Need: In statistical inference we have different test statistics(test) which will differ by different differs of variables as well as for different types of analysis ,hence choosing correct statistical test is really important, wrongly chosen statistical tests might not give correct analysis as well as inferences.
Approach:
The following approach illustrates how to choose the correct statistical test for hypothesis testing.
There are 5 factors we need examine before choosing the statistical test which are
- Type of the variable
- Type of analysis
- Number of groups and data sets
- Study design comparison
- Distribution of the data
And the characteristics for these 5 groups mentioned below
While choosing the statistical test we need to examine all these 5 factors and then on the basis of their characteristics we will choose the statistical test using the below illustration.
Example: if you want to test is there any significance difference BMI score vs Gender this us how we will choose
- Type of variable – Continuous.
- # of groups – Two
- Type of analysis – Comparison
- Study design – un Paired matched
- Distribution of the data – Normal
We will use this information and follow through the above illustration and identified test as Un-paired t test and here we wanted to test the mean , we can select the t test for means
