I have two treatments A and B which have been applied to tree buds to see if they will burst in a set time period. The outcome is binary - burst or not burst (yes or no, 1 or 0). There are an uneven number of buds in each group. In group A there were 100 buds and 38 burst. In group B there were 79 buds and 0 burst. I have been advised to do a 2 sample T test using minitab but I get an error message as all values in group B are 0 - the test simply wont run. What test can I do that will measure the validity of these stats? The null hypothesis is that there will be no difference in the proportion of bud burst between the two treatments.
2
-
2$\begingroup$ you should be doing a proportion test. $\endgroup$– AdamOCommented Apr 10, 2018 at 13:43
-
$\begingroup$ The normal distribution cannot apply in this case so the t test is not appropriate. $\endgroup$– Michael R. ChernickCommented Apr 10, 2018 at 14:22
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Per AdamO's comment I have now done a proportion test. For minitab (18.0) users this is under stat > basic statistics > 2 proportions, it includes a Fisher's exact test. For my case the null hypothesis was rejected p<0.00.