Quick one: is ANOVA the right test?

#1
Hi everyone,

I was wondering if some generous soul would invest a minute and a half to help me with this. I'm new to statistics and have a simple demonstration I would like to undertake.

50 states of the US = 50 observations
var1 = percentage of energy consumption each state derives from coal
var2 = 0 to 3 (4 categories) according to how many of a specified range of policies each state has undertaken to curb co2 emissions from power plants.

So my wife is a psych student and she says that what the ANOVA would tell me is if the means are significantly different between categories of states (i.e. between those who do no policies v. 1 policy v. etc.)

From what I read, the causality relation in projects using ANOVAs usually run the other way (e.g. difference between groups affect a continous variable). For me obviously, the interval is the independent variable. So is the test appropriate for me as well?

I don't need an irrefutable test, simply something that illustrates a relationship I already intend to analyze qualitatively. So far, the results of the test in Stata seem promising, significant and all. I just want to make sure I'm not way off track.

Thanks for your time.

Vincent

edit: And here's the stata output http://vincent.arel.googlepages.com/home
 
Last edited:

Xenu

New Member
#2
The direction of the relationship doesn't really matter if you just want to test whether there is a relation between the variables. If there is a relationship in one direction, then surely there is a (numerical) relationship in the other direction as well.

If I were you I would try to make a make a plot or table of some sort to see if it is possible to show the relationship graphically.
 
#3
Thank you very much for your answer, confirms my guess.

yes, you are right, i'll probably include a table or an histogram like this.



Thanks again, have a good week!

Vincent