LOGIT SPECIFICATION

Here I repeat the same analysis as in the paper but now the dependent variable is logit transformed to make sure that all model predictions are bouded by 0 and 1. If e is the turnout (in percentages) in the EP elections its logit is ln((e/100)/(1-(e/100))). Similar transformation is applied to the turnout in national parliamentary elections.

NOTE! in the following models (the "b" variants) the dummy variables may be different than in the paper. The reason is that I used the modelling strategy explained in the paper.

All in all, these results are very similar to the ones in the paper (with the untransformed variables).

Summary:

Compulsory voting         ++++++
Simultanous elections     ++++++
Weekend voting            ++++++
Country net payer         ---- 
Country net receiver      ++ 
Share of seats            + 
Strict party lists        --
Multiple constituencies   ++ 
New member country        + 

MODEL 1A

--> REGRESS;Lhs=lnep;Rhs=ONE, compulso, arrival, concurr, seatshar, restday, ...
    ,constitu, netpayer, netrecei$

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNEP     Mean=   .5234934857    , S.D.=   .9388390734     |
| Model size: Observations =      64, Parameters =  10, Deg.Fr.=     54 |
| Residuals:  Sum of squares= 9.766228328    , Std.Dev.=         .42527 |
| Fit:        R-squared=  .824125, Adjusted R-squared =          .79481 |
| Model test: F[  9,     54] =   28.12,    Prob value =          .00000 |
| Diagnostic: Log-L =    -30.6536, Restricted(b=0) Log-L =     -86.2690 |
|             LogAmemiyaPrCrt.=   -1.565, Akaike Info. Crt.=      1.270 |
| Autocorrel: Durbin-Watson Statistic =   1.09979,   Rho =       .45010 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -.3106464100      .14082711       -2.206   .0317
 COMPULSO  1.224447945      .17027337        7.191   .0000  .28125000
 ARRIVAL   .8963516067E-03  .20935305         .004   .9966  .93750000E-01
 CONCURR   .4284986239      .14446750        2.966   .0045  .26562500
 SEATSHAR  .1969184716E-01  .16860132E-01    1.168   .2480  8.2995483
 RESTDAY   .6399315287      .16027420        3.993   .0002  .68750000
 LIST     -.5642262840      .20577312       -2.742   .0083  .43750000
 CONSTITU  .9454306491E-01  .17400070         .543   .5891  .31250000
 NETPAYER -.1592396331      .15709063       -1.014   .3153  .18750000
 NETRECEI  .6981437013E-01  .18670003         .374   .7099  .28125000

MODEL 1B


--> REGRESS;Lhs=lnep;Rhs=ONE, compulso, arrival, concurr, seatshar, restday, ...
    ,constitu, netpayer, netrecei,  y1999,uk, italcomp, greece, ireland, finland
    , sweden, france$

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNEP     Mean=   .5234934857    , S.D.=   .9388390734     |
| Model size: Observations =      64, Parameters =  18, Deg.Fr.=     46 |
| Residuals:  Sum of squares= 2.342300872    , Std.Dev.=         .22565 |
| Fit:        R-squared=  .957819, Adjusted R-squared =          .94223 |
| Model test: F[ 17,     46] =   61.44,    Prob value =          .00000 |
| Diagnostic: Log-L =     15.0359, Restricted(b=0) Log-L =     -86.2690 |
|             LogAmemiyaPrCrt.=   -2.730, Akaike Info. Crt.=       .093 |
| Autocorrel: Durbin-Watson Statistic =   2.01335,   Rho =      -.00668 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -.1551573026      .10733574       -1.446   .1551
 COMPULSO  1.520633228      .21527747        7.064   .0000  .28125000
 ARRIVAL   .2316629623      .13764282        1.683   .0991  .93750000E-01
 CONCURR   .4082080127      .95953081E-01    4.254   .0001  .26562500
 SEATSHAR  .3767162851E-01  .16960306E-01    2.221   .0313  8.2995483
 RESTDAY   .4120888358      .16039394        2.569   .0135  .68750000
 LIST     -.2384888878      .25172900        -.947   .3484  .43750000
 CONSTITU  .3280713194      .18503370        1.773   .0828  .31250000
 NETPAYER -.3712288611      .13235064       -2.805   .0074  .18750000
 NETRECEI -.1999182618      .22822028        -.876   .3856  .28125000
 Y1999    -.2912158473      .80121462E-01   -3.635   .0007  .23437500
 UK       -1.252533502      .29230159       -4.285   .0001  .78125000E-01
 ITALCOMP -1.323495737      .25905617       -5.109   .0000  .46875000E-01
 GREECE   -.5395499348      .22106071       -2.441   .0186  .78125000E-01
 IRELAND  -.7140490344E-01  .28011057        -.255   .7999  .78125000E-01
 FINLAND  -.7460820942      .21466491       -3.476   .0011  .31250000E-01
 SWEDEN   -.3963434129      .23953044       -1.655   .1048  .31250000E-01
 FRANCE   -.4597390513      .19335269       -2.378   .0216  .78125000E-01

MODEL 2A

--> REGRESS;Lhs=LNDIFF;Rhs=ONE, compulso, arrival, concurr, seatshar, restday...
    ,constitu, netpayer, netrecei$

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNDIFF   Mean=  -.9717786972    , S.D.=   .6841621581     |
| Model size: Observations =      64, Parameters =  10, Deg.Fr.=     54 |
| Residuals:  Sum of squares= 5.989583614    , Std.Dev.=         .33304 |
| Fit:        R-squared=  .796887, Adjusted R-squared =          .76303 |
| Model test: F[  9,     54] =   23.54,    Prob value =          .00000 |
| Diagnostic: Log-L =    -15.0085, Restricted(b=0) Log-L =     -66.0163 |
|             LogAmemiyaPrCrt.=   -2.054, Akaike Info. Crt.=       .782 |
| Autocorrel: Durbin-Watson Statistic =   1.43528,   Rho =       .28236 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -1.715878829      .11028619      -15.558   .0000
 COMPULSO  .4044068919      .13334650        3.033   .0037  .28125000
 ARRIVAL  -.9215830738E-01  .16395104        -.562   .5764  .93750000E-01
 CONCURR   .6642769614      .11313710        5.871   .0000  .26562500
 SEATSHAR  .4344681341E-04  .13203706E-01     .003   .9974  8.2995483
 RESTDAY   .6341938320      .12551583        5.053   .0000  .68750000
 LIST     -.1008869411      .16114748        -.626   .5339  .43750000
 CONSTITU  .1429460369      .13626549        1.049   .2988  .31250000
 NETPAYER -.3374411246      .12302267       -2.743   .0082  .18750000
 NETRECEI  .3161625952      .14621074        2.162   .0350  .28125000

MODEL 2B

--> REGRESS;Lhs=LNDIFF;Rhs=ONE, compulso, arrival, concurr, seatshar, restday...
    ,constitu, netpayer, netrecei, y1999, sweden, italcomp, france$

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNDIFF   Mean=  -.9717786972    , S.D.=   .6841621581     |
| Model size: Observations =      64, Parameters =  14, Deg.Fr.=     50 |
| Residuals:  Sum of squares= 2.685506595    , Std.Dev.=         .23175 |
| Fit:        R-squared=  .908932, Adjusted R-squared =          .88525 |
| Model test: F[ 13,     50] =   38.39,    Prob value =          .00000 |
| Diagnostic: Log-L =     10.6604, Restricted(b=0) Log-L =     -66.0163 |
|             LogAmemiyaPrCrt.=   -2.726, Akaike Info. Crt.=       .104 |
| Autocorrel: Durbin-Watson Statistic =   1.96305,   Rho =       .01848 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -1.780610638      .93332121E-01  -19.078   .0000
 COMPULSO  .4900456025      .12201058        4.016   .0002  .28125000
 ARRIVAL   .1517122057      .13154412        1.153   .2543  .93750000E-01
 CONCURR   .5251938122      .83718367E-01    6.273   .0000  .26562500
 SEATSHAR  .3609608081E-02  .14805397E-01     .244   .8084  8.2995483
 RESTDAY   .6549210674      .96025689E-01    6.820   .0000  .68750000
 LIST     -.3944938317      .15518835       -2.542   .0142  .43750000
 CONSTITU  .2794475956      .10896176        2.565   .0134  .31250000
 NETPAYER  .1588513974E-02  .10714972         .015   .9882  .18750000
 NETRECEI  .5460715874      .12226257        4.466   .0000  .28125000
 Y1999    -.1293367237      .78219676E-01   -1.654   .1045  .23437500
 SWEDEN   -.9875988284      .21067254       -4.688   .0000  .31250000E-01
 ITALCOMP -.5104663134      .23881740       -2.137   .0375  .46875000E-01
 FRANCE    .6854362028      .14206344        4.825   .0000  .78125000E-01

MODEL 3A

--> REGRESS;Lhs=lnep;Rhs=ONE, compulso, arrival, concurr, seatshar, restday, ...
    ,constitu, netpayer, netrecei, lnnp$

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNEP     Mean=   .5234934857    , S.D.=   .9388390734     |
| Model size: Observations =      64, Parameters =  11, Deg.Fr.=     53 |
| Residuals:  Sum of squares= 5.497892490    , Std.Dev.=         .32208 |
| Fit:        R-squared=  .900991, Adjusted R-squared =          .88231 |
| Model test: F[ 10,     53] =   48.23,    Prob value =          .00000 |
| Diagnostic: Log-L =    -12.2675, Restricted(b=0) Log-L =     -86.2690 |
|             LogAmemiyaPrCrt.=   -2.107, Akaike Info. Crt.=       .727 |
| Autocorrel: Durbin-Watson Statistic =   1.40146,   Rho =       .29927 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -1.359793995      .19525857       -6.964   .0000
 COMPULSO  .6122046760      .16043497        3.816   .0004  .28125000
 ARRIVAL  -.6857832702E-01  .15892193        -.432   .6678  .93750000E-01
 CONCURR   .6045309086      .11280075        5.359   .0000  .26562500
 SEATSHAR  .5022336633E-02  .12972106E-01     .387   .7002  8.2995483
 RESTDAY   .6356477600      .12138467        5.237   .0000  .68750000
 LIST     -.2182967810      .16490835       -1.324   .1913  .43750000
 CONSTITU  .1306807603      .13189889         .991   .3263  .31250000
 NETPAYER -.2922850012      .12076619       -2.420   .0190  .18750000
 NETRECEI  .2537381409      .14427417        1.759   .0844  .28125000
 LNNP      .7466007546      .11639098        6.415   .0000  1.4952722

MODEL 3B

--> REGRESS;Lhs=lnep;Rhs=ONE, compulso, arrival, concurr, seatshar, restday, ...
    ,constitu, netpayer, netrecei, lnnp, y1999, italcomp,uk, luxembou, sweden
    $

+-----------------------------------------------------------------------+
| Ordinary    least squares regression    Weighting variable = none     |
| Dep. var. = LNEP     Mean=   .5234934857    , S.D.=   .9388390734     |
| Model size: Observations =      64, Parameters =  16, Deg.Fr.=     48 |
| Residuals:  Sum of squares= 2.020048105    , Std.Dev.=         .20514 |
| Fit:        R-squared=  .963622, Adjusted R-squared =          .95225 |
| Model test: F[ 15,     48] =   84.77,    Prob value =          .00000 |
| Diagnostic: Log-L =     19.7723, Restricted(b=0) Log-L =     -86.2690 |
|             LogAmemiyaPrCrt.=   -2.945, Akaike Info. Crt.=      -.118 |
| Autocorrel: Durbin-Watson Statistic =   2.03481,   Rho =      -.01740 |
+-----------------------------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient  | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
 Constant -1.098761439      .15118105       -7.268   .0000
 COMPULSO  .7342851113      .13722281        5.351   .0000  .28125000
 ARRIVAL   .2600033269      .12913409        2.013   .0497  .93750000E-01
 CONCURR   .3857261057      .86647526E-01    4.452   .0001  .26562500
 SEATSHAR  .1408316616E-01  .14183266E-01     .993   .3257  8.2995483
 RESTDAY   .3803437397      .12093550        3.145   .0028  .68750000
 LIST      .7729731584E-01  .23361384         .331   .7422  .43750000
 CONSTITU  .5903790503      .14958587        3.947   .0003  .31250000
 NETPAYER -.2782095054      .90363656E-01   -3.079   .0034  .18750000
 NETRECEI  .3234576870E-02  .15331452         .021   .9833  .28125000
 LNNP      .5881444408      .82296688E-01    7.147   .0000  1.4952722
 Y1999    -.1854150273      .73003454E-01   -2.540   .0144  .23437500
 ITALCOMP -.6993530377      .22283279       -3.138   .0029  .46875000E-01
 UK       -1.151199970      .27578130       -4.174   .0001  .78125000E-01
 LUXEMBOU  .5812383689      .19121835        3.040   .0038  .78125000E-01
 SWEDEN   -.4861514350      .19722959       -2.465   .0173  .31250000E-01