Correlation Across Time Help?

[bataille]

Hi I'm a bit lost as to how to calculate the correlation between two groups of numbers across time.

I'd like to see if one group, taken from the start of the time frame, has a significant difference or similarity to another group, which represents the remainder of the time. This may be a total nonsensical thing to do however when I graph the averages of my data, the data is higher at the start than at the end and I would like to know if this is significant at all.

I've posted my data here if anyone wants to take a look:

http://pastebin.com/CNjuskFP

and I've been advised to use kendall's tau, however I'm unsure if this is correct. I'm also unsure about what information to post here so any help on that would also be massively appreciated.

 

[noetsi]

If you want to see if there is a correlation you can run any of a number of autoregressive tests (Durbin Watson, Durbin's special test statistic which test for general autocorrlation etc). If you want to know if you can use earlier values to predict later values you can use an autoregressive model (regression where you use earlier lags of Y to predict latter values).

Or you can just take a mean of the earlier values and compare them to the mean of the later values. Because these sets of values are not independent I don't think you can do a statistical test (I don't think you can do a pooled test either such as repeated measure ANOVA because you really don't have multiple values to test a change in. You have a single value that changes over time I think).