Part A 

---

First click on the series "Marry", and then use "Unit root test" to examine
the data stationarity, and from below we can see that the series is stationary.
   
Then Plot the Series "Marry" by using "quick", "graph" option.
And then see the results like that below, an intervention seem to
occurred from observation 21 and go on.
   
And then now plot the correlogram of series "Marry", you can see that several
ACFs and PACFs are possible outcomes of this ARIMA model.  
To test the models, the best way is to examine their AICs and SICs:
   
From the above, ARIMA(2,0,2) is the best model.
And then add the dummy variable "inter" into the regression model to indicate
the intervention effect, then here are the results:
   

---