*
* Example 10.6 from page 343
*
* This is a specialized calendar setting for data which has several years (in
* this case 4) between observations.
*
calendar(years=4) 1916
open data fair.raw
data(org=columns) 1916:1 1992:1 year V I DPER DUR g3 p15 n $
  g2 gYR p8 p2YR Ig2 Ip8 demwins In $
  d Id Ig3 Ip151md In1md demvote incum gnews $
  inf partyWH pWHgnews pWHinf
*
linreg demvote
# constant partywh incum pWHgnews pWHinf
*
* The SUMMARIZE instruction with the VECTOR option evaluates a specific linear
* combination of the regression coefficients. This gives the sum and standard
* error.
*
summarize(vector=||1.0,1,1,3,3.019||)
*
* The actual percentage was 54.65, so this shows the number of standard errors
* above the prediction that was actually observed. This uses the variables %SUMLC
* (sum of linear combination) and %VARLC (variance of linear combination) defined
* by SUMMARIZE.
*
disp "Z Score for Actual Data" (.5465-%sumlc)/sqrt(%varlc)