* File name: montecarlo.sps

* Copyright (c) 2008 Applied Personnel Research
* A limited license to use and modify this SPSS syntax is granted for non-commercial use only, and only with the following acknowledgement:
* "Thanks to Applied Personnel Research and Joel Wiesen for providing sample SPSS commands for Monte Carlo research."

* This is a (lightly edited) sample of the SPSS programs used for the 2008 APA Poster Session titled "Six Alternative Methods of Use of Employee Selection Test Data"
* by Joel P. Wiesen, Ph.D. and Aaron A. Brown

* This SPSS program was authored by Joel Wiesen (creation of variables and calculation of grade scores) and Aaron Brown (plotting results).
* Many thanks to Raynald (Ray) Levesque for providing the original Monte Carlo macro syntax, and to Marcin (of Poland) for showing my how to use that syntax.

*****************************.
* Outline

* 1. Read 8 random variables for specified number of cases.

* 2a. Create gender and EEO as categorical variables.
* 2b. Create a known proportion of white and minority applicants.

* 3a. Create Ad Impact and save new data set.
* 3b. Selection ratio and number selected variables.
* 3c. Save new data set.

* 4. Calculate overall scores for people in six ways: 
* 4a. Equal composite 
* 4b. GSM 
* 4c. Equal composite without written test
* 4d. 2 GSM
* 4e. Drop lowest score
* 4f. Random selection
* 4g. Fix for low scores

* 5. Select based on selection ratio.

* 6. Save all the corresponding means and correlations in .sav files.

* 7a. Read data from "means.sav" file in order to calculate adverse impact
* 7b. Drop all unneeded variables
* 7c. Calculate adverse impact

* 8a. Read data from "correlations.sav" file in order to find validity
* 8b. Drop all unneeded variables

* 9. Combine adverse impact and validity and save in "Adverse_Impact_Validity.sav" file

* 10. Identify adverse impact and validity values by grading method.  This makes graphing by grading method simpler.
	* (Create several temporary files (Adverse_Impact_Validity.txt), with one grading method in each one)
* 10b. Put all of the temporary files together, to prepare the data for graphing

* 11. Graphing Syntax

* Note: all "temp" files are overwritten each replication and only hold important data while the syntax is running.
*****************************.


*****************************.
* 1. Read 8 random variables for specified number of cases.

*Multivar produces a file called data.out."

DATA LIST FILE= 'C:\DATA.OUT'
   / written      1-6
     oral          8-13
     ppt           15-20
     consc       22-27
     gender      29-34
     eeo_gp     36-41
     job_per     43-48
     random     50-55.
     
* Sample data lines:     
*	 1.306  0.797  0.747  0.223 -0.922 -1.655  0.473  1.142 
*	-0.206  1.493  0.352  0.676  1.730 -0.808  1.939 -1.464 
*	 0.406  0.279  0.330  1.855  1.302  0.155  1.261 -0.369

FORMAT written   oral    ppt    consc     job_per      (F6.3).

CACHE.
EXECUTE.

VARIABLE LABEL  written "M/C Cog Ability"
                oral "Oral Exam"
                ppt 'Physical Performance Test'
                consc "Conscientiousness"
                eeo_gp 'EEO Group' 
                job_per  "Job Performance" .

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


*****************************.
* 2a. Create gender and eeo as categorical variables.

*** Create a known proportion of men and women.
* this creates a population with half men and half women.
IF (gender LE 0) gender = 0.
IF (gender GT 0 ) gender = 1.
VALUE LABEL gender 0 "Male" 1 "Female".

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

*****************************.
*** 2b. Create a known proportion of white and minority applicants.
* this creates a population with half minority and half white.
IF (eeo_gp LE 0) eeo_gp = 0.
IF (eeo_gp GT 0) eeo_gp = 1.
VALUE LABEL eeo_gp 0 "White" 1 "Minority".

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


*****************************.
* 3a. Create Ad Impact and Save new data set.

*** Set adverse impact.
COMPUTE dgenppt = 1.25.
COMPUTE deeocog = .72.
COMPUTE deeojob = .27.
COMPUTE deeocon = .07.
COMPUTE deeooral = .31.
IF (gender = 1) ppt= ppt-dgenppt .
IF (eeo_gp = 1) written = written -deeocog .
IF (eeo_gp = 1) job_per = job_per - deeojob .
IF (eeo_gp = 1) consc = consc + deeocon .
IF (eeo_gp = 1) oral = oral - deeooral .

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


*****************************.
* 3b. Selection ratio and number selected variables.

COMPUTE selratio = .01.
VARIABLE LABEL selratio "Selection Ratio" .
COMPUTE groupsize = 750000.
COMPUTE numselected = selratio * groupsize.
COMPUTE numselected_adj = numselected + ((1/4)*(.9*4)*groupsize) * selratio.
COMPUTE numfail = groupsize * .1.

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


*****************************.
* 3c. Save new data set.

SAVE OUTFILE =  'C:\DATA.OUT.sav'.

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


*****************************.
* 4. Calculate overall scores for people in six ways

GET FILE = 'C:\DATA.OUT.sav'.

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


*****************************.
* 4a. Equal Composite

COMPUTE p_trac4 = (oral   + written +  ppt + consc)/4.
VARIABLE LABEL p_trac4 "Predictor, Traditional Composite (oral written ppt consc)".

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


*****************************.
* 4b. GSM

COMPUTE highest = MAX (oral, written, ppt, consc).
*LIST highest oral written ppt consc /CASES = 10.
IF (highest = written) hivar = 1.
IF (highest = oral) hivar = 2.
IF (highest = ppt) hivar = 3.
IF (highest = consc) hivar = 4.
VARIABLE LABEL hivar 'Area of Highest Score'.
VALUE LABEL hivar 1 'Written' 2 'Oral' 3 'Physical Ability' 4 'Conscientiousness'.
    
VARIABLE LABEL highest "Predictor, Highest Score".

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


*****************************.
* 4c. Equal composite without written test

COMPUTE comp2 = (oral + ppt + consc)/3.
VARIABLE LABEL comp2 "Equal Composite Without Written Test".

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


*****************************.
* 4d. 2 GSM

COMPUTE hivar2 = 0.
IF (hivar = 1) hivar2 = MAX (oral, ppt, consc) .
IF (hivar = 2) hivar2 = MAX (written, ppt, consc) .
IF (hivar = 3) hivar2 = MAX (oral, written, consc) .
IF (hivar = 4) hivar2 = MAX (oral, ppt, written) .

COMPUTE twohighest = (highest + hivar2)/2.
VARIABLE LABEL twohighest  "Two Highest Scores".

*****************************.
* 4e. Drop lowest score

COMPUTE lowest = MIN (oral, written, ppt, consc).
IF (lowest = written) lovar = 1.
IF (lowest = oral) lovar = 2.
IF (lowest = ppt) lovar = 3.
IF (lowest = consc) lovar = 4.
VARIABLE LABEL lovar 'Area of Lowest Score'.
VALUE LABEL lovar 1 'Written' 2 'Oral' 3 'Physical Ability' 4 'Conscientiousness'.

COMPUTE lovareq = 0.
IF (lovar = 1) lovareq = written.
IF (lovar = 2) lovareq = oral.
IF (lovar = 3) lovareq = ppt.
IF (lovar = 4) lovareq = consc.

COMPUTE droplowest = ((oral + written + ppt + consc) - (lovareq))/3.

VARIABLE LABEL droplowest "Drop Lowest Score".

*****************************.
* 4f. Random selection

* Rank based on variable "random"

*****************************.
* 4g. Create a variable to identify (and later OMIT) lowest 10% on any predictor from innovative scoring.

RANK  VARIABLES = written oral ppt consc (a).
COMPUTE low = 0.
IF (rwritten GE numfail AND roral GE numfail AND rppt GE numfail AND rconsc GE numfail) low = 1.
VALUE LABEL  low  0 'Has a low score'  1  'No low score'.
VARIABLE LABEL low "Indicator for any low score".

CROSSTABS  /TABLES=low  BY eeo_gp gender /FORMAT= AVALUE TABLES    /CELLS= COUNT    /COUNT ROUND CELL .

*****************************.
SAVE OUTFILE = 'C:\DATA.OUT.sav' /DROP rwritten.

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


*****************************.
* 5. Select Based on selection ratio.

GET FILE = 'C:\DATA.OUT.sav'.

* Assume will appoint top 01%.
RANK  VARIABLES = p_trac4  highest written comp2 droplowest twohighest random (D).

* GSM
* selected variable = 1 if selected, otherwise, 0:

COMPUTE selected_women_rhighest_01 = 0.
IF (rhighest LE numselected_adj AND gender = 1 AND low = 1) selected_women_rhighest_01 = 1.

COMPUTE selected_men_rhighest_01 = 0.
IF (rhighest LE numselected_adj AND gender = 0 AND low = 1) selected_men_rhighest_01 = 1.

COMPUTE selected_minority_rhighest_01 = 0.
IF (rhighest LE numselected_adj AND eeo_gp = 1 AND low = 1) selected_minority_rhighest_01 = 1.

COMPUTE selected_white_rhighest_01 = 0.
IF (rhighest LE numselected_adj AND eeo_gp = 0 AND low = 1) selected_white_rhighest_01 = 1.


*Written
*set the appropiate "selected" variable to 1, otherwise, leave it at 0:

COMPUTE selected_women_rwritten_01 = 0.
IF (rwritten LE numselected AND gender = 1 ) selected_women_rwritten_01 = 1.

COMPUTE selected_men_rwritten_01 = 0.
IF (rwritten LE numselected AND gender = 0 ) selected_men_rwritten_01 = 1.

COMPUTE selected_minority_rwritten_01 = 0.
IF (rwritten LE numselected AND eeo_gp = 1) selected_minority_rwritten_01 = 1.

COMPUTE selected_white_rwritten_01 = 0.
IF (rwritten LE numselected AND eeo_gp = 0) selected_white_rwritten_01 = 1.


*Composite
*set the appropiate "selected" variable to 1, otherwise, leave it at 0:

COMPUTE selected_women_rp_trac4_01 = 0.
IF (rp_trac4 LE numselected AND gender = 1 ) selected_women_rp_trac4_01 = 1.

COMPUTE selected_men_rp_trac4_01 = 0.
IF (rp_trac4 LE numselected AND gender = 0 ) selected_men_rp_trac4_01 = 1.

COMPUTE selected_minority_rp_trac4_01 = 0.
IF (rp_trac4 LE numselected AND eeo_gp = 1) selected_minority_rp_trac4_01 = 1.

COMPUTE selected_white_rp_trac4_01 = 0.
IF (rp_trac4 LE numselected AND eeo_gp = 0) selected_white_rp_trac4_01 = 1.

*Comp2

COMPUTE selected_women_rcomp2_01 = 0.
IF (rcomp2 LE numselected AND gender = 1) selected_women_rcomp2_01 = 1.

COMPUTE selected_men_rcomp2_01 = 0.
IF (rcomp2 LE numselected AND gender = 0) selected_men_rcomp2_01 = 1.

COMPUTE selected_minority_rcomp2_01 = 0.
IF (rcomp2 LE numselected AND eeo_gp = 1) selected_minority_rcomp2_01 = 1.

COMPUTE selected_white_rcomp2_01 = 0.
IF (rcomp2 LE numselected AND eeo_gp = 0) selected_white_rcomp2_01 = 1.

*Twohighest

COMPUTE selected_women_rtwohighest_01 = 0.
IF (rtwohigh LE numselected_adj AND gender = 1 AND low = 1) selected_women_rtwohighest_01 = 1.

COMPUTE selected_men_rtwohighest_01 = 0.
IF (rtwohigh LE numselected_adj AND gender = 0 AND low = 1) selected_men_rtwohighest_01 = 1.

COMPUTE selected_minority_rtwohighest_01 = 0.
IF (rtwohigh LE numselected_adj AND eeo_gp = 1 AND low = 1) selected_minority_rtwohighest_01 = 1.

COMPUTE selected_white_rtwohighest_01 = 0.
IF (rtwohigh LE numselected_adj AND eeo_gp = 0 AND low = 1) selected_white_rtwohighest_01 = 1.

*Droplowest

COMPUTE selected_women_rdroplowest_01 = 0.
IF (rdroplow LE numselected_adj AND gender = 1 AND low = 1) selected_women_rdroplowest_01 = 1.

COMPUTE selected_men_rdroplowest_01 = 0.
IF (rdroplow LE numselected_adj AND gender = 0 AND low = 1) selected_men_rdroplowest_01 = 1.

COMPUTE selected_minority_rdroplowest_01 = 0.
IF (rdroplow LE numselected_adj AND eeo_gp = 1 AND low = 1) selected_minority_rdroplowest_01 = 1.

COMPUTE selected_white_rdroplowest_01 = 0.
IF (rdroplow LE numselected_adj AND eeo_gp = 0 AND low = 1) selected_white_rdroplowest_01 = 1.

*Random

COMPUTE selected_women_rrandom_01 = 0.
IF (rrandom LE numselected AND gender = 1) selected_women_rrandom_01 = 1.

COMPUTE selected_men_rrandom_01 = 0.
IF (rrandom LE numselected AND gender = 0) selected_men_rrandom_01 = 1.

COMPUTE selected_minority_rrandom_01 = 0.
IF (rrandom LE numselected AND eeo_gp = 1) selected_minority_rrandom_01 = 1.

COMPUTE selected_white_rrandom_01 = 0.
IF (rrandom LE numselected AND eeo_gp = 0) selected_white_rrandom_01 = 1.

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


*****************************.
* 6. Save all the corresponding means and correlations in .sav files.

/*This is a macro definition
/*The last line of this syntax invokes this macro with proper parameters

DEFINE abc(sample_size=!TOKENS(1) /no_of_samples=!TOKENS(1) /vars = !CMDEND)

/*IF statement creates variables sample_no
/*variable sample_no contains number of sample that each case belongs
IF ($CASENUM<=!sample_size*!no_of_samples) sample_no=TRUNC(($CASENUM-1)/!sample_size)+1.

/*The OMS function saves output tables to sav file .
/*Only correlation tables are selected
OMS
  /SELECT TABLES
  /IF SUBTYPES=['Correlations']
  /DESTINATION FORMAT=SAV NUMBERED='runID'
               OUTFILE='C:\correlations.sav'.

/*This loop computes a filter variable and correlations for each sample
!DO !idx=1 !TO !no_of_samples
/*computes a filter and select cases belong to sample number !idx
  COMPUTE fi=(sample_no=!idx).
  FILTER by fi.

/*Correlation tables are created for each sample and all variables declared in parameter 'vars'
/*These tables are saved by OMS function
  CORRELATIONS
    /VARIABLES= !vars         /*change number of variables 
    /PRINT=TWOTAIL NOSIG
    /MISSING=PAIRWISE .
!DOEND.
OMSEND.

OMS
  /SELECT TABLES
  /IF SUBTYPES=['Report']
  /DESTINATION FORMAT=SAV NUMBERED='runID'
               OUTFILE='C:\means.sav'.

!DO !idx=1 !TO !no_of_samples
  COMPUTE fi=(sample_no=!idx).
  FILTER by fi.
  MEANS VARIABLES= !vars BY eeo_gp.
!DOEND.
OMSEND.
/*Opens SPSS data file
GET FILE='C:\means.sav'.

!ENDDEFINE.

abc sample_size=1000 no_of_samples=1000
vars=written oral ppt consc gender eeo_gp job_per p_trac4 comp2 highest twohighest droplowest random
selected_women_rhighest_01 selected_men_rhighest_01 selected_minority_rhighest_01 selected_white_rhighest_01
selected_women_rwritten_01 selected_men_rwritten_01 selected_minority_rwritten_01 selected_white_rwritten_01
selected_women_rp_trac4_01 selected_men_rp_trac4_01 selected_minority_rp_trac4_01 selected_white_rp_trac4_01 
selected_women_rrandom_01 selected_men_rrandom_01 selected_minority_rrandom_01 selected_white_rrandom_01
selected_women_rdroplowest_01 selected_men_rdroplowest_01 selected_minority_rdroplowest_01 selected_white_rdroplowest_01
selected_women_rtwohighest_01 selected_men_rtwohighest_01 selected_minority_rtwohighest_01 selected_white_rtwohighest_01
selected_women_rcomp2_01 selected_men_rcomp2_01 selected_minority_rcomp2_01 selected_white_rcomp2_01
.

/*change size of each sample, number of samples and list of variables above.
/*If there are 3 original variables (V1 V2 V3) and 2 new variables (V4 V5) use form: 'V1 TO V5' or 'V1 V2 V3 V4 V5'
.

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

*****************************.
* 7a. Read data from "means.sav" file in order to calculate adverse impact

GET FILE='C:\means.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\means.txt'
  /TYPE=TAB   
  /KEEP=runID Var1 Var2 MCCogAbility PredictorTraditionalCompositeoralwrittenpptconsc EqualCompositeWithoutWrittenTest PredictorHighestScore TwoHighestScores DropLowestScore random selected_women_rhighest_01
    selected_men_rhighest_01 selected_minority_rhighest_01 selected_white_rhighest_01 selected_women_rwritten_01 selected_men_rwritten_01 selected_minority_rwritten_01 selected_white_rwritten_01 selected_women_rp_trac4_01
    selected_men_rp_trac4_01 selected_minority_rp_trac4_01 selected_white_rp_trac4_01 selected_women_rrandom_01 selected_men_rrandom_01 selected_minority_rrandom_01 selected_white_rrandom_01 selected_women_rdroplowest_01
    selected_men_rdroplowest_01 selected_minority_rdroplowest_01 selected_white_rdroplowest_01 selected_women_rtwohighest_01 selected_men_rtwohighest_01 selected_minority_rtwohighest_01 selected_white_rtwohighest_01
    selected_women_rcomp2_01 selected_men_rcomp2_01 selected_minority_rcomp2_01 selected_white_rcomp2_01
  /REPLACE.
  
  EXECUTE.


GET DATA  /TYPE = TXT
 /FILE = 'C:\means.txt'
 /DELCASE = VARIABLES 342
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 run F2.1
 V2 A5
 V3 A4
 V4 F19.2
 V5 F17.2
 V6 F17.2
 V7 F16.2
 V8 F16.2
 V9 F16.2
 V10 F21.2
 V11 F18.2
 V12 F18.2
 V13 F1.0
 sel_whi_highest F18.2
 V15 F18.2
 V16 F18.2
 V17 F1.0
 sel_whi_written F18.2
 V19 F18.2
 V20 F18.2
 V21 F1.0
 sel_whi_comp F18.2
 V23 F18.2
 V24 F18.2
 V25 F1.0
 sel_whi_random F18.2
 V27 F18.2
 V28 F18.2
 V29 F1.0
 sel_whi_droplowest F18.2
 V31 F18.2
 V32 F18.2
 V33 F1.0
 sel_whi_twohighest F18.2
 V35 F18.2
 V36 F18.2
 V37 F1.0
 sel_whi_comp2 F18.2
 V39 F2.1
 V40 A5
 V41 A1
 V42 F3.2
 V43 F3.2
 V44 F3.2
 V45 F3.2
 V46 F3.2
 V47 F3.2
 V48 F3.2
 V49 F3.2
 V50 F3.2
 V51 F3.2
 V52 F3.2
 V53 F3.2
 V54 F3.2
 V55 F3.2
 V56 F3.2
 V57 F3.2
 V58 F3.2
 V59 F3.2
 V60 F3.2
 V61 F3.2
 V62 F3.2
 V63 F3.2
 V64 F3.2
 V65 F3.2
 V66 F3.2
 V67 F3.2
 V68 F3.2
 V69 F3.2
 V70 F3.2
 V71 F3.2
 V72 F3.2
 V73 F3.2
 V74 F3.2
 V75 F3.2
 V76 F3.2
 V77 F2.1
 V78 A5
 V79 A14
 V80 F16.2
 V81 F16.2
 V82 F16.2
 V83 F16.2
 V84 F16.2
 V85 F16.2
 V86 F16.2
 V87 F17.2
 V88 F17.2
 V89 F1.0
 V90 F17.2
 V91 F17.2
 V92 F17.2
 V93 F1.0
 V94 F17.2
 V95 F17.2
 V96 F17.2
 V97 F1.0
 V98 F17.2
 V99 F17.2
 V100 F17.2
 V101 F1.0
 V102 F17.2
 V103 F17.2
 V104 F17.2
 V105 F1.0
 V106 F17.2
 V107 F17.2
 V108 F17.2
 V109 F1.0
 V110 F17.2
 V111 F17.2
 V112 F17.2
 V113 F1.0
 V114 F17.2
 V115 F2.1
 V116 A8
 V117 A4
 V118 F17.2
 V119 F17.2
 V120 F17.2
 V121 F16.2
 V122 F16.2
 V123 F19.2
 V124 F19.2
 V125 F18.2
 V126 F18.2
 sel_min_highest F18.2
 V128 F1.0
 V129 F17.2
 V130 F3.2
 sel_min_written F17.2
 V132 F1.0
 V133 F17.2
 V134 F18.2
 sel_min_comp F18.2
 V136 F1.0
 V137 F18.2
 V138 F18.2
 sel_min_random F18.2
 V140 F1.0
 V141 F17.2
 V142 F17.2
 sel_min_droplowest F17.2
 V144 F1.0
 V145 F18.2
 V146 F17.2
 sel_min_twohighest F18.2
 V148 F1.0
 V149 F17.2
 V150 F18.2
 sel_min_comp2 F18.2
 V152 F1.0
 V153 F2.1
 V154 A8
 V155 A1
 V156 F3.2
 V157 F3.2
 V158 F3.2
 V159 F3.2
 V160 F3.2
 V161 F3.2
 V162 F3.2
 V163 F3.2
 V164 F3.2
 V165 F3.2
 V166 F3.2
 V167 F3.2
 V168 F3.2
 V169 F3.2
 V170 F3.2
 V171 F3.2
 V172 F3.2
 V173 F3.2
 V174 F3.2
 V175 F3.2
 V176 F3.2
 V177 F3.2
 V178 F3.2
 V179 F3.2
 V180 F3.2
 V181 F3.2
 V182 F3.2
 V183 F3.2
 V184 F3.2
 V185 F3.2
 V186 F3.2
 V187 F3.2
 V188 F3.2
 V189 F3.2
 V190 F3.2
 V191 F2.1
 V192 A8
 V193 A14
 V194 F16.2
 V195 F16.2
 V196 F16.2
 V197 F16.2
 V198 F16.2
 V199 F16.2
 V200 F16.2
 V201 F17.2
 V202 F17.2
 V203 F17.2
 V204 F1.0
 V205 F17.2
 V206 F2.1
 V207 F17.2
 V208 F1.0
 V209 F16.2
 V210 F17.2
 V211 F17.2
 V212 F1.0
 V213 F17.2
 V214 F17.2
 V215 F17.2
 V216 F1.0
 V217 F16.2
 V218 F17.2
 V219 F17.2
 V220 F1.0
 V221 F17.2
 V222 F17.2
 V223 F17.2
 V224 F1.0
 V225 F16.2
 V226 F17.2
 V227 F17.2
 V228 F1.0
 V229 F2.1
 V230 A5
 V231 A4
 written F20.2
 comp F10.2
 comp2 F17.2
 highest F16.2
 twohighest F16.2
 droplowest F16.2
 random F19.2
 sel_wom_highest F18.2
 sel_men_highest F18.2
 V241 F4.2
 V242 F18.2
 sel_wom_written F18.2
 sel_men_written F18.2
 V245 F4.2
 V246 F18.2
 sel_wom_comp F4.2
 sel_men_comp F18.2
 V249 F4.2
 V250 F18.2
 sel_wom_random F18.2
 sel_men_random F18.2
 V253 F4.2
 V254 F18.2
 sel_wom_droplowest F18.2
 sel_men_droplowest F18.2
 V257 F4.2
 V258 F18.2
 sel_wom_twohighest F18.2
 sel_men_twohighest F18.2
 V261 F4.2
 V262 F18.2
 sel_wom_comp2 F4.2
 sel_men_comp2 F18.2
 V265 F4.2
 V266 F18.2
 V267 F2.1
 V268 A5
 V269 A1
 V270 F4.2
 V271 F4.2
 V272 F4.2
 V273 F4.2
 V274 F4.2
 V275 F4.2
 V276 F4.2
 V277 F4.2
 V278 F4.2
 V279 F4.2
 V280 F4.2
 V281 F4.2
 V282 F4.2
 V283 F4.2
 V284 F4.2
 V285 F4.2
 V286 F4.2
 V287 F4.2
 V288 F4.2
 V289 F4.2
 V290 F4.2
 V291 F4.2
 V292 F4.2
 V293 F4.2
 V294 F4.2
 V295 F4.2
 V296 F4.2
 V297 F4.2
 V298 F4.2
 V299 F4.2
 V300 F4.2
 V301 F4.2
 V302 F4.2
 V303 F4.2
 V304 F4.2
 V305 F2.1
 V306 A5
 V307 A14
 written_dev F16.2
 comp_dev F16.2
 comp2_dev F16.2
 highest_dev F16.2
 twohighest_dev F16.2
 droplowest_dev F16.2
 random_dev F16.2
 V315 F17.2
 V316 F17.2
 V317 F17.2
 V318 F17.2
 V319 F17.2
 V320 F17.2
 V321 F17.2
 V322 F17.2
 V323 F17.2
 V324 F17.2
 V325 F17.2
 V326 F17.2
 V327 F17.2
 V328 F17.2
 V329 F17.2
 V330 F17.2
 V331 F17.2
 V332 F17.2
 V333 F17.2
 V334 F17.2
 V335 F17.2
 V336 F17.2
 V337 F17.2
 V338 F17.2
 V339 F17.2
 V340 F17.2
 V341 F17.2
 V342 F17.2
 .

CACHE.
EXECUTE.
SAVE OUTFILE = 'C:\temp.sav'.

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

*****************************.
* 7b. Drop all unneeded variables

GET FILE = 'C:\temp.sav'
 /DROP=V2
V3
V4
V5
V6
V7
V8
V9
V10
V11
V12
V13
V15
V16
V17
V19
V20
V21
V23
V24
V25
V27
V28
V29
V31
V32
V33
V35
V36
V37
V39
V40
V41
V42
V43
V44
V45
V46
V47
V48
V49
V50
V51
V52
V53
V54
V55
V56
V57
V58
V59
V60
V61
V62
V63
V64
V65
V66
V67
V68
V69
V70
V71
V72
V73
V74
V75
V76
V77
V78
V79
V80
V81
V82
V83
V84
V85
V86
V87
V88
V89
V90
V91
V92
V93
V94
V95
V96
V97
V98
V99
V100
V101
V102
V103
V104
V105
V106
V107
V108
V109
V110
V111
V112
V113
V114
V115
V116
V117
V118
V119
V120
V121
V122
V123
V124
V125
V126
V128
V129
V130
V132
V133
V134
V136
V137
V138
V140
V141
V142
V144
V145
V146
V148
V149
V150
V152
V153
V154
V155
V156
V157
V158
V159
V160
V161
V162
V163
V164
V165
V166
V167
V168
V169
V170
V171
V172
V173
V174
V175
V176
V177
V178
V179
V180
V181
V182
V183
V184
V185
V186
V187
V188
V189
V190
V191
V192
V193
V194
V195
V196
V197
V198
V199
V200
V201
V202
V203
V204
V205
V206
V207
V208
V209
V210
V211
V212
V213
V214
V215
V216
V217
V218
V219
V220
V221
V222
V223
V224
V225
V226
V227
V228
V229
V230
V231
V241
V242
V245
V246
V249
V250
V253
V254
V257
V258
V261
V262
V265
V266
V267
V268
V269
V270
V271
V272
V273
V274
V275
V276
V277
V278
V279
V280
V281
V282
V283
V284
V285
V286
V287
V288
V289
V290
V291
V292
V293
V294
V295
V296
V297
V298
V299
V300
V301
V302
V303
V304
V305
V306
V307
V315
V316
V317
V318
V319
V320
V321
V322
V323
V324
V325
V326
V327
V328
V329
V330
V331
V332
V333
V334
V335
V336
V337
V338
V339
V340
V341
V342
.

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


*****************************.
* 7c. Calculate adverse impact

COMPUTE ad_imp_eeo_written = sel_min_written / sel_whi_written .
COMPUTE ad_imp_gen_written = sel_wom_written / sel_men_written .

COMPUTE ad_imp_eeo_comp = sel_min_comp / sel_whi_comp .
COMPUTE ad_imp_gen_comp = sel_wom_comp / sel_men_comp .

COMPUTE ad_imp_eeo_comp2 = sel_min_comp2 / sel_whi_comp2 .
COMPUTE ad_imp_gen_comp2 = sel_wom_comp2 / sel_men_comp2 .

COMPUTE ad_imp_eeo_highest = sel_min_highest / sel_whi_highest.
COMPUTE ad_imp_gen_highest = sel_wom_highest / sel_men_highest.

COMPUTE ad_imp_eeo_twohighest = sel_min_twohighest / sel_whi_twohighest.
COMPUTE ad_imp_gen_twohighest = sel_wom_twohighest / sel_men_twohighest.

COMPUTE ad_imp_eeo_droplowest = sel_min_droplowest / sel_whi_droplowest.
COMPUTE ad_imp_gen_droplowest = sel_wom_droplowest / sel_men_droplowest.

COMPUTE ad_imp_eeo_random = sel_min_random / sel_whi_random .
COMPUTE ad_imp_gen_random = sel_wom_random / sel_men_random .

CACHE.
EXECUTE.

SAVE OUTFILE = 'C:\temp.sav'.

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


*****************************.
* 8a. Read data from "correlations.sav" file in order to find validity

GET FILE='C:\correlations.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\correlations.txt'
  /TYPE=TAB   
  /KEEP=Var1 JobPerformance
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\correlations.txt'
 /DELCASE = VARIABLES 246
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
V1 A8
validity_written F8.4
V3 A8
V4 F8.4
V5 A8
V6 F8.4
V7 A8
V8 F8.4
V9 A8
V10 F8.4
V11 A8
V12 F8.4
V13 A8
V14 F8.4
V15 A8
V16 F8.4
V17 A8
V18 F8.4
V19 A8
V20 F8.4
V21 A8
V22 F8.4
V23 A8
V24 F8.4
V25 A8
V26 F8.4
V27 A8
V28 F8.4
V29 A8
V30 F8.4
V31 A8
V32 F8.4
V33 A8
V34 F8.4
V35 A8
V36 F8.4
V37 A8
V38 F8.4
V39 A8
V40 F8.4
V41 A8
V42 F8.4
V43 A8
validity_comp F8.4
V45 A8
V46 F8.4
V47 A8
V48 F8.4
V49 A8
validity_comp2 F8.4
V51 A8
V52 F8.4
V53 A8
V54 F8.4
V55 A8
validity_highest F8.4
V57 A8
V58 F8.4
V59 A8
V60 F8.4
V61 A8
validity_twohighest F8.4
V63 A8
V64 F8.4
V65 A8
V66 F8.4
V67 A8
validity_droplowest F8.4
V69 A8
V70 F8.4
V71 A8
V72 F8.4
V73 A8
validity_random F8.4
V75 A8
V76 F8.4
V77 A8
V78 F8.4
V79 A8
V80 F8.4
V81 A8
V82 F8.4
V83 A8
V84 F8.4
V85 A8
V86 F8.4
V87 A8
V88 F8.4
V89 A8
V90 F8.4
V91 A8
V92 F8.4
V93 A8
V94 F8.4
V95 A8
V96 F8.4
V97 A8
V98 F8.4
V99 A8
V100 F8.4
V101 A8
V102 F8.4
V103 A8
V104 F8.4
V105 A8
V106 F8.4
V107 A8
V108 F8.4
V109 A8
V110 F8.4
V111 A8
V112 F8.4
V113 A8
V114 F8.4
V115 A8
V116 F8.4
V117 A8
V118 F8.4
V119 A8
V120 F8.4
V121 A8
V122 F8.4
V123 A8
V124 F8.4
V125 A8
V126 F8.4
V127 A8
V128 F8.4
V129 A8
V130 F8.4
V131 A8
V132 F8.4
V133 A8
V134 F8.4
V135 A8
V136 F8.4
V137 A8
V138 F8.4
V139 A8
V140 F8.4
V141 A8
V142 F8.4
V143 A8
V144 F8.4
V145 A8
V146 F8.4
V147 A8
V148 F8.4
V149 A8
V150 F8.4
V151 A8
V152 F8.4
V153 A8
V154 F8.4
V155 A8
V156 F8.4
V157 A8
V158 F8.4
V159 A8
V160 F8.4
V161 A8
V162 F8.4
V163 A8
V164 F8.4
V165 A8
V166 F8.4
V167 A8
V168 F8.4
V169 A8
V170 F8.4
V171 A8
V172 F8.4
V173 A8
V174 F8.4
V175 A8
V176 F8.4
V177 A8
V178 F8.4
V179 A8
V180 F8.4
V181 A8
V182 F8.4
V183 A8
V184 F8.4
V185 A8
V186 F8.4
V187 A8
V188 F8.4
V189 A8
V190 F8.4
V191 A8
V192 F8.4
V193 A8
V194 F8.4
V195 A8
V196 F8.4
V197 A8
V198 F8.4
V199 A8
V200 F8.4
V201 A8
V202 F8.4
V203 A8
V204 F8.4
V205 A8
V206 F8.4
V207 A8
V208 F8.4
V209 A8
V210 F8.4
V211 A8
V212 F8.4
V213 A8
V214 F8.4
V215 A8
V216 F8.4
V217 A8
V218 F8.4
V219 A8
V220 F8.4
V221 A8
V222 F8.4
V223 A8
V224 F8.4
V225 A8
V226 F8.4
V227 A8
V228 F8.4
V229 A8
V230 F8.4
V231 A8
V232 F8.4
V233 A8
V234 F8.4
V235 A8
V236 F8.4
V237 A8
V238 F8.4
V239 A8
V240 F8.4
V241 A8
V242 F8.4
V243 A8
V244 F8.4
V245 A8
V246 F8.4
.
CACHE.
EXECUTE.

SAVE OUTFILE = 'C:\temp1.sav'.

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


*****************************.
* 8b. Drop all unneeded variables

GET FILE = 'C:\temp1.sav'
 /DROP=V1
V3
V4
V5
V6
V7
V8
V9
V10
V11
V12
V13
V14
V15
V16
V17
V18
V19
V20
V21
V22
V23
V24
V25
V26
V27
V28
V29
V30
V31
V32
V33
V34
V35
V36
V37
V38
V39
V40
V41
V42
V43
V45
V46
V47
V48
V49
V51
V52
V53
V54
V55
V57
V58
V59
V60
V61
V63
V64
V65
V66
V67
V69
V70
V71
V72
V73
V75
V76
V77
V78
V79
V80
V81
V82
V83
V84
V85
V86
V87
V88
V89
V90
V91
V92
V93
V94
V95
V96
V97
V98
V99
V100
V101
V102
V103
V104
V105
V106
V107
V108
V109
V110
V111
V112
V113
V114
V115
V116
V117
V118
V119
V120
V121
V122
V123
V124
V125
V126
V127
V128
V129
V130
V131
V132
V133
V134
V135
V136
V137
V138
V139
V140
V141
V142
V143
V144
V145
V146
V147
V148
V149
V150
V151
V152
V153
V154
V155
V156
V157
V158
V159
V160
V161
V162
V163
V164
V165
V166
V167
V168
V169
V170
V171
V172
V173
V174
V175
V176
V177
V178
V179
V180
V181
V182
V183
V184
V185
V186
V187
V188
V189
V190
V191
V192
V193
V194
V195
V196
V197
V198
V199
V200
V201
V202
V203
V204
V205
V206
V207
V208
V209
V210
V211
V212
V213
V214
V215
V216
V217
V218
V219
V220
V221
V222
V223
V224
V225
V226
V227
V228
V229
V230
V231
V232
V233
V234
V235
V236
V237
V238
V239
V240
V241
V242
V243
V244
V245
V246
.
SAVE OUTFILE = 'C:\temp1.sav'.

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


*****************************.
* 9. Combine adverse impact and validity and saive in "Adverse_Impact_Validity.sav" file

MATCH FILES FILE = 'C:\temp2.sav'
     / FILE = 'C:\temp3.sav'
.

SAVE OUTFILE = 'C:\Adverse_Impact_Validity.sav'.

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


*****************************.
* 10. Identify adverse impact and validity values by grading method.  This makes graphing by grading method simpler.
	* (Create several temporary files (Adverse_Impact_Validity.txt), with one grading method in each one)

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

**************************************

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_1.txt'
  /TYPE=TAB   
  /KEEP=validity_written ad_imp_eeo_written ad_imp_gen_written
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_1.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_written F16.2
 ad_imp_eeo_written F16.2
 ad_imp_gen_written F16.2
 .
COMPUTE method=1

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_1.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_2.txt'
  /TYPE=TAB   
  /KEEP=validity_comp ad_imp_eeo_comp ad_imp_gen_comp
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_2.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_comp F16.2
 ad_imp_eeo_comp F16.2
 ad_imp_gen_comp F16.2
 .
COMPUTE method=2

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_2.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_3.txt'
  /TYPE=TAB   
  /KEEP=validity_comp2 ad_imp_eeo_comp2 ad_imp_gen_comp2
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_3.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_comp2 F16.2
 ad_imp_eeo_comp2 F16.2
 ad_imp_gen_comp2 F16.2
 .
COMPUTE method=3

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_3.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_4.txt'
  /TYPE=TAB   
  /KEEP=validity_highest ad_imp_eeo_highest ad_imp_gen_highest
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_4.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_highest F16.2
 ad_imp_eeo_highest F16.2
 ad_imp_gen_highest F16.2
 .
COMPUTE method=4

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_4.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_5.txt'
  /TYPE=TAB   
  /KEEP=validity_twohighest ad_imp_eeo_twohighest ad_imp_gen_twohighest
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_5.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_twohighest F16.2
 ad_imp_eeo_twohighest F16.2
 ad_imp_gen_twohighest F16.2
 .
COMPUTE method=5

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_5.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_6.txt'
  /TYPE=TAB   
  /KEEP=validity_droplowest ad_imp_eeo_droplowest ad_imp_gen_droplowest
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_6.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_droplowest F16.2
 ad_imp_eeo_droplowest F16.2
 ad_imp_gen_droplowest F16.2
 .
COMPUTE method=6

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_6.sav'.

********************************************

GET FILE = 'C:\Adverse_Impact_Validity.sav'.

SAVE TRANSLATE
  /OUTFILE= 'C:\Adverse_Impact_Validity_7.txt'
  /TYPE=TAB   
  /KEEP=validity_random ad_imp_eeo_random ad_imp_gen_random
  /REPLACE.
  
  EXECUTE.

GET DATA  /TYPE = TXT
 /FILE = 'C:\Adverse_Impact_Validity_7.txt'
 /DELCASE = LINE
 /DELIMITERS = "\t"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 1
 /IMPORTCASE = ALL
 /VARIABLES =
 validity_random F16.2
 ad_imp_eeo_random F16.2
 ad_imp_gen_random F16.2
 .
COMPUTE method=7

CACHE.
EXECUTE.

SAVE OUTFILE='C:\Adverse_Impact_Validity_7.sav'.

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


*****************************.
* 10b. Put all of the temporary files together, to prepare the data for graphing

MATCH FILES FILE ='C:\Adverse_Impact_Validity_1.sav'
     / FILE ='C:\Adverse_Impact_Validity_2.sav'
     / FILE ='C:\Adverse_Impact_Validity_3.sav'
     / FILE ='C:\Adverse_Impact_Validity_4.sav'
     / FILE ='C:\Adverse_Impact_Validity_5.sav'
     / FILE ='C:\Adverse_Impact_Validity_6.sav'
     / FILE ='C:\Adverse_Impact_Validity_7.sav'
     /BY method.

VARIABLE LABEL method 'Grading method'.

VALUE LABEL method 1 'Written' 2 'Composite' 3 'Composite Without Written' 4 'Greatest Strength' 5 'Two Greatest Strengths' 6 'Drop Lowest Score' 7 'Random Selection'.

COMPUTE sr=.01.
COMPUTE eeo=.5.
COMPUTE ss=10000.

CACHE.
EXECUTE.

SAVE OUTFILE='C:\graphdata_sr1.sav'
.

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


*****************************.
* 11. Graphing Syntax


***********************************************
***********************************************
* Note the Igraph command below presumes you have run 3 replications of the above with 3 selection ratios.
* (The 3 graphdata.sav files from the 3 replications are combined in the syntax immediately below.)
***********************************************
***********************************************


*****************************.
* a. Combine means for all replications that have different selection ratios but similar everything else.  This makes it easier to plot the data by selection ratio.  

ADD FILES FILE ='C:\graphdata_sr1.sav'
     / FILE='C:\graphdata_sr2.sav'
     / FILE='C:\graphdata_sr3.sav'
     / BY method.

VARIABLE LABEL sr 'Selection Ratio'.
VALUE LABEL sr .9 'All Passers'.

CACHE.
EXECUTE.

SAVE OUTFILE='C:\graphdata_sr_all.sav'.

*****************************.
* b. Graph Adverse Impact eeo

GET FILE='C:\graphdata_sr_all.sav'.

IGRAPH /VIEWNAME='Boxplot' /X1 = VAR(category_set1) TYPE = CATEGORICAL 
  /Y = VAR(value_set1) TYPE = SCALE  /COLOR = VAR(sr)
  TYPE = CATEGORICAL /COORDINATE = VERTICAL 
  /COMBINE  GROUP=VAR(category_set1) TYPE=CATEGORICAL 'Grading Method' RESPONSE=VAR(value_set1) TYPE=SCALE 'Adverse Impact EEO' VAR(ad_imp_eeo_highest) 'Greatest Strength' 
  VAR(ad_imp_eeo_twohighest) 'Two Greatest Strengths' VAR(ad_imp_eeo_droplowest) 'Drop Lowest Score' VAR(ad_imp_eeo_comp2) 'Composite Without M/C' 
  VAR(ad_imp_eeo_random) 'Random Selection' VAR(ad_imp_eeo_comp) 'Composite' VAR(ad_imp_eeo_written) 'M/C Cognitive Ability'  
  /TITLE='Figure 1. Adverse Impact EEO by Grading Method' /SUBTITLE='Proportion of Minorities = .1, Applicant Group Size = 1000, Number of Replications = 750'
 /X1LENGTH=4.0 /YLENGTH=4.0 /X2LENGTH=4.0 /CHARTLOOK='C:\Program Files\SPSS\Looks\Grayscale.clo' /CATORDER VAR(category_set1) (ASCENDING VALUES OMITEMPTY)
 /SCALERANGE = VAR(value_set1) MIN=0.000000 MAX=2 /BOX OUTLIERS = OFF EXTREME = OFF MEDIAN = ON WHISKER = T.
EXE.

*****************************.
* c. Graph Adverse Impact gender

GET FILE='C:\graphdata_sr_all.sav'.

IGRAPH /VIEWNAME='Boxplot' /X1 = VAR(category_set1) TYPE = CATEGORICAL /Y = VAR(value_set1) TYPE = SCALE /COLOR = VAR(sr)
  TYPE = CATEGORICAL  /COORDINATE = VERTICAL /COMBINE  GROUP=VAR(category_set1) TYPE=CATEGORICAL 'Grading Method' RESPONSE=VAR(value_set1) TYPE=SCALE 'Adverse Impact Gender' 
  VAR(ad_imp_gen_highest) 'Greatest Strength' VAR(ad_imp_gen_twohighest) 'Two Greatest Strengths' VAR(ad_imp_gen_droplowest) 'Drop Lowest Score' 
  VAR(ad_imp_gen_comp2) 'Composite Without M/C' VAR(ad_imp_gen_random) 'Random Selection' VAR(ad_imp_gen_comp) 'Composite' VAR(ad_imp_gen_written) 'M/C Cognitive Ability'  
   /TITLE='Figure 2. Adverse Impact Gender by Grading Method' /SUBTITLE='Proportion of Minorities = .1, Applicant Group Size = 1000, Number of Replications = 750' /X1LENGTH=4.0
 /YLENGTH=4.0 /X2LENGTH=4.0 /CHARTLOOK='C:\Program Files\SPSS\Looks\Grayscale.clo' /CATORDER VAR(category_set1) (ASCENDING VALUES OMITEMPTY) /SCALERANGE =
  VAR(value_set1) MIN=0.000000 MAX=2.000000 /BOX OUTLIERS = OFF EXTREME = OFF MEDIAN = ON WHISKER = T.
EXE.

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