Hi All,
I just posted to SSC a completely revamped version of the-itsa- package (ITSA stands for "interrupted time series analysis").
What makes this revised version so special is that it now utilizes GLM as the analytic engine. That allows users to specify models for most any type of data, such as counts and fractional data, etc. Newey-West standard errors are computed for these models, so users of previous versions of -itsa- will not see any difference in the results when estimating a -newey- model. Now, however, a user is not limited to a continuous outcome, as before. The help file also gives some examples with Poisson and fractional data types.
-itsa- also has a new option to superimpose lowess lines on the graph as a helpful visual.
Finally, Kit Baum has just posted a revised version of his -actest- command. I have always advocated the use of -actest- after running -itsa- for a single-group analysis. Kit's revised command will now work after -itsa- when other model types are specified (e.g. poisson, binomial , etc.).
In short, a lot of good stuff!
To download these programs, type:
Cheers,
Ariel
I just posted to SSC a completely revamped version of the-itsa- package (ITSA stands for "interrupted time series analysis").
What makes this revised version so special is that it now utilizes GLM as the analytic engine. That allows users to specify models for most any type of data, such as counts and fractional data, etc. Newey-West standard errors are computed for these models, so users of previous versions of -itsa- will not see any difference in the results when estimating a -newey- model. Now, however, a user is not limited to a continuous outcome, as before. The help file also gives some examples with Poisson and fractional data types.
-itsa- also has a new option to superimpose lowess lines on the graph as a helpful visual.
Finally, Kit Baum has just posted a revised version of his -actest- command. I have always advocated the use of -actest- after running -itsa- for a single-group analysis. Kit's revised command will now work after -itsa- when other model types are specified (e.g. poisson, binomial , etc.).
In short, a lot of good stuff!
To download these programs, type:
Code:
ssc install itsa, replace
Code:
ssc install actest, replace
Ariel