[MT] CHI SQUARE TEST

Test used to compare 2 or more proportions

(%) or to describe the

relationship/association between 2 variables

CHI-SQUARE TEST

Test used to compare 2 or more proportions

(%) or to describe the

relationship/association between 2 variables

CHI-SQUARE TEST

Test compares the observed frequencies (O) in

each cell with the expected frequencies (E)

under the null hypothesis

CHI-SQUARE TEST

no association/relationship or no significant difference exists between row and column variables IF the ________ frequency will be very close to the ______ frequencies (very small difference), and the resulting value of chi square test statistic is expected to be small

observed

expected

• comparing 2 or more proportions (%)

Chi-square Test of homogeneity 

determining the association between 2 qualitative variables

Chi-square Test of association/independence

df of chi square test

Df = (#rows – 1) x (#columns -1)

Chi square has no ____

negative values

Higher degrees of freedom, the curve moves to the right, higher SD =

wider curve

Used to determine whether or not two or more populations have the same proportions for the different categories of another variable

X2 Test of homogeneity

For X2 of homgeneity If used for 2 sample groups and the variable is ______, it can be used interchangeably with z-test (same hypotheses and conclusions)

dichotomous

For x2 of homegeneity Samples should be ________ and may be classified into 1 of 2 _____ categories (with or without characteristic)

independent

mutually exclusive

what X2 Test of homogeneity hypothesis is this?

The proportion of elements falling in each category of the variable of interest is the same for all the group

Null Hypothesis

what X2 Test of homogeneity hypotheses is this?

At least 2 of the population proportions are not equal. (Not all proportions are equal)

Alternative Hypotheses

Used to test for the existence of a relationship or association between 2 qualitative variables

X2 Test of association

X2 Test of association uses what samples?

random samples

Observations may be classified according to 2 variables

X2 Test of association

Data has 2x2 tables (2 variables with 2 groups/categories)

X2 Test of association

what X2 Test of association hypotheses is this ?

There is no association between the first variable and second variable (no relationship)

Null Hypotheses

what X2 Test of association hypotheses is this ?

There is an association between the first variable and second variable (no relationship)

Alternative Hypotheses

determining if there is significant difference of mean or proportion between comparison groups

Comparison

determining if there is a significant association/relationship between 2 variables

Association

in 2x2 (2 rows and 2 columns) tables:

expected frequencies should be ____________ 5 in all cells

greater than or equal to

if expected frequency is not satisfied for 2×2 tables what probability test should be used

Fisher’s Exact Probability Test

For larger tables All expected frequencies should be ________1

greater than or equal to

Not more than___ of the cells have expected frequencies less than 5, therefore at least ____ should have expected frequencies of ≥5

20%

80%