Impact of internal migration on political participation in Turkey

Table 3 Estimated robust regressions (Method: LTS)

Independent variables	Regression 1	Regression 2	Regression 3
CONSTANT	68.424 (0.00)	51.516 (0.00)	68.717 (0.00)
PRIMARY	0.227 (0.00)	0.128 (0.00)	0.210 (0.00)
HIGHER	−0.355 (0.02)	−0.222 (0.07)	−0.342 (0.03)
UNDER30	0.456 (0.03)	0.046 (0.58)	0.487 (0.00)
OVER60	0.325 (0.00)	0.061 (0.44)	0.325 (0.01)
URBAN	0.052 (0.01)	0.038 (0.02)	0.049 (0.02)
MP	−0.358 (0.00)	−0.204 (0.04)	−0.359 (0.00)
PARTIES	−0.766 (0.16)	−0.619 (0.15)	−0.794 (0.18)
NOCOMPETITION	−0.682 (0.28)	−0.278 (0.58)	−0.744 (0.27)
EMIGRANT	−0.113 (0.00)	−0.055 (0.00)	−0.109 (0.00)
IMMIGRANT	−0.074 (0.02)	−0.034 (0.19)	−0.072 (0.03)
(MP×IMMIGRANT)/100	0.366 (0.05)	0.222 (0.13)	0.364 (0.06)
TURNOUT 2007		0.348 (0.00)
WEST			0.485 (0.63)
NORTH			0.321 (0.73)
CENTRAL			0.748 (0.40)
R-SQUARE	0.79	0.90	0.81

Notes: The dependent variable is the turnout rate in the 2011 parliamentary election. Definitions of independent variables are given in section 2. Estimates are obtained using the ROBUSTREG procedure of the SAS statistical package with METHOD = LTS (least trimmed squares) and FWLS (final weighted least squares) options. The following observations are picked up by the algorithm as outliers: Ağrı, Gümüşhane, Tunceli, Yozgat, Aksaray, and Iğdır in the first regression; Ağrı, Amasya, Gümüşhane, Yozgat, Aksaray, and Iğdır in the second regression; Ağrı, Gümüşhane, Tunceli, Yozgat, Aksaray, and Iğdır in the third regression. The numbers in parantheses next to the parameter estimates are the probabilities relevant to the chi-square test of whether the associated coefficient is equal to zero.