Is One Way ANOVA the proper test that I should use?

#1
Hi!

I have precipitation data ~20 years. For each year, I have the total precipitation for a given month for that specific year. I am trying to find if there there is a statistically significant difference between the years in the total precipitation. I have organized my data such that each column represents a specific year (i.e. 2000) with each column representing aspecific month (i.e. Jan). I am trying to do ONE way ANOVA using SPSS, however, the program is not giving me any results. Additionally, in the output data, the program is saying that it is unable to do post-hoc tests as "Post hoc tests are not performed for variable1 because at least one group has fewer than two cases".

I have a couple of questions about this:

Firstly, am I using the appropriate test (ANOVA) or should I be looking at an alternative statistic test?

Secondly, why am I am unable to get post-hot test results? Is the problem with the way I have organized my data OR is there a better way to preform the post-hoc tests?

Thanks in advance!!:)
 

noetsi

Fortran must die
#2
With one way ANOVA you are saying some variable (here rainfall) varies at given levels of another variable. Or that at least was the way I was taught that (ANOVA terminology varies widely by author unfortunately). And you don't appear to have another variable. Repeated measure anova looks at whether there is a change over time, usually after some intervention. But again you don't seem to have an intervention you are testing.
 
#3
So I wouldn't be able to ANOVA with total rainfall (dependent variable) and year (independent variable)?

Sorry, I am just very lost with how to approach my data at this point.
 

noetsi

Fortran must die
#4
Sure you can. The question is if you should do cross sectional (in this case one way)ANOVA or repeated measure ANOVA (which looks at change over time). I think you probably want to do repeated measure ANOVA, but I would look at a discussion of that ANOVA approach and see if it fits your needs.