Probability and significance

Probability definition

Measure of likelihood an event will happen

0= impossible 1= certain

What do 0 an 1 mean in probability?

0 = impossible

1 = certain

Significance definition

How sure u are that a statistical difference or correlation exits

(how sure u are that the study actually discovered something)

How do you decide if a study has discovered something or not?

Using significance level

What is a critical value?

The numerical cut off point b/t acceptance or rejection of null hypothesis

What is

The numerical cut off point b/t acceptance or rejection of null hypothesis

Critical value

What is the observed/calculated value?

Value calculated for a set of data. Used in stat tests

When is the null hypothesis accepted?

Only if study is found to be insignificant

What does H1 stand for?

Directional hypothesis

What does H2 stand for?

Non-directional hypothesis

What is the usual level of significance?

5%

What does the accepted level of significance mean?

% probability results occurred by chance

Why is there an accepted level of significance?

Can never claim 100% certainty in results

When is a more stringent level of significance used? (1%)

Drug therapies or non-repeatable studies

When is a more lenient level of significance used? (10%)

Results have power but shouldn’t be seen as significant

When do Results have power but shouldn’t be seen as significant ?

When a more lenient level of significance is used (10%)

Why is 5% usually used as the level of significance?

Halfway b/t 1 + 10 so reduces risk of making Type I or Type II error

What is a Type I error?

False positive

Alternative hypothesis is accepted when it should’ve been rejected

What is a False positive

Alternative hypothesis is accepted when it should’ve been rejected

Type I error

What is a Type II error?

False negative

Null hypothesis is accepted when it should’ve been rejected

What is a False negative

Null hypothesis is accepted when it should’ve been rejected ?

Type II error

When are you more likely to make a Type I error?

When level of significance is too lenient

When are you more likely to make a Type II error?

When significance level is too stringent