Bhavya Chebrolu wrote: ↑19 Apr 2020, 10:17
What is a Cochran - Armitage trend test and how is it different from the conventional chi square test?
Chi Square is useful to tests the association whereas Cochran-Armitage test of trend is used to detect the linearity [trends ] between variables
Lets say, X and Y are 2 variables
Chi square is used to test the association between X and Y. Here, both X and Y must be categorical variable
For Cochran-Armitage test is used to detect the linearity [trends ] between the variables.Here, X must be independent ordinal variable and Y must be binary response variable
Lets discuss with a funny example, Suppose i conduct a study where i ask the students that whether they play PubG ?
Now i shall have two variables
X = Gender (here the values will be Male or Female)
Y= PubG Playing (here the values will be Yes or No)
I can conclude my study with the findings whether gender is associated with PubG playing?
In this case, I can perform a simple Chi Square and report whether gender and pubG playing are associated
In my another study I collect the PubG players levels and the result of same year University Exam?
Here X= PubG Players level [so the values will be: Bronze, Silver, Gold or Platinum] and
Y = Pass in University Exam [Yes or No]
I can use Cochran-Armitage Test of trend to find the linear trend between different levels of PubG Players and University Results
Domain Examples
Example 1 Chi Square test
X= Gender [ values shall be Male or Female]
Y = Recovery from Corona [ Values will be Yes or No]
Chi square is used to find the association between gender and recovery from Corona
Example 2 Cochran-Armitage Test of trend
X= dose of a particular antiviral [ values are low dose, moderate dose or high dose]
Y = recovery from Corona [ Values will be Yes or No]
Cochran-Armitage Test of trend is used to find the trend between dose of the antiviral used against recovery of corona
We can't perform Cochran Armitage test using SPSS. It can be performed using SAS