Thank you for using the site. I'll be happy to help you. Please post the question information at your convenience.
Ok, what about the subway and metro rail do you want to consider? I'm not clear on what you need to produce eventually. Do you just need one of each type of hypothesis set?
I would suggest that we rephrase that slightly to something like "Do subway riders feel better about saving money than non-subway riders!" Rephrasing it in this manner avoids the word "cause". Since causation is so difficult to prove, it's a word that is best avoided.
Assuming that we had two samples (one of subway riders, and one of non-subway riders), and some measure of the average sense of how the members of each group felt, then your t-test hypotheses would be:
H0: μ1 = μ2
Ha: μ1 > μ2
where μ1 is the mean for the subway riders and μ2 is the mean for non-subway riders.
You can expand this to an ANOVA by adding a third group, say metro raill riders. The hypotheses would then be:
H0: μ1 = μ2 = μ3
Ha: At least one of the group means is different.
Please feel free to ask if you have any questions about these hypotheses.
There is a null hypothesis and an alternative hypothesis in each set. The null hypothesis is designated as H0, and the alternative hypothesis is designated Ha.