GET FILE='C:\data\PISA2000_OECD\Data\Pupil\pupil2000.sav'. * Create flag for whether CCC module was taken. mean cc01q01 by country. recode country (8 36 40 56 76 100 152 203 208 246 276 344 348 352 372 380 410 428 442 350 484 528 554 578 620 642 643 752 756 764 826 840=1) (else=0) into cccflag. /*Also deleted Liechtenstein since n=310. select if cccflag=1 and not miss(WLEread). /*303 cases had missing values on WLERead. * Get overall frequency distribution of supplementary domain items. freq st31q01 to st31q06. factor var st31q01 to st31q06 /print=corr. * Create independent item differentiation ERS ARS measures on polarity-balanced set of items. COUNT nsdQ31=st31q01 to st31q06(1). count ndQ31=st31q01 to st31q06(2). count naQ31=st31q01 to st31q06(3). COUNT nsaQ31=st31q01 to st31q06(4). count nmissQ31=st31q01 to st31q06(miss). count ersQ31=st31q01 to st31q06(1,4). compute sdQ31=sd(st31q01 to st31q06). do if nmissQ31 lt 6. compute nsdQ31=100*nsdQ31/(6-nmissQ31). compute ndQ31=100*ndQ31/(6-nmissQ31). compute naQ31=100*naQ31/(6-nmissQ31). compute nsaQ31=100*nsaQ31/(6-nmissQ31). ELSE. compute nsdQ31=999. compute ndQ31=999. compute naQ31=999. compute nsaQ31=999. end if. do if nmissQ31 lt 4. compute ersQ31=100*ersQ31/(6-nmissQ31). ELSE. compute ersQ31=999. end if. compute iidq31=sum(.0001*nsdQ31*nsdQ31, .0001*ndq31*ndq31,.0001*naq31*naq31,.0001*nsaQ31*nsaq31). compute iidq31=1-iidq31. compute nmissq31=100*nmissq31/6. Var lab nsdQ31 "Percentage of strongly disagrees to school integration-exclusion items" nsaQ31 "Percentage of strongly agrees to school integration-exclusion items" ersQ31 "Percentage of extreme responses to school integration-exclusion items" iidq31 "Index of item discrimination to school integration-exclusion items" nmissq31 "Percentage of non-response to school integration-exclusion items" sdQ31 "Standard deviation of school integration-exclusion items". missing values nsdQ31 ndQ31 naQ31 nsaQ31 ersQ31 (999). freq iidq31. split file by country. corr wleread iidq31 nmissQ31 nsdQ31 nsaQ31 ersQ31 sdQ31. split file off. save outfile="c:\sage\pisa2000\ccc.sav" /keep=stidstd country wleread cc01q01 cc01q03 cc01q05 cc01q07 cc01q09 cc01q10 cc01q12 cc01q13 cc01q15 cc01q17 cc01q19 cc01q20 cc01q21 cc01q23 cc01q25 cc01q27 cc01q28 nmissQ31 nsdQ31 nsaQ31 ersQ31 sdQ31 iidQ31. get file="c:\sage\pisa2000\ccc.sav". * Focal domain items. FREQUENCIES VARIABLES=CC01Q01 CC01Q03 CC01Q05 CC01Q07 CC01Q09 CC01Q10 CC01Q12 CC01Q13 CC01Q15 CC01Q17 CC01Q19 CC01Q20 CC01Q21 CC01Q23 CC01Q25 CC01Q27 CC01Q28. * Get country specific reading deciles. split file by country. freq wleread /ntiles=10 /format=notable. * Read the PISA2000 excel file in which we stored the deciles created in the above command. GET DATA /TYPE=XLSX /FILE='C:\Sage\PISA2000\Country.xlsx' /SHEET=name 'Deciles' /CELLRANGE=full /READNAMES=on /ASSUMEDSTRWIDTH=32767. FLIP VARIABLES=Albania Australia Austria Belgium Brazil Bulgaria Chile CzechRep Denmark Finland Germany HongKong Hungary Iceland Ireland Italy Korea Latvia Luxemburg Mexico Netherlands NewZealand Norway Portugal Romania Russia Sweden Switzerland Thailand UK USA /NEWNAMES=Name. SAVE OUTFILE='C:\Sage\PISA2000\countryReadDeciles.sav'. get file="c:\sage\pisa2000\ccc.sav". match files file=* /table='C:\Sage\PISA2000\countryReadDeciles.sav' /by country. execute. * Create country specific deciles as a single variable. do if wleread le dec1. compute readdec=1. else if wleread le dec2. compute readdec=2. else if wleread le dec3. compute readdec=3. else if wleread le dec4. compute readdec=4. else if wleread le dec5. compute readdec=5. else if wleread le dec6. compute readdec=6. else if wleread le dec7. compute readdec=7. else if wleread le dec8. compute readdec=8. else if wleread le dec9. compute readdec=9. else if wleread gt dec9. compute readdec=10. end if. var lab readdec "Country-specific reading achievement decile". freq readdec. save outfile="c:\sage\pisa2000\ccc.sav" /drop=dec1 to dec9. get file="c:\sage\pisa2000\ccc.sav". missing values cc01q01 TO cc01q28 (). Count n7=cc01q01 TO cc01q28(7). * It appears that codes of 7 indicate that the question wasn't asked, so delete all 10,654 cases where all 17 codes are 7. select if n7 lt 17. missing values cc01q01 TO cc01q28 (7 8 9). save outfile="c:\sage\pisa2000\ccc.sav" /drop=n7. get file="c:\sage\pisa2000\ccc.sav". * Calculate percent choosing never and always. TEMPORARY. recode cc01q01 TO cc01q28 (4=100)(1 thru 3=0). mean cc01q01 TO cc01q28 /cells=mean. * Calculate missings per country. COUNT nmissccc=cc01q01 TO cc01q28 (miss,sysmis). freq nmissccc. compute pmissccc=100*nmissccc/17. compute sdccc=sd(cc01q01 TO cc01q28). * Leave out the item where never exceeds always. Count pnever =cc01q01 TO cc01q09 cc01q12 to cc01q28 (1). Count psome =cc01q01 TO cc01q09 cc01q12 to cc01q28 (2). Count poften =cc01q01 TO cc01q09 cc01q12 to cc01q28 (3). Count palways =cc01q01 TO cc01q09 cc01q12 to cc01q28(4). if miss(cc01q10)nmissccc=nmissccc-1. freq nmissccc pnever to palways. compute meanccc=mean(cc01q01 TO cc01q28). do if nmissccc lt 16. compute pnever=100*pnever/(16-nmissccc). compute psome=100*psome/(16-nmissccc). compute poften=100*poften/(16-nmissccc). compute palways=100*palways/(16-nmissccc). compute normccc=palways-pnever. ELSE. compute pnever=999. compute psome=999. compute poften=999. compute palways=999. end if. missing values pnever to palways (999). compute iidccc=sum(pnever to palways). freq iidccc. compute iidccc=sum (.0001*pnever*pnever, .0001*psome*psome,.0001*poften*poften,.0001*palways*palways). compute iidccc=1-iidccc. freq iidccc. corr iidccc iidq31 sdQ31 sdccc. Var lab pnever "Proportion of 'never' responses to study habit items" palways "Proportion of 'always' responses to study habit items" normccc "Proportion of normative (always-never) responses" iidccc "Index of item discrimination for study habit items" meanccc "Mean of study habit items" sdccc "Standard deviation of study habit items" pmissccc "Proportion of non-response to study habit items". corr ersq31 pnever palways normccc nsdq31 nsaq31 iidccc iidq31 sdQ31 sdccc nmissQ31 pmissccc . means nmissQ31 to iidq31 iidccc pmissccc sdccc pnever palways normccc by readdec /cells=mean /stat=anova. means nmissQ31 to sdq31 pmissccc sdccc pnever palways normccc by country /cells=mean /stat=anova. * Do higher achievers endorse more of each study habit?. factor var=cc01q01 TO cc01q28 /select readdec(10) /print = correlation extraction. factor var=cc01q01 TO cc01q28 /print = correlation extraction. sort cases by country readdec. save outfile="c:\sage\pisa2000\ccc.sav". split file by country readdec. FACTOR /VARIABLES cc01q01 to cc01q28 /MISSING LISTWISE /ANALYSIS cc01q01 to cc01q28 /PRINT INITIAL EXTRACTION /CRITERIA FACTORS(4) ITERATE(25) /EXTRACTION PC /ROTATION NOROTATE /METHOD=CORRELATION. split file off. ************************************ CatPCA - to be done for all countries and all levels ********************************. TEMPORARY . select if (country =56 and readdec=10). CATPCA VARIABLES=CC01Q01 CC01Q03 CC01Q05 CC01Q07 CC01Q09 CC01Q10 CC01Q12 CC01Q13 CC01Q15 CC01Q17 CC01Q19 CC01Q20 CC01Q21 CC01Q23 CC01Q25 CC01Q27 CC01Q28 /ANALYSIS=CC01Q01(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q03(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q05(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q07(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q09(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q10(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q12(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q13(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q15(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q17(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q19(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q20(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q21(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q23(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q25(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q27(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) CC01Q28(WEIGHT=1,LEVEL=SPORD,DEGREE=2,INKNOT=2) /MISSING=CC01Q01(LISTWISE) CC01Q03(LISTWISE) CC01Q05(LISTWISE) CC01Q07(LISTWISE) CC01Q09(LISTWISE) CC01Q10(LISTWISE) CC01Q12(LISTWISE) CC01Q13(LISTWISE) CC01Q15(LISTWISE) CC01Q17(LISTWISE) CC01Q19(LISTWISE) CC01Q20(LISTWISE) CC01Q21(LISTWISE) CC01Q23(LISTWISE) CC01Q25(LISTWISE) CC01Q27(LISTWISE) CC01Q28(LISTWISE) /DIMENSION=3 /NORMALIZATION=VPRINCIPAL /MAXITER=100 /CRITITER=.00001 /PRINT=LOADING QUANT(CC01Q01 CC01Q03 CC01Q05 CC01Q07 CC01Q09 CC01Q10 CC01Q12 CC01Q13 CC01Q15 CC01Q17 CC01Q19 CC01Q20 CC01Q21 CC01Q23 CC01Q25 CC01Q27 CC01Q28) /PLOT=NONE. aggregate outfile="c:\sage\pisa2000\countrydecile.sav" /break = country readdec /aggsdccc=mean(sdccc) /aggsdq31=mean(sdq31) /aggiidccc=mean(iidccc) /aggiidq31=mean(iidq31) /aggpmissccc=mean(pmissccc) /aggnmissQ31=mean(nmissq31) /aggpalways=mean(palways) /aggpnever=mean(pnever) /aggnsaq31=mean(nsaq31) /aggnsdq31=mean(nsdq31) /aggersq31=mean(ersq31) /aggnormccc=mean(normccc) /aggmeanccc=mean(meanccc). aggregate outfile="c:\sage\pisa2000\countryTotals.sav" /break = country /Tsdccc=mean(sdccc) /Tsdq31=mean(sdq31) /Tiidccc=mean(iidccc) /Tiidq31=mean(iidq31) /Tpmissccc=mean(pmissccc) /TnmissQ31=mean(nmissq31) /Tpalways=mean(palways) /Tpnever=mean(pnever) /Tnsaq31=mean(nsaq31) /Tnsdq31=mean(nsdq31) /Tersq31=mean(ersq31) /Tnormccc=mean(normccc) /Tmeanccc=mean(meanccc). get file "c:\sage\pisa2000\countrydecile.sav". var lab aggsdccc "Standard deviation of study habit items" aggsdq31 "Standard deviation of school integration-exclusion items" aggiidccc "Index of item discrimination for study habit items" aggiidq31 "Index of item discrimination for school integration-exclusion items" aggpmissccc "Proportion non-response for study habit items" aggnmissQ31 "Proportion non-response for school integration-exclusion items" aggpalways "ARS for study habit items" aggpnever "Proportion 'never' to study habit items" aggnsaq31 "ARS for school integration-exclusion items" aggnsdq31 "Proportion 'strongly disagree' to school integration-exclusion items" aggersq31 "ERS for school integration-exclusion items" aggnormccc "Proportion normative response (always - never) to study habit items" aggmeanccc "Mean response to study habit items". match files file=* /table='C:\Sage\PISA2000\countryTotals.sav' /by country. * Make measures relative to country-specific means. do repeat I=AGGsdccc, AGGsdq31, AGGiidccc, AGGiidq31, AGGpmissccc, AGGnmissQ31, AGGpalways, AGGpnever, AGGnsaq31, AGGnsdq31, AGGersq31, AGGnormccc, AGGmeanccc /J=Tsdccc, Tsdq31, Tiidccc, Tiidq31, Tpmissccc, TnmissQ31, Tpalways, Tpnever, Tnsaq31, Tnsdq31, Tersq31, Tnormccc, Tmeanccc /K=Relsdccc, Relsdq31, Reliidccc, Reliidq31, Relpmissccc, RelnmissQ31, Relpalways, Relpnever, Relnsaq31, Relnsdq31, Relersq31, Relnormccc, Relmeanccc. compute k=i-j. end repeat. var lab Relsdccc "Standard deviation of study habit items relative to country mean" Relsdq31 "Standard deviation of school integration-exclusion items relative to country mean" Reliidccc "Index of item discrimination for study habit items relative to country mean" Reliidq31 "Index of item discrimination for school integration-exclusion items relative to country mean" Relpmissccc "Proportion non-response for study habit items relative to country mean" RelnmissQ31 "Proportion non-response for school integration-exclusion items relative to country mean" Relpalways "ARS for study habit items relative to country mean" Relpnever "Proportion 'never' to study habit items relative to country mean" Relnsaq31 "ARS for school integration-exclusion items relative to country mean" Relnsdq31 "Proportion 'strongly disagree' to school integration-exclusion items relative to country mean" Relersq31 "ERS for school integration-exclusion items relative to country mean" Relnormccc "Proportion normative response (always - never) to study habit items relative to country mean" Relmeanccc "Mean response to study habit items relative to country mean". mean relsdccc to relmeanccc by readdec /cells=mean /stat=anova. mean aggsdccc to aggmeanccc by readdec /cells=mean /stat=anova. save outfile= "c:\sage\pisa2000\countrydecile.sav".