Help with statistical significance testing and null/alternative hypothesis! thx
In a clinical trail, 60 patients involved in a study, all of them are suffering from a brain injury.
30 patients got "NewDrug" and 30 got "Placebo".
I need to assess the efficacy of the drug treatment by comparison with the placebo treated by statistical significance testing for the entire sample.
I have a table for the placebo group, where first column has the data of the values of the volume of injury before receivingplacebo, second column has values of the volume injury after receiving placebo, and the third column with the differences (column 1 - column 2)
The same table for the newdrug group, where first column has the data of the values of the volume of injury before receivingdrug, second column has values of the volume injury after receiving drug, and the third column with the differences (column 1 - column 2).
I dont know which test to use ? is it paried or un-paried t-test (i tried to do paried t-test but i dont know what to write in the null hypothesis and alternative hypothesis), so if someone can help me with that it will be great!
Re: Help with statistical significance testing and null/alternative hypothesis! thx
If you want to test whether there is a statistically significant difference between two means, then you will have a hypothesis H0: mu1 = mu2 and H1: mu1 != mu2.
For paired t-test I may recommend it if for example, the same person was involved in each pair of observations or if the other attributes were largely the same or similar enough which is not the case.
It seems given the data, it would make sense to use a standard 2 sample t-test with the above hypotheses and then if they are unequal, see which mean is the biggest to answer your question.
You should have enough data to satisfy t-test assumptions if all observations are independent and the process to collect the data ensures this (you will have to decide if this is the case since I don't know enough details).