Jaime Díez MedranoDirector of the WVS Data Archive
Researchers have gained free access during the last years to a valuable number of data sources. Among them is the complete collection of files from the World Values Surveys from 1981 to 2009, available separately or aggregated in a single file at www.worldvaluessurvey.org.
An agreement with EVS five years ago also made it possible to build an aggregate of the 4 initial waves of the WVS and EVS surveys, covering the 1981-2004 period. Unfortunately this agreement has not been renewed yet for the 1981-2009 period and so it is not possible to download directly the file with all data from both studies covering the period 1981-2009.
The five-wave WVS/EVS aggregate can however be constructed using existing files by any researcher using the procedure we describe in this paper.
COMPATIBILITY OF 4-WAVE AND 5-WAVE AGGREGATES
The 4-wave aggregate of the values surveys includes both WVS and EVS surveys ranging from 1981 to 2004. Both studies had an initial common start but later they took a separate path. In an effort to build a common and official aggregate of all the studies, a joint project was carried in 2005 between Tilburg University (Netherlands) and ASEP/JDS (Spain).
The main result from this project was the definition of a global dictionary, which is a listing of all variables asked in any of the surveys included in the aggregate file (EVS 1981, WVS 1981, EVS 1990, WVS 1990, WVS 1995, EVS 1999 and WVS 2000). The global dictionary also includes information about the correspondence of each individual wave variable wuth other wave variables and the recodings that are necessary to make them compatible. This was covered in Human Beliefs and Values: A Cross-Cultural Sourcebook, Loek Halman, Ronald Inglehart, Jaime Díez-Medrano, Ruud Luijkx, Editorial Siglo XXI, 2004.
When processing the WVS 2005, the variables have also been added to the global dictionary and, at the same, this dictionary was extended to include the new variables added in 2005. A few variables have been reorganized and some minor mistakes of the previous aggregate corrected. The new WVS 5-wave aggregate available on the web is built from the WVS 1981, WVS 1990, WVS 1995, WVS 2000 and WVS 2005 files using the same technique as the 4-wave, and using the same global dictionary. It is, as a consequence, compatible with the 4-wave aggregate.
The lack of an agreement between the interested parts prevents publishing a 5-wave WVS-EVS aggregate nowadays, but it is possible to extract the EVS files from the 4-wave aggregate and insert them in the existing 5-wave WVS aggregate. The procedure is described below.
PROCEDURE FOR BUILDING A 5-WAVE WVS-EVS AGGREGATE
The procedure is described in terms of SPSS syntax, available for download in the following link: 5waveWVSEVSprocedure.sps
The methodology is easy enough to be replicated in any other statistical environment.
There are basically four steps:
Before you start, the following files need to be downloaded from the WVS site: www.worldvaluessurvey.org:
b- Extracting EVS data from the 1981-2004 aggregate file
Variable S001 tells if the data comes from EVS or WVS. S001 equals 1 for all EVS data. Also note that Sweden made the 1999 survey for both studies, and so it needs not be extracted from the file since it is already present in the WVS 5-wave aggregate.
The syntax in SPSS for doing this task is just this:
* Open the 4-wave aggregate in the workspace. NEW FILE.GET FILE='c:\your_data_folder\XWVSEVS_1981_2000_v20060423.sav'.* Filter for EVS, excluding Sweden 1999 which is already in the WVS 5-wave aggregate.SELECT IF (s001=1 AND s021<>75204111999).EXECUTE.
c- Adapting EVS data to the current global dictionary:
Some EVS variables weren’t properly recoded in the 4-wave aggregate. More specifically, some Yes/No variables with Yes coded 1 and No coded 2 weren’t recoded to 1 Yes and 0 No in the aggregate and can be corrected using the following syntax:
* Adapt S009 format to 3 characters instead of 2.alter type s009 (A3).* Correct a few recoding mistakes in the official version.recode a010 to a019,C059,D020A (2=0).
Also, a number of variables were renamed. The battery of variables about neighbours, confidence in organizations and trust were renamed to a common variable name, followed by a number for each of the groups mentioned.
The following table summarizes the changes:
Neighbours: People with a criminal record
Neighbours: People of a different race
Neighbours: Heavy drinkers
Neighbours: Emotionally unstable people
Neighbours: Immigrants/foreign workers
Neighbours: People who have AIDS
Neighbours: Drug addicts
Neighbours: People of a different religion
Neighbours: People of the same religion
Neighbours: Militant minority
Neighbours: People not from country of origin
Neighbours: Political Extremists
Neighbours: Indians or Lebanese
Neighbours: Chinese or Philippine Chinese
Neighbours: Witchdoctors and related labels
Neighbours: Left wing extremists
Neighbours: Right wing extremists
Neighbours: People with large families
Neighbours: North-American persons
Neighbours: Members of new religious movements
Neighbours: Jews, Arabs, Asians, gypsies, etc
Neighbours: Black people
Neighbours: White people
Neighbours: Coloured people
Neighbours: Kurds, Esids
Neighbours: Unmarried mothers
Neighbours: Members of minority religious sects or cults
Neighbours: Members of ETA (terrorists)
Why people work: work is like a business transaction
Why people work: I do the best I can regardless of pay
Why people work: I wouldn’t work if I didn’t have to
Why people work: I wouldn´t work if work interfered my life
Why people work: work most important in my life
Why people work: I never had a paid job
Why people work: don’t know
Confidence: Armed Forces
Confidence: Education System
Confidence: The Press
Confidence: Labour Unions
Confidence: The Police
Confidence: The Civil Services
Confidence: Social Security System
Confidence: The Government
Confidence: The Political Parties
Confidence: Major Companies
Confidence: The Environmental Protection Movement
Confidence: The Women´s Movement
Confidence: Health Care System
Confidence: Justice System
Confidence: The European Union
Confidence: The United Nations
Confidence: The Arab League
Confidence: The Association of South East Asian Nations -ASEAN
Confidence: The Organization for African Unity-OAU
Confidence: The NAFTA
Confidence: The Andean pact
Confidence: The Mercosur
Confidence: The SAARC
Confidence: The ECO
Confidence: The APEC
Confidence: The Free Commerce Treaty (Tratado de libre comercio)
Confidence: The United American States Organization (Organización de Estados Unidos Americanos - OEA)
Confidence: The “Movimiento en pro de Vieques”(Puerto Rico)
Confidence: Local/Regional Government
Confidence: East African Cooperation (EAC)
Confidence: Other television
Confidence: The Americans
Confidence: The Presidency
Confidence: The Civil Society Groups
Trust: Other people in country
Trust: French Canadians
Trust: Korean residents in Japan
Trust: Chinese residents in Japan
Trust: Central Americans
Trust: East Germans
Trust: West Germans
Trust: Your friends
Trust: Your neighbourhood
Trust: White South Africans
Trust: Black South Africans
Trust: Coloured South Africans
Trust: Asian South Africans
Trust: Soviet Union people
Trust: Latin Americans
Trust: Mapuche Indians
Trust: Chinese Zhuan Nationality
Trust: Chinese Hui Nationality
Trust: Indian Hindus
Trust: Indian Non-Hindus
The SPSS syntax to rename the above variables is this:
* Rename the few variable names that have changed in the current aggregate dictionnary.RENAME VARIABLES (A124=A124_01) (A125=A124_02) (A126=A124_03) (A127=A124_04) (A128=A124_05) (A129=A124_06) (A130=A124_07) (A131=A124_08) (A132=A124_09) (A133=A124_10) (A134=A124_11) (A135=A124_12)(A136=A124_13) (A137=A124_14) (A138=A124_15) (A139=A124_16) (A140=A124_17) (A141=A124_18) (A142=A124_19) (A143=A124_20) (A144=A124_21) (A145=A124_22) (A146=A124_23) (A147=A124_24)(A148=A124_25) (A149=A124_26) (A150=A124_27) (A151=A124_28) (A152=A124_29) (A153=A124_30) (A154=A124_31) (A155=A124_32) (A156=A124_33) (A157=A124_34) (A158=A124_35) (A159=A124_36)(A160=A124_37) (A161=A124_38) (A162=A124_39) (A163=A124_40) (A164=A124_41) (A177=A124_44) (A178=A124_45) (A180=A124_46) (A182=A124_47) (A183=A124_48) (A184=A124_49) (A185=A124_50) (A186=A124_51) (A187=A124_52) (A188=A124_53) (C042=C042B1) (C043=C042B2) (C044=C042B3) (C045=C042B4) (C046=C042B5) (C047=C042B6) (C048=C042B7) (E069=E069_01) (E070=E069_02) (E071=E069_03) (E072=E069_04) (E073=E069_05) (E074=E069_06) (E075=E069_07) (E076=E069_08)(E077=E069_09) (E078=E069_10) (E079=E069_11) (E080=E069_12) (E081=E069_13) (E082=E069_14) (E083=E069_15) (E084=E069_16) (E085=E069_17) (E086=E069_18) (E087=E069_19) (E088=E069_20)(E089=E069_21) (E090=E069_22) (E091=E069_23) (E092=E069_24) (E093=E069_25) (E094=E069_26) (E095=E069_27) (E096=E069_28) (E097=E069_29) (E098=E069_30) (E099=E069_31) (E100=E069_32)(E101=E069_33) (E102=E069_34) (E103=E069_35) (E199=E069_36) (E200=E069_37) (E201=E069_38) (E202=E069_39) (G007=G007_01) (G008A=G007_02) (G008AA=G007_03) (G008AB=G007_04) (G008AC=G007_05) (G008AD=G007_06) (G008AE=G007_07) (G008AF=G007_08) (G008AG=G007_09) (G008AH=G007_10) (G008AI=G007_11) (G008AJ=G007_12) (G008AK=G007_13) (G008AL=G007_14) (G008AM=G007_15) (G008AN=G007_16) (G008AO=G007_17) (G008AP=G007_18) (G008AQ=G007_19) (G008AR=G007_20) (G008AS=G007_21) (G008AT=G007_22) (G008AU=G007_23) (G008AV=G007_24) (G008AW=G007_25) (G008AX=G007_26) (G008AY=G007_27) (G008AZ=G007_28) (G008B=G007_29) (G008BA=G007_30) (G008BB=G007_31) (G008BC=G007_32) (G008C=G007_37) (G008D=G007_38) (G008E=G007_39) (G008F=G007_40) (G008G=G007_41) (G008H=G007_42) (G008I=G007_43) (G008J=G007_44) (G008K=G007_45) (G008L=G007_46) (G008M=G007_47) (G008N=G007_48) (G008O=G007_49) (G008P=G007_50) (G008Q=G007_51) (G008R=G007_52) (G008S=G007_53) (G008T=G007_54) (G008U=G007_55) (G008V=G007_56) (G008W=G007_57) (G008X=G007_58) (G008Y=G007_59) (G008Z=G007_60).execute.
d- Saving the EVS part for later use:
You can save the result of the previous step to a file in the same disk folder using the following syntax (or manually if you prefer). Let’s save it under the name evs_1981_1999.sav.
* The file is ready to be merged. Save the EVS 4-wave under evs_1981_1999.sav name excluding some variables that will not be included in the aggregate. SAVE OUTFILE='c:\your_data_folder\evs_1981_1999.sav' /DROP a008a,a059a,a060a,a061a,c036a,c037a,c038a,c039a,c040a,c041a, c042a,c060a,d020a,d044a,d056a,d057a,d058a,d061a,d062a,d063a,d064a, d065a,e003a,e004a,e146a,e147a,e148a,e149a,e168a,e169a,e170a,e171a, e172a,e173a,e174a,e175a,e176a,e177a,f145a,f146a,f147a,f148a,f149a,f150a, f151a,f152a, f153a,f154a,f155a,f160a,f161a,f163a.
In the above command, we have omitted some variables that are not common to any WVS study. If you want to keep them, simply remove the /DROP subcommand from the SPSS command.
e- Merging together 5-wave WVS and 4-wave EVS files
The last step is to merge the two files with WVS and EVS only aggregates. This can be done in SPSS with the following command:
* Now merge the WVS 5-wave and the EVS 4-wave.NEW FILE.ADD FILES/FILE='c:\your_data_folder\wvs1981_2008_v20090914.sav'/FILE='c:\your_data_folder\evs_1981_1999.sav' .
We now have an aggregate of the WVS+EVS, covering 1981-2009 and with 355,298 individual cases from 98 world societies!
f- Final adjustments of the aggregate
As a final step, we need to rebuild the Y003 Autonomy index, which is reversed in the 4-wave aggregate, and rebuild the TradRat/SurvSelf factors for the whole file:
* Compute again the Autonomy Index (Auton4) to make sure it is correct.COMPUTE Y003=-4. if (A040>=0 and A042>=0 and A029>=0 and A039>=0) y003=-(a040 + a042)+(a029 +A039).execute.* Compute again the TradRat and SurvSelf scores.FACTOR /VARIABLES a008 a165 e018 e025 f063 f118 f120 g006 y002 y003 /MISSING LISTWISE /ANALYSIS a008 a165 e018 e025 f063 f118 f120 g006 y002 y003/PRINT INITIAL EXTRACTION ROTATION/FORMAT SORT BLANK(.3)/CRITERIA FACTORS(2) ITERATE(25)/EXTRACTION PC/CRITERIA ITERATE(25)/ROTATION VARIMAX/SAVE REG(ALL)/METHOD=CORRELATION.EXECUTE .COMPUTE TRADRAT5=fac1_1.COMPUTE SURVSELF=-fac2_1.COMPUTE S007=$CASENUM.EXECUTE.
g- Save your results
The resulting file is ready. Let’s save it for future use in our research:
* Save the resulting 5 wave EVS/WVS aggregate under the name wvsevs_1981_2009, excluding the auxiliar variables that aren't needed.SAVE OUTFILE='c:\your_data_folder\wvsevs_1981_2009.sav' /DROP fac1_1,fac2_1.
For those willing to create smaller files for analyzing subsets of such a big file, you may consider the following tricks:
a) Variable S001 codes the study (1= EVS 2=WVS)b) Variable S002 codes the wave (1=1981, 2=1990, 3=1995, 4=1999/2000 and 5=2005)c) Variable S003 codes the countryd) Variable S004 codes the survey number in the same wave, for the few cases were more than a survey was carried.e) Variable S021 is a compound of the above variables (study-wave-country-sample)You can use any of these variables to filter the file and build a smaller one.