**********************************************.
* Final production run of program to          .
*  merge AG and case datasets in SPSS format  .
*  recode or create  variables                .
*  save analysis dataset and                  .
*  run logit regressions reported in SPPQ     .
* Leonard Ray October 2004                    .
**********************************************.

*************************************************************.
***            DATA MERGE STEP                             **.
*************************************************************.


MATCH FILES FILE='B:\AGCASES.sav'
           /TABLE='B:\fullags.sav'
           /BY state year.

*************************************************************.

***             VARIABLE CREATION STEP(S)                  **.
*************************************************************.

** AG/Panel Interaction term requires case and AG data       .

RECODE
  party
  (1=-1)  (0=1)  INTO  agptyrev .
VARIABLE LABELS agptyrev 'ag party with dems as 1'.
execute.
RECODE
  demmaj
  (1=1)  (0=-1)  INTO  dempmajr .
VARIABLE LABELS agptyrev 'ag party with dems as 1 and reps as -1'.
VARIABLE LABELS dempmajr 'panel party with reps as -1'.
compute agpanel2=agptyrev*dempmajr.
VARIABLE LABELS agpanel2 'ag and panel of judges are same party' .
execute.

* removing the DC circuit.
SELECT IF (circuit NE 0).

*compress scales to make parameters visible.
compute staff_i=staff_i/100.
compute salary_i=salary_i/1000.
compute attny_i=attny_i/100.
compute entry_i=entry_i/1000.
compute entsal_x=entsal_x/1000.
compute statepop=statepop/1000000.

SAVE
 OUTFILE='B:\analysis.sav'.

GET
  FILE='B:\analysis.sav'.


*************************************************************.
**              Running Logits for SPPQ paper                .
*************************************************************.
* SPPQ model Table 1 Left hand collumn.

* Prediciting AG victory.
* using w-nominate scores for ideology.
* using wasby suggested staff (attny only) and salary (starting) data.
* using state / panel agreement.
* using budget adjusted for national CPI and state population.
* using salary adjusted for state CPI.
* and state population
* linear service variable.
* all cases.
* all variables suggested by reviewers included.

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER 
tottrain appoint service attny_i entsal_x bgtpcap statepop agree_p agree_c proussc
circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
 strgopp criminal economic crimpet petition pveconap complex 
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .





* SPPQ model Table 1 Right hand column.

*policy cases only.
*using w-nominate scores for ideology.
*using wasby suggested staff (attny only) and salary (starting) data.
*using state / panel agreement.
* using budget per capita.
* using salary adjusted for state CPI.
* and state population.
* linear service variable.
* all reviewer suggestions included.

temporary.
select if (policy eq 1).
LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER 
tottrain appoint service attny_i entsal_x bgtpcap statepop agree_p agree_c proussc
circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
 strgopp criminal economic crimpet petition pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .


* code to test for circuit differences reported in SPPQ table 2.
LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ7 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ11 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

LOGISTIC REGRESSION VAR=agwin
  /METHOD=ENTER tottrain service attny_i entsal_x bgtpcap statepop agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 
proussc strgopp criminal economic crimpet petition complex appoint pveconap
  /CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .


*writing out small version of dataset for analysis with CLARIFY.
* SAVE OUTFILE='B:\clarify.sav' 
 /KEEP= appoint pveconap agwin service tottrain term1 entsal_x bgtpcap statepop  
 attny_i entry_i biggest agree_p agree_c
circ1 circ2 circ3 circ4 circ5 circ6 circ7 circ8 circ9 circ10 circ11 
proussc strgopp criminal economic crimpet petition complex policy.

EXECUTE .

********************** END OF SYNTAX **************************.