T racing the Contours of Turkish Ideological Space, 2001-2004

1. Tracing the Contours of Turkish Ideological Space, 2001-2004 Melvin J. Hinich Ali Çarkoğlu Austin, Texas April, 2006

2. Vote Share on a Left - Right Scale 1950-2002

5. Province clusters in November 2002 • AKP is by far the largest party, CHP and DYP are second and third far behind • AKP is by far the largest party, CHP is the second with considerable votes, DYP and MHP are following • AKP is by far the largest party, ANAP is the second far behind • AKP is by far the largest party, there are considerable votes for the independent candidate • KIRKLARELİ • SİNOP • BARTIN • İSTANBUL • 34 • RİZE • ARDAHAN • EDİRNE • 15 • 248 • KASTAMONU • SAMSUN • KOCAELİ • ZONGULDAK • ARTVİN • TEKİRDAĞ • 14 • KARABÜK • TRABZON • DÜZCE • ORDU • 116 • GİRESUN • KARS • SAKARYA • 65 • YALOVA • 27 • AMASYA • BOLU • ÇANKIRI • ÇORUM • GÜMÜŞHANE • 41 • TOKAT • 30 • IĞDIR • BAYBURT • BURSA • 45 • 23 • ÇANAKKALE • ERZURUM • BİLECİK • 36 • 165 • ANKARA • 24 • AĞRI • KIRIKKALE • ERZİNCAN • SİVAS • ESKİŞEHİR • BALIKESİR • 203 • YOZGAT • 9 • 47 • KÜTAHYA • TUNCELİ • KIRŞEHİR • MUŞ • BİNGÖL • 54 • VAN • MANİSA • ELAZIĞ • AFYON • NEVŞEHİR • BİTLİS • KAYSERİ • 54 • 27 • UŞAK • MALATYA • 23 • 107 • AKSARAY • DİYARBAKIR • İZMİR • BATMAN • KONYA • 106 • SİİRT • 69 • 118 • AYDIN • ADIYAMAN • HAKKARİ • ISPARTA • NİĞDE • ŞIRNAK • 50 • DENİZLİ • KAHRAMANMARAŞ • 50 • ADANA • MARDİN • ŞANLIURFA • BURDUR • MUĞLA • OSMANİYE • ANTALYA • 45 • 44 • GAZİANTEP • KARAMAN • 63 • 26 • İÇEL • KİLİS • 22 • HATAY • DEHAP is by far the largest party, AKP is the second far behind • 28 1st Province Cluster 4th Province Cluster • CHP is the first party, AKP is the second with considerable votes and DYP is the tird party 2nd Province Cluster 5th Province Cluster 3rd Province Cluster 6th Province Cluster

6. Euclidean Distance Model Suppose that there are N observers and M targets. Each observer at position reports the squared Euclidean distance to the targets at locations

7. Squared Distances with Errors For each error

8. Removing the Nonlinear Terms The nonlinearity is removed by subtracting the distances to one target, target m=0 from the distances to the other targets. Then compute the sample covariance matrix of the differences

9. Assume that the errors are independently & identically distributed and that they are independent of the observer positions

10. Covariance Matrix of Distance Differences • Assume that the observer positions are uncorrelated random variables whose variances are Then the covariance matrix of the distance differences is

11. is a Mx2 matrix of target positions

12. 2001 Survey A nation-wide representative survey of urban population conducted during the chaotic weeks of the second economic crisis of February 2001 1201 face-to-face interviews were conducted in 12 of the 81 provinces of Turkey The survey was run during 2/20 – 3/16 using a random sampling method that represents the nationwide voting age urban population based on the urban population figures of 1997 census data.

16. KEY DATA ABOUT THE SURVEY • The target of the sample was the nation-wide urban and rural settlers who are 18 years or older. • The sample consisted of a total of 2028 face-to-face interviews conducted in 54 districts, 291 neighborhoods and 95 villages of a total 33 provinces. • Under the restrictive assumption of simple random sampling this sample has a confidence interval of 95% with an error margin of +/- 2,2%. • Provinces chosen according to probability proportionate to size (PPS) principle on the basis of their registered voter population in 1999 general election. • The questionnaire was administered between the 10th and 25th of October 2002.

17. FACTS ABOUT THE FIELD RESEARCH • Training for questionnaire implementation was given by Ali Çarkoğlu and Ersin Kalaycıoğlu on October 4. The questionaire was tested and necessay corrections made after a pilot study on October 5-6. • During the project 9 experts, 35 supervisers and 186 interviewers were assigned. • 45% of the interviews were completed in the first trial. The rest is completed by either trying for a second time or selecting new streets and households. · Interviews lasted 35 minutes on average. • 45% of the interviews were randomly controlled. Controls are made either by phoning or going to the households one more time. 332 interviews were cancelled and conducted with replacements.

18. Basic Independent Variables

19. Vote Intentions for the November 3rd Election • % 29,4 • AKP • % 14,4 • CHP • % 9,9 • GP • % 5,0 • DYP • % 4,1 • MHP • % 3,3 • DEHAP • % 2,4 • ANAP • % 1,5 • SP • % 1,0 • YTP • % 0,7 • DSP • % 0,6 • BBP • % 1,3 • Other • Will not vote • % 1,9 • Will not vote for the • existing parties • % 7,1 • % 8,7 • Undecided • % 8,7 • DK/NA • %0 • %10 • %20 • %30

20. Xenophobia & Political Efficacy • Xenophobia • Foreigners who settle in our country harm our culture. • Foreigners who settle in our country make our chances of finding a job • more difficult • Some should either love Turkey or leave it. • I would not want a foreigner to be my neighbor • Political efficacy • Regular citizens like me have no power for changing political decisions in Turkey for their advantage. • Turkey is being ruled by a small and powerful group. • Whatever I do I don't think I can reach a better position in society

24. Valence Question - Revitalizing the Economy

31. March-2004 survey of nation-wide representative urban population (N=1,232)

33. ● 1st D. 2nd D. Alevis -1.5 -3.8 Non-Alevis +0.3 -1.7