14 - Chi Square

Question 1

Chi square

Answer 1

• Non parametric
• nominal data
• observations can be classified into discrete categories and treated as frequencies
• categories should be mutually exclusive
• the more categories the larger the sample
• 2-tailed (sensitive to difference but not to the direction of difference)

1. Goodness-of-fit test
2. Cross tabulation between two categories

Question 2

Chi square formula

Answer 2

Somme (observed - expected)^2/expected

Question 3

Goodness of fit test

Answer 3

• Chi square
• H0 = normal distribution of variables and uniform (equally split)
• degrees of freedom = nb categories - 1
• effect size = chi square / (total sample size)(degrees of freedom)

Question 4

Cross-tabulation test

Answer 4

• Contingency tables : 2x2 (2 variables divided into 2 categories)
• expected frequencies= (row tot * column tot)/tottot
• df = (row-1)*(colum -1)

Question 5

Chi square = 0

Answer 5

No difference between O and E frequencies