Chi Square Analysis chi
square test in spss interpretation interpretation of chi square test
results in spss with Example Summary
If sig value > 0.05
then accept H0
If sig value <= 0.05
then accept H1
Assumptions
· Sample
must be random and independently distributed
· Each
unit in the cell has the same probability being from a particular cell
· Each
observation may be categorized either into class 1 or class 2 etc
Solution
1. H0 : Sample is independent
2. H1 :
There is association
3. Sig value: 0.05
For
simple category Analysis
Check expected count if
any value is less than 5 then use
“Likehood/contingency table” than pearson chi square
method
· If
2*2 then use risk and see phi
· If
2*3, 2*4 then use contingency table
· If
3*3, 4*4, 5*5 then use cramer-v
Kappa
Applied
When use:
· when rows
and columns have same name.
· use
for qualitative variable want to see index of agreement b/t observation
Solution:
· Click
on statistics “kappa”
· In
kappa column value tell (index of agreement)
· If
want to see sig value then see “Asymp std error” value.
Nominal
Data
When use:
Category is not properly
defined e.g
· Over,
proper, under
· High
dose, recommended dose, low dose
If data is not in order
then see “Linear by Linear” in chi square table.
Kandall-tau
b test
When use
To check one value is
decreasing & another value is increasing. For qualitative
Solution:
· Click
on “co-relation”
· Click
on “kandrall-tau-b”
· Check
kandrall value in table if value is (-ve) mean
one value is decreasing & another value is increasing
Cochran-Mantel-Haenszel
When use:
· When
there is group like “city” etc
· Want
to group/pool element
· When
there is more than rows and column in table
Solution
· Enter
city column in pool
· Click
on Mental-haenszel in statistics
0 comments:
Post a Comment