Back to FAQs

Joinpoint: Frequently Asked Questions

Estimating the Autocorrelation Parameter

Can the Joinpoint software estimate the autocorrelation parameter?

Answer: No, the software cannot estimate the autocorrelation parameter. In our paper in Statistics in Medicine (Kim, et al, 2000, 335-351) we outlined a method for estimating the autocorrelation parameter (see section 2.3). We found in our simulations on table IV of that paper that adjusting for autocorrelation was helpful in maintaining proper size of the tests of joinpoints when there was large autocorrelation. We also found that if there was no autocorrelation that the adjustment seriously affected the power of the test to detect joinpoints. For example we see in Table IV (b) with phi=0, the power goes from 90% to 68 %. Thus we chose not to include an option in the joinpoint software for estimating autocorrelation as done in the paper. This is to avoid this serious loss of power when there is no autocorrelation present. Instead we allow the user to input values for autocorrelation as a type of sensitivity analysis. The way this option would be used is as follows:

1. Fit the model with the uncorrelated errors option.
2. If the user suspects that there is positive autocorrelation in the data, then repeat the analysis trying several values of the autocorrelation parameter, say for example .1, .2, and .3. If the results are very similar with different values of the autocorrelation parameter, then the user knows their results will still hold if there is autocorrelation present. If the results change as the autocorrelation parameter changes then the user may end up presenting the series of results, to show how the results depend on different assumptions about the autocorrelation.
3. If the user suspects negative autocorrelation, there is need to do any further analysis (see Kim, et al., 2000).

Back to FAQs