Summary Of All Stats Tests Etc So Far Mocks

Question 1

PMCC

Answer 1

Is used to find a number to see how closely correlated thins are

Question 2

PMCC test
Null
Consideration

Answer 2

To check if the data found in a sample suggests if there is a correlation in the parent population
Null is that there is no hypothesis

2)only works if data normal distribution eliptical

Question 3

Spearman’s rank

Answer 3

Shows if there’s association between rankings

Question 4

Spearman’s rank tedt
Null hypothesis

Answer 4

To see if there is association between two ranks , SOME
Null = there is NO association

Question 5

Regression lines

Answer 5

Use model to find regression to predict values

Sum of residuals is 0 and these are distances away

Question 6

Chi 2 contingency table test

Null hypothesis

Considerstions

Answer 6

Is a tedt to determine association between two variables not necessarily rwnled

Null hypothesis = there is no association
Alternative = there is some

Consideration = assumes independent, work out independent first
- assumes normal
- also small frewuency consideration

Question 7

Chinsqaured to see if uniform binomial or poison?

Answer 7

Possion
- use poison to generates expected frequencies. Might predict lambda
- null hypothesis = it does fit

Binomial
- use p to work out edpected , or work it out yourself
- null= there is

Question 8

ALWAYS REMEMBER SMALL FREWUENCY THING FOR

Answer 8

ANY CHINSQAURED TEST

Question 9

If distribution nit clear for goodness, use chi sqaured how?

Answer 9

Null hypothesis = assume true

Use cintext to find expected

Question 10

What’s the null Hypotheken overall

Answer 10

Goodness if fit = assume it does fit

For correlation / association, assume no associstion

Question 11

What is n in binomial case and what about poission

Answer 11

Binomial case = what’s constant for each thing, so 30 samples for each hospital, n= 40

So the maximum value your model can take

Poission this goes to infinity remember

Question 12

If youreconfused about total frewuency what to do?

Answer 12

Just add the frequencies up and stick to this value

And keep n the number collumn , take care if this says 8+ might have to extend it

Question 13

What does a very low chi 2 statistic. Mean

Answer 13

It means you could’ve habe rigged data, because the lower it is it means to closer to actual values your expected values are

So you could ,are up valued to fit model, or omitted, or chested etc, these are considerations when cin2 is ver small

Question 14

Finally remember what to do with small frewuency

Answer 14

Less than 5 = ADD THEM UP