6 - chi squared tests

Question 1

hypothesis test of independence

Answer 1

H0 are independent - no association
H1 arent - is association

Question 2

expended values

Answer 2

find the expected proportion by doing the total of row/ total * total of column

Question 3

chi squared

Answer 3

tests goodness of fit
sum ( (O-E)^2 / E)

Question 4

v in an n by m table

Answer 4

v = (n-1)(m-1)

Question 5

all E need to be

Answer 5

> = 5
if less then 5 combine rows/ columns with next lowest

Question 6

critical value chi squared

Answer 6

if 5% look at the 0.95 - look at the upper tail always - to find the critical acleu

Question 7

yates correction

Answer 7

used when v = 1 (2*2) as chi squared isn’t a good approximation
= sum( (O-E - 0.5)^2 /E)

Question 8

fitting a distribution hypothesis

Answer 8

H0 follows distribution
H1 doesn’t follow distribution

Question 9

v in goodness of fit tests

Answer 9

number of data - number of constraints
(normally -1 unless you estimate a parameter then it’s -2)

Question 10

how to estimate p binomial

Answer 10

E(X) = sum of (fx)/ sum of (f) = np
p = E(X) /n