34
Appendix: Individual-Specific Results for Party Selection Model (cont.)
Ind.
Uncorrected Model
Corrected Model
Party Variable
Coef. R.S.E. Party Variable Coef. R.S.E.
Democratic
SMD (1,0)
1.795***
0.682 0.420 0.202 0.655 0.001
Initiatives
Prev. Fail.
1.305*** 0.478 0.307 2.891*** 0.522 0.015
Mills
-11.81***
1.255
-0.011
Constant
-0.767 0.546 -0.124 19.422 2.276 0.999
European
SMD (1,0)
2.351***
0.795 0.526 0.949 0.733 0.001
Choice
Prev. Fail.
1.227** 0.548 0.288 2.669*** 0.603 0.012
Mills
-10.50***
1.070
-0.009
Constant
-1.346 0.690 -0.185 16.784 2.042 0.999
People’s
SMD (1,0)
2.571***
0.906 0.560 1.024 0.879 0.002
Will
Prev. Fail.
-0.044 0.990 -0.009 1.474 0.988 0.003
Mills
-11.47***
1.121
-0.010
Constant
-1.682 0.797 -0.209 17.947 2.128 0.999
People’s 2
SMD (1,0)
0.568 0.668 0.125 -0.897 0.648 -0.001
Democratic
Prev. Fail.
-17.18*** 0.335 -0.272 -21.73*** 0.456 -0.001
Party ‘02
Mills
-10.55***
1.127
-0.009
Constant
-0.138 0.467 -0.026 18.110 2.024 0.999
People’s
SMD (1,0)
1.466** 0.668 0.345 -0.124 0.669 0.001
Party
Prev. Fail.
1.688*** 0.445 0.397 3.283*** 0.590 0.022
Mills
-11.74***
1.500
-0.011
Constant
-0.723 0.492 -0.119 19.349 2.631 0.999
Party of
SMD (1,0)
0.972* 0.543 0.224
-2.445***
0.628 -0.001
Industrialists
Prev. Fail.
-0.500 0.704 -0.087
3.021***
1.033 0.017
and
Mills
-27.93***
4.051
-0.025
Entrepreneurs Constant
0.587 0.376 0.129 45.042 6.128 1.000
Regions of
SMD (1,0)
2.049*** 0.588 0.472 -2.326*** 0.597 -0.001
Ukraine ‘02
Prev. Fail.
1.098*** 0.430 0.256 6.620*** 1.257 0.400
Mills
-41.98***
5.636
-0.038
Constant
-0.208 0.449 -0.039 64.523 8.159 1.000
***p ≤ .01, **p ≤ .05, *p ≤ .1
Notes: The equations were estimated with a conditional fixed-effects logit model. The corrected model
includes the inverse Mills ratio from the first-stage population-average probit model. The dependent
variable in each model indicated a deputy’s parliamentary party choice at the end of the legislative session.
The marginal effects represent the probability of a positive outcome for each continuous variable at its
mean level and for each dummy variable the change from going from 0 to 1. Robust standard errors
calculated clustering on individual deputies. Results from natural cubic splines not reported. The reference
group is the SDPU
(o)
.