Voter turnout is a broad topic that has been described and analysed by academics in Poland (Buć 2015; Cześnik 2009; ed. Kowalski 2003; Zarycki 2008) and elsewhere (Avery 2015; Kostadinova 2003; Mansley & Demšar 2015; Salim Saib 2017). Voter turnout – its increase or decrease from one election to another, and its potential determinants – can be crucial to the outcome. This article examines the record increase in Polish turnout in the 2019 European Parliament (EP) election in relation to the two largest electoral committees,
An electoral committee is a committee that proposes candidates, usually on behalf of a political party or coalition of political parties, although they can also be independent.
Studies conducted to date argue for various interpretations of, and prognoses for, the connections between these indicators. One such interpretation is the party influence hypothesis. This posits that there are parties whose constituents evince similar socio-economic traits to people who vote regularly. They are generally well-educated, high earning, advanced in years (Hansford & Gomez 2010), urban dwelling and religious (Skorupska 2019). PO constituents are typically well-educated, high earning and urban dwelling, whereas PiS constituents are typically religious and advanced in years. It can therefore be assumed that PO would more likely win in the event that only those who never fail to vote went to the polls (i.e. if turnout was low). However, the higher the turnout, the greater will be the number of voters exhibiting the opposite characteristics, and who are therefore more likely to support the party opposed to them (i.e. PiS). Moreover, studies by Pacek and Radcliff (Hansford & Gomez 2010) and others (Najbar 2017) show that people of low socio-economic status tend to vote for left-leaning parties (in Poland, this includes the prosocial PiS). The party influence hypothesis predicts that PiS would benefit most from a higher turnout.
De Nardo (Hansford & Gomez 2010) views the relationship between turnout and support from the opposite standpoint. He contends that electorates consist of core and peripheral voters. The former tend to have a strong political affiliation, while the latter are less constant. Furthermore, a high turnout is favourable to parties that do not have a core constituency (the deeply conservative and prosocial PiS has a core constituency [Cybulska 2011; Cybulska 2015; Maliszewski et al. 2016]). This is known as the two-effects hypothesis, and it predicted that a high turnout would favour the opposition PO-led coalition KE. Polish studies are ambiguous on the relationship between frequency and party support. The sticking point lies in defining a party with a core constituency. Once PiS established its manifesto, while PO’s remained unclear, especially on social mores, a high turnout favoured PO (which had more support in the parliamentary elections in 2007 than it did in 2005). However, once PO stabilised its constituency as PiS had done, higher turnout began to cost it votes (Gendźwiłł et al. 2014). The prevailing view in Poland of late has been that a high turnout favours the governing party and that this is due to its successful campaigning.
E.g. Mikołaj Cześnik: “This was evident in the European Parliament elections, where there would not have been such a high turnout had PiS not mobilized its constituency.” (Terlik 2019) Further: p.p. From 23.83% in 2014 to 45.68% in 2019 Poland is administratively subdivided into voivodeships, which are subdivided into poviats, which are in turn subdivided into gminas.
The following control variables were used to gauge the impact of other factors that potentially altered the level of support for the two major electoral committees from 2014 to 2019 (i.e. other than the difference in voter turnout): the 2014 committee result and the difference in the: (i) number of people per sq. km. franchised to vote in each EP election;
A normal distribution was achieved by using the logarithm of the density of the number of franchised voters. This is a synthetic indicator calculated using the linear allocation method with the following variables: percentage of people with sewerage; percentage of people with running water; number of people per household; and rate of urbanisation. For this reason, the model includes such control variables as difference in support for the opposition party, degree of feminisation, and gmina size as measured by number of franchised voters. Nor was it possible to design a model with difference in voter turnout between 2014 and 2019 as a control variable.
European Parliament elections are held every five years, and Poland has participated since joining the EU in 2004, i.e. in 2004, 2009, 2014, and 2019. In Poland, the Electoral Code (Dz. U. 2011 Nr 21 poz. 112 ze zm. b.d.) provides that EP elections are held on a non-working day. This has so far always been a Sunday. EP elections have never attracted a great deal of interest in Poland. Although the mood following the 2003 accession referendum was upbeat, and 59% of Poles surveyed declared their intention to vote in the first election (Badora 2019), only 20.87% of registered voters actually did so (Państwowa Komisja Wyborcza 2004). The turnout was similarly low in subsequent years (Table 1).
Declaration of intention to vote in EP elections vs actual turnout in 2004–19
45%* | 36% | 40% | 60% | |
20.87% | 24.53% | 23.83% | 45.68% | |
24.13 p.p. | 11.47 p.p. | 16.17 p.p. | 14.32 p.p. |
Poles most often attribute their failure to vote in EP elections to a lack of interest in politics, dissatisfaction with the domestic political situation and the belief that politicians treat these elections as a means to a well-paid sinecure. Moreover, Poles regard EP elections as the elections of least relevance to them (Cybulska 2014). For this reason, these elections are considered second-order, both in Poland and internationally (Dagnis Jensen 2020).
The turnout in Poland varies by area, as it does in EP elections and national parliamentary and presidential elections (i.e. in central elections).
The voting trends in local elections are the inverse of those in general elections.
The increased turnout in the 2019 EP election was uniform throughout the country (the standard deviation was 1.5), although it was higher in some areas. The national average increase in turnout was 21.85 p.p. The increases were largest in the voivodeships of Mazowieckie (24.46 p.p.), Świętokrzyskie (23.46 p.p) and Łódzkie (23.2 p.p.), and smallest in Kujawsko-Pomorskie (19.16 p.p.), Podlaskie (19.31 p.p.) and Warmińsko-Mazurskie (19.84 p.p.). The largest increase was recorded in Gmina Wiśniowa in Małopolskie Voivodeship (31.44 p.p.) and the smallest in Gmina Kikół in Kujawsko-Pomorskie Voivodeship (4.69 p.p.) (Figure 3).
It can be instructive to analyse the change in support for particular parties against the background of this increase in voter turnout. Twelve electoral committees, of which ten were party committees and two were coalition committees, stood for the 2014 EP election.
There is a threshold of 5% for political parties and 8% for committees in Poland.
Nationally, an increase in turnout of 6,459,324 votes, i.e. 21.85 p.p, can be observed. Both electoral committees recorded a considerable increase in the number of voters as a result. However, PiS gained 6.46 p.p. and KE lost 17.37 p.p. nationally. These results clearly indicate that the increase in turnout was primarily due to PiS’s constituency being mobilised.
The voivodeship-level analysis also confirms the thesis that the “right-leaning” electorate, was mobilised, albeit not uniformly. In Pomorskie Voivodeship, PiS barely gained 1 p.p. more than KE, while in Podkarpackie it gained almost seven times as much as PO. PiS gained at least twice as much as KE in only five voivodeships, and these voivodeships had the greatest bearing on the nationwide increase in its support. However, a high degree of mobilisation does not necessarily go in tandem with a high turnout. PiS gained at least twice as much as KE in only one (Świętokrzyskie) of the five voivodeships with the largest increase in turnout. It is worth examining the gmina-level results in order to see the full picture. As shown on the map below (Figure 4), the largest increase in turnout was recorded in the south of Mazowieckie and Świętokrzyskie voivodeships, the north of Śląskie and Małopolskie voivodeships and the west of Łódzkie voivodeship. The lowest increase in turnout was along the country’s eastern border, especially in those parts of Podlasie with Orthodox Christian populations. A modest increase in turnout can also be observed along the border with Kalingrad Oblast. There was not a single gmina in the entire country in which the turnout was lower than it had been in 2014. As mentioned, support for PiS increased in most gminas. Most of the new votes for PiS were in Świętokrzyskie Voivodeship, the north and west of Lubelskie and the south of Małopolskie. It is interesting to note that those gminas that witnessed the greatest increase in support for PiS in 2014 had voted PSL in 2014 (ed. Solarz 2018). This is confirmed by exit polls. An Ipsos survey found that more than 70% of those who described themselves as farmers (PSL is regarded as an agricultural party) voted for PiS (Wilgocki et al. 2019). This suggests that the new votes for PiS mostly came from voters who had previously voted PSL, rather than from new voters. The smallest increase in support for PiS was recorded in most voivodeship capitals, and in the party’s existing strongholds, viz. central Podlasie, northern Podkarpacie, and north-west Lubelskie Voivodeship, and north-east Łódzkie voivodeship. It may well be that these places, with their preponderance of PiS supporters (support for this party has never fallen below 50% and often exceeds 80%), would find it difficult to attract new voters. Moreover, the increase in turnout compared with 2014 was approximately equal to the national average in these places (Figure 5). Support for PiS rose in 2,370 gminas, the largest increase being in Gmina Borki, Lubelskie Voivodeship (49.06 p.p.), and fell in 107 gminas, the largest decrease being in Gmina Dąbrowice, Łódkie Voivodeship (10.73 p.p.).
There was a definite drop in support for KE in 2019 compared with 2014. Only 13 gminas recorded an increase in support for KE, the largest increase being a mere 4.55 p.p. (in Gmina Lublin, Lubelskie Voivodeship). Support fell in 2,464 gminas, the largest decrease being in Gmina Świedziebnia, Kujawsko-Pomorskie Voivodeship (51.74 p.p.). KE lost least support in the voivodeship capitals (and actually gained support in Lublin and Białystok), and in central Podlasie (probably due to the increasing sprawl of the Białystok metropolitan area) and the north of Podkarpackie Voivodeship. The most interesting of this group is Małopolskie Voivodeship, which defied the trend with a rise in support for PiS without any fall in support for KE. As a result, this voivodeship recorded one of the highest increases in voter turnout in the country.
The interdependence described above is even more evident on the maps illustrating the changes in turnout and support vis-à-vis the national averages (Figure 5). Every gmina with a greater-than-average rise in turnout is described as an increase, and every gmina with a lower rise in turnout is described as a decrease. Those gminas in which the difference exceeds plus-or-minus 2 p.p. are also marked. When the maps illustrating the changes in party support are compared, it becomes clear that the rise in turnout was not always accompanied by a significant increase in support for PiS. On the contrary, it is plain to see that the rise in turnout was well below the national average in most regions that recorded an exceptionally high increase in support for PiS. These regions include the north and east of Lubelskie Voivodeship, eastern Świętokrzyskie, the west of Podlaskie, and northern Mazowsze. At the same time, regions whose rise in turnout exceeded the national average by more than 5 p.p. often recorded an increase in support for KE. These included most voivodeship capitals, Małopolska region, southern Mazowieckie and Łódzkie. This implies that even though PiS won the 2019 PE election, the higher turnout was largely due to the mobilisation of the KE constituency. This is clearly evident in the linear regression model designed for the two party groupings (Table 3). Of the 15 voivodeship models designed for PiS,
The model for Łódź voivodeship was not statistically significant.
Breakdown and comparison of votes cast in the 2019 EP and 2014 EP elections
7,301,650 | 13,760,974 | 6,459,324 | |
2,751,682 | 6,192,780 | 3,441,098 | |
3,441,861 | 5,249,935 | 1,808,074 | |
875,942 | 2,204,596 | 1,328,654 | |
232,165 | 113,663 | −118,502 |
Linear regression model for support for PiS in 2014–19 and support for KE in 2014–19
Dolnośląskie, N=169 | R2 | ||||||
Constant | −0.236 | 0.051 | 0.288 | 0.062 | |||
difference in turnout 2014–2019 | 0.874 | 0.213 | −0.863 | 0.231 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.064 | 0.326 | −0.026 | −0.316 | 0.351 | −0.120 | |
difference in standard of living 2014–2019 | −0.491 | 0.210 | 0.505 | 0.226 | |||
difference in % of regular churchgoers 2014–2017 | 0.000 | 0.000 | 0.086 | 0.000 | 0.000 | −0.059 | |
difference in number of franchised voters 2014–2019 | −7.00E-06 | 0.000 | −0.108 | 1.214E-05 | 0.000 | 0.169 | |
2014 party result | −0.183 | 0.064 | −0.036 | 0.068 | −0.041 | ||
Kujawsko-Pomorskie, N=144 | R2 | ||||||
Constant | −0.336 | 0.088 | 0.474 | 0.047 | |||
difference in turnout 2014–2019 | 1.665 | 0.255 | −1.564 | 0.232 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.350 | 0.411 | −0.065 | 0.097 | 0.398 | 0.020 | |
difference in standard of living 2014–2019 | −0.111 | 0.687 | −0.012 | −0.104 | 0.664 | −0.012 | |
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.054 | 0.001 | 0.001 | 0.050 | |
difference in number of franchised voters 2014–2019 | −7.944E-06 | 0.000 | −0.055 | 9.395E-06 | 0.000 | 0.071 | |
2014 party result | −0.306 | 0.104 | −0.092 | 0.105 | −0.066 | ||
Łódzkie, N=177 | R2 | 0.09 | 0.08 | ||||
Constant | −0.122 | 0.054 | 0.262 | 0.075 | |||
difference in turnout 2014–2019 | 0.338 | 0.216 | 0.118 | −0.173 | 0.253 | −0.053 | |
difference in logarithm of density of franchised voters 2014–2019 | −0.019 | 0.392 | −0.004 | −0.170 | 0.456 | −0.029 | |
difference in standard of living 2014–2019 | −1.041 | 0.469 | 0.960 | 0.548 | 0.134 | ||
difference in % of regular churchgoers 2014–2017 | 0.000 | 0.001 | −0.036 | 0.000 | 0.001 | 0.044 | |
difference in number of franchised voters 2014–2019 | −7.345E-06 | 0.000 | −0.052 | 3.046E-07 | 0.000 | 0.002 | |
2014 party result | −0.281 | 0.069 | −0.214 | 0.074 | |||
Lubelskie, N=213 | R2 | ||||||
Constant | 0.061 | 0.052 | 0.511 | 0.044 | |||
difference in turnout 2014–2019 | 0.170 | 0.190 | 0.047 | −0.215 | 0.206 | −0.065 | |
difference in logarithm of density of franchised voters 2014–2019 | 0.791 | 0.335 | −0.878 | 0.374 | |||
difference in standard of living 2014–2019 | −1.714 | 0.369 | 1.702 | 0.417 | |||
difference in % of regular churchgoers 2014–2017 | 0.001 | 0.001 | 0.043 | 0.001 | −0.055 | ||
difference in number of franchised voters 2014–2019 | 1.918E-07 | 0.000 | 0.011 | −1.622E-07 | 0.000 | −0.010 | |
2014 party result | −0.736 | 0.054 | −0.560 | 0.063 | |||
Lubuskie, N=82 | R2 | ||||||
Constant | −0.227 | 0.072 | 0.382 | 0.068 | |||
difference in turnout 2014–2019 | 1.319 | 0.273 | −1.151 | 0.273 | |||
difference logarithm of in density of franchised voters 2014–2019 | 0.023 | 0.021 | 0.108 | −0.009 | 0.021 | −0.047 | |
difference in standard of living 2014–2019 | −0.027 | 0.247 | −0.011 | 0.055 | 0.246 | 0.025 | |
difference in % of regular churchgoers 2014–2017 | 0.001 | 0.001 | 0.070 | 0.000 | 0.001 | 0.053 | |
difference in number of franchised voters 2014–2019 | −7.343E-07 | 0.000 | −0.027 | 1.387E-06 | 0.000 | 0.056 | |
2014 party result | −0.326 | 0.095 | −0.121 | 0.098 | −0.134 | ||
Małopolskie, N=182 | R2 | ||||||
Constant | 0.120 | 0.036 | 0.053 | 0.057 | |||
difference in turnout 2014–2019 | −0.415 | 0.161 | 0.683 | 0.186 | |||
difference in logarithm of density of franchised voters 2014–2019 | 0.596 | 0.328 | 0.112 | −0.825 | 0.374 | ||
difference in standard of living 2014–2019 | −0.068 | 0.193 | −0.022 | −0.125 | 0.223 | −0.041 | |
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | 0.002 | 0.001 | |||
difference in number of franchised voters 2014–2019 | 2.767E-06 | 0.000 | 0.024 | −3.099E-06 | 0.000 | −0.028 | |
2014 party result | −0.424 | 0.045 | −0.218 | 0.047 | |||
Mazowieckie, N=314 | R2 | ||||||
Constant | −0.216 | 0.050 | 0.407 | 0.046 | |||
difference in turnout 2014–2019 | 0.698 | 0.162 | −0.925 | 0.192 | |||
difference in logarithm of density of franchised voters 2014–2019 | 1.531 | 0.222 | −2.140 | 0.277 | |||
difference in standard of living 2014–2019 | −1.257 | 0.233 | 1.320 | 0.288 | |||
difference in % of regular churchgoers 2014–2017 | 0.000 | 0.001 | 0.020 | 0.000 | 0.001 | −0.004 | |
difference in number of franchised voters 2014–2019 | −3.230E-06 | 0.000 | −0.041 | 4.898E-06 | 0.000 | 0.053 | |
2014 party result | −0.331 | 0.050 | −0.054 | 0.058 | −0.049 | ||
Opolskie, N=71 | R2 | ||||||
Constant | −0.179 | 0.090 | 0.106 | 0.039 | |||
difference in turnout 2014–2019 | −0.070 | 0.223 | −0.044 | −0.035 | 0.230 | −0.020 | |
difference in logarithm of density of franchised voters 2014–2019 | −0.734 | 0.286 | 0.817 | 0.309 | |||
difference in standard of living 2014–2019 | −0.885 | 0.280 | 0.828 | 0.307 | |||
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.175 | 0.001 | 0.001 | 0.164 | |
difference in number of franchised voters 2014–2019 | −5.488E-06 | 0.000 | −0.064 | 6.659E-06 | 0.000 | 0.072 | |
2014 party result | 0.036 | 0.105 | 0.048 | 0.091 | 0.112 | 0.110 | |
Podkarpackie, N=160 | R2 | ||||||
Constant | −0.020 | 0.050 | 0.319 | 0.046 | |||
difference in turnout 2014–2019 | 0.137 | 0.200 | 0.051 | −0.186 | 0.226 | −0.069 | |
difference in logarithm of density of franchised voters 2014–2019 | −0.118 | 0.310 | −0.025 | −0.040 | 0.364 | −0.008 | |
difference in standard of living 2014–2019 | −0.120 | 0.211 | −0.038 | 0.109 | 0.252 | 0.034 | |
difference in % of regular churchgoers 2014–2017 | 0.000 | 0.001 | 0.028 | 0.000 | 0.001 | 0.026 | |
difference in number of franchised voters 2014–2019 | 3.163E-07 | 0.000 | 0.009 | 3.436E-07 | 0.000 | 0.009 | |
2014 party result | −0.441 | 0.055 | −0.297 | 0.062 | |||
Podlaskie, N=118 | R2 | ||||||
Constant | −0.241 | 0.090 | 0.507 | 0.086 | |||
difference in turnout 2014–2019 | 1.035 | 0.381 | −1.127 | 0.432 | |||
difference in logarithm of density of franchised voters 2014–2019 | 1.030 | 0.590 | 0.143 | −1.285 | 0.684 | −0.173 | |
difference in standard of living 2014–2019 | −2.742 | 0.690 | 3.067 | 0.798 | |||
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.039 | 0.001 | 0.001 | 0.067 | |
difference in number of franchised voters 2014–2019 | 9.003E-06 | 0.000 | 0.044 | −1.379E-05 | 0.000 | −0.066 | |
2014 party result | −0.371 | 0.061 | −0.234 | 0.064 | |||
Pomorskie, N=123 | R2 | ||||||
Constant | −0.260 | 0.060 | 0.277 | 0.042 | |||
difference in turnout 2014–2019 | 1.026 | 0.203 | −0.836 | 0.198 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.164 | 0.184 | −0.076 | 0.117 | 0.180 | 0.059 | |
difference in standard of living 2014–2019 | −0.510 | 0.257 | 0.497 | 0.249 | |||
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.111 | 0.001 | 0.001 | 0.073 | |
difference in number of franchised voters 2014–2019 | 1.191E-06 | 0.000 | 0.035 | −6.215E-07 | 0.000 | −0.020 | |
2014 party result | −0.167 | 0.068 | −0.024 | 0.062 | −0.032 | ||
Śląskie, N=167 | R2 | ||||||
Constant | 0.077 | 0.054 | 0.077 | 0.046 | |||
difference in turnout 2014–2019 | −0.439 | 0.176 | 0.494 | 0.196 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.229 | 0.328 | −0.054 | 0.222 | 0.375 | 0.048 | |
difference in standard of living 2014–2019 | −0.237 | 0.209 | −0.087 | 0.126 | 0.235 | 0.042 | |
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.076 | 0.001 | 0.001 | 0.077 | |
difference in number of franchised voters 2014–2019 | 1.807E-06 | 0.000 | 0.118 | −1.775E-06 | 0.000 | −0.107 | |
2014 party result | −0.261 | 0.054 | −0.159 | 0.056 | |||
Świętokrzyskie, N=102 | R2 | ||||||
Constant | 0.172 | 0.092 | 0.478 | 0.066 | |||
difference in turnout 2014–2019 | −0.166 | 0.267 | −0.055 | 0.177 | 0.307 | 0.059 | |
difference in logarithm of density of franchised voters 2014–2019 | −0.397 | 0.549 | −0.056 | 0.421 | 0.660 | 0.059 | |
difference in standard of living 2014–2019 | −0.089 | 0.357 | −0.019 | −0.114 | 0.427 | −0.024 | |
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.025 | 0.000 | 0.002 | 0.024 | |
difference in number of franchised voters 2014–2019 | −3.274E-06 | 0.000 | −0.064 | 1.867E-06 | 0.000 | 0.037 | |
2014 party result | −0.773 | 0.088 | −0.672 | 0.111 | |||
Warmińsko-Mazurskie, N=116 | R2 | ||||||
Constant | −0.121 | 0.068 | 0.396 | 0.055 | |||
difference in turnout 2014–2019 | 0.748 | 0.252 | −1.117 | 0.256 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.019 | 0.432 | −0.004 | −0.582 | 0.438 | −0.119 | |
difference in standard of living 2014–2019 | −0.826 | 0.469 | −0.167 | 1.312 | 0.471 | ||
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.138 | 0.001 | 0.001 | 0.085 | |
difference in number of franchised voters 2014–2019 | −3.611E-06 | 0.000 | −0.021 | −9.050E-06 | 0.000 | −0.050 | |
2014 party result | −0.342 | 0.095 | −0.135 | 0.093 | −0.121 | ||
Wielkopolskie, N=226 | R2 | ||||||
Constant | −0.423 | 0.070 | 0.412 | 0.052 | |||
difference in turnout 2014–2019 | 1.445 | 0.235 | −1.517 | 0.239 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.102 | 0.316 | −0.022 | 0.017 | 0.327 | 0.004 | |
difference in standard of living 2014–2019 | −0.378 | 0.338 | −0.069 | 0.358 | 0.349 | 0.064 | |
difference in % of regular churchgoers 2014–2017 | −0.001 | 0.001 | −0.062 | 0.001 | 0.001 | 0.067 | |
difference in number of franchised voters 2014–2019 | 2.812E-08 | 0.000 | 0.000 | 2.045E-06 | 0.000 | 0.019 | |
2014 party result | −0.157 | 0.070 | 0.195 | 0.073 | |||
Zachodniopomorskie, N=113 | R2 | ||||||
Constant | −0.103 | 0.055 | 0.442 | 0.045 | |||
difference in turnout 2014–2019 | 0.798 | 0.162 | −0.877 | 0.172 | |||
difference in logarithm of density of franchised voters 2014–2019 | −0.407 | 0.193 | 0.434 | 0.208 | |||
difference in standard of living 2014–2019 | −1.857 | 0.446 | 1.824 | 0.481 | |||
difference in % of regular churchgoers 2014–2017 | −0.003 | 0.001 | 0.003 | 0.001 | |||
difference in number of franchised voters 2014–2019 | −2.465E-06 | 0.000 | −0.049 | 4.314E-06 | 0.000 | 0.081 | |
2014 party result | −0.435 | 0.085 | −0.340 | 0.087 |
Dependent variable: difference in support for KE 2019-2014
p-values (
This is also confirmed by exit polls conducted by Ipsos for TVN24. Ipsos found that new voters supported KE more strongly than any other party (35.1%), with considerably fewer voting for PiS (22.5%) (Wilgocki et al. 2019). Stated in numbers, 2,267,223 of the 6,459,324 new votes went to KE and 1,453,348 went to PiS. Even after the effectiveness of the survey is validated,
The Ipsos survey of the 2019 EP election was not exact. The result for PiS was underestimated by 3 p.p. and the result for KE was overestimated by 1 p.p. (Goczał 2019). See Table 2.
It is worth paying attention to the control variables when analysing the regression results. The 2014 party result was the most frequently significant of them. While it was a destimulant for both electoral committees, in the case of PiS it was decidedly rarer (7 voivodeships) and weaker (the maximum value was −0.620 for Świętokrzystkie). The 2014 result for KE significantly destimulated the change in support in 15 models (the maximum value was −0.727 for Lubelskie). Change in standard of living was also a major predictor. This was a stimulant for the change in PiS support and a destimulant for the change in KE support in eight cases. However, it was a significantly weaker determinant than those described above (the maximum
Voter turnout may have a crucial impact on the result of an election. This impact, however, may take different forms. Poland’s electoral history indicates that a high turnout tends to favour socially progressive parties with an inconstant constituency, especially KE. However, the high turnout in the most recent EP election resulted in a PiS victory. This is consistent with the party influence hypothesis. There was nothing untoward about the rise in frequency in and of itself, as the political situation in Poland in 2019 was such as to have keenly aroused the interest of the electorate (social polarisation, consolidation of political parties, the results of opinion polls prior to the parliamentary election). The biggest surprise, however, was that this turnout affected different regions in different ways. Most importantly, there was only an average rise in turnout in those areas where support for PiS was greatest, while the largest rise in turnout was accompanied by an increase (or only a slight decrease) in support for KE. PiS gained least where it had always enjoyed strong support, as well as where it had always had least support. The increase in the number of votes for PiS was largely the result of winning over people who had voted for PSL in 2014 (mainly in Świętokrzyskie Voivodeship). Those voters were probably averse to supporting the remaining coalition members, as they had completely different political outlooks (e.g. PSL is traditional, while SLD is hard-left and post-communist).
An analysis of the election results, the opinion polls, and the comparative maps clearly shows that, at the local level, there is no justification for concluding that the high turnout assisted PiS. The governing party decidedly won the most recent EP election, gaining almost 3.5 million new voters. However, while it cannot be conclusively stated that the increase in turnout was attributable to these same voters, the argument that PiS would have won with an even larger majority had the turnout been typical (i.e. approx. 20–25%) can definitely be supported. This is because the increase in support for PiS was mainly due to shifting allegiances in the electorate, while the higher turnout was largely due to the mobilisation of KE’s constituency. Moreover, the hypothesis concerning the geographic differentiation of the relationship between voter turnout and support for the two largest committees was confirmed. This differentiation is not very strong, but it indicates the necessity to search for contextual explanations in research in the field of electoral geography. The confirmation of this hypothesis also suggests that the use of the above conclusions regarding party influence effect should not be applied uniformly throughout Poland.