it isn’t)

Question 3

type 2 error (beta)

Answer 3

we retain the null when it’s actually false (we’re too conservative and don’t recognise that something is happening)

Question 4

power

Answer 4

• sensitivity
• PROBABILITY OF AVOIDING A TYPE 2 ERROR (likelihood that a significance test will detect an effect when one truly exists)

Question 5

things that affect power

Answer 5

• sample size (larger samples increase power)
• effect size (stronger effects easier to detect)
• significance level (usually set at 0.05)
• variability of effect (more variability = lower power)

Question 6

statistical significance

Answer 6

• proclaimed when p-value < 0.05

Question 7

practical significance

Answer 7

• determined by size of effect and its application
• does it matter in the REAL WORLD??

Question 8

appeal to ignorance

Answer 8

argument from ignorance
* thrusting an assertion into a region of uncertainty
* assertion that a claim is true simply because it has not been proven false - rely on absence of evidence as if it were compelling evidence itself
- you can’t prove I’m wrong, so I must be right

Question 9

false dichotomy

Answer 9

logical fallacy when someone falsely asserts there are only two possible options or outcomes, ignoring other potential choices
* situation: want to be able to attack other side w/o establishing own case
- modern: denialism