L9- Probability and Significance; Type I & Type II errors

Question 1

Define level of statistical significance

Answer 1

Level at which decision made to reject null hypothesis in favour of experimental/alternate hypothesis- states how sure we can be that IV having an effect on DV and ✖️ due to chance

Question 2

Define chance

Answer 2

Something has no real cause- just happens

Question 3

How do we know if results are significant?

Answer 3

From results gained from experiment- compare control and experimental conditions- AND look for whether real difference exists between 2 sets of data
- If 2 sets of data similar then statistical test might no real difference and
… we accept null hypothesis
- BUT if there is a real difference between the 2 conditions (proved by conducting statistical tests) then we accept experimental and reject null hypothesis

Question 4

What is probability?

Answer 4

Probability- numerical measure- determines whether results due to chance or due to real difference existing between experimental and control conditions
If real difference exists (calculated statistically)- we say results significant, null hypothesis rejected and experimental hypothesis accepted

Question 5

What is the standard level of significance used in psychology and why?

Answer 5

p<0.05 (5% level)
Used in Psychology because:
- not too strict or too lenient but middle, fair value of significance
- minimises chances of making Type 1 or Type 2 error

Question 6

What does the 5% significance level (p<0.05) actually mean?

Answer 6

If significance level achieved then probability (p) of 5% or less that results due to chance/fluke … 95% or more certainty that results showing real difference between control and experimental conditions

Question 7

When are 5% significance levels usually used?

Answer 7

Usually used when there is a directional 1 tailed hypothesis clearly stated in the research

Question 8

What does the 10% level of significance signify?

Answer 8

Expressed as p<0.10 (10%)- used when we allow 10% or less margin of error and … 90% or more certainty that results really showing significant difference

Question 9

What does a 1% level of significance signify?

Answer 9

Expressed as p<0.01- 1% or less probability that results due to chance and … 99% or more certainty that real difference exists between control and experimental conditions
- Often used when research findings critical- e.g. when testing effect of drugs on humans

Question 10

What is a type 1 error?

Answer 10

Type I error (false positive)- occurs when researcher incorrectly rejects true null hypothesis (… incorrectly accepts alternate hypothesis)-
… you report that your findings are significant BUT in fact occurred by chance

Question 11

What is a type 2 error?

Answer 11

Type II (false negative)- occurs when researcher incorrectly accepts false null hypothesis … researcher concludes that results ✖️ significant BUT they actually are