Skip to contents

From two parties to four: Quebec's party system, 1998-2022

Source: vignettes/articles/realignment.Rmd
realignment.Rmd

Version française

This page downloads 11 studies (about 16 MB) the first time it runs. options(qesR.cache = "disk") keeps them on disk for later sessions.

For about three decades, Quebec’s party system rested on the sovereignty question. From the 1970s on, the Parti Québécois (PQ) and the Parti libéral du Québec (PLQ) alternated in power, and where Quebecers stood on independence went a long way toward explaining how they voted. In 1998 the two parties still took 86% of the valid votes. However, after the breakthrough of the Action démocratique du Québec (ADQ) in 2007, Bélanger and Nadeau (2009) asked whether the recent change in voters’ behaviour was a passing incident or a lasting transformation of Quebec’s political landscape. The rise of Québec solidaire (QS) and of the Coalition avenir Québec (CAQ) has since settled part of that question. In 2022 five parties each took more than 12% of the vote, and the effective number of parties (Laakso and Taagepera 1979) rose from 2.6 in 1998 to 4.0, the equivalent of four parties of equal size. It is now commonplace to say that the system has fragmented, and the 2018 election, which ended the alternation of the PQ and the PLQ in power, is the starting point of recent work on the “new” Quebec voter (Bélanger et al. 2022). The official results tell us how much the system fragmented. They do not tell us among whom, nor whether the sovereignty question still organizes the vote.

The Quebec Election Studies give a fairly clear answer to both questions. First, the fragmentation is concentrated among francophones. From the first Quebec Election Study, in 2007, to 2022, their effective number of parties went from 3.5 to 3.9. The 1998 polls put it lower, at 2.8, but they surveyed francophones only and have no validated weight, so that point is a rough guide. Non-francophones, by contrast, kept voting Liberal until 2018: 72% of them supported the PLQ that year, compared to 12% of francophones. Second, the sovereignty question still divides the francophone vote, but it no longer channels it into two parties. Since 2012 sovereigntists have almost never voted PLQ and federalists have rarely voted PQ, yet the share of francophones who voted for the party of their own camp fell from 55% in 2012 to 22% in 2022. In other words, knowing where a francophone voter stands on independence now tells us more about the party they will not choose than about the one they will. The idea that the sovereignty question has faded from Quebec elections therefore finds only mixed support.

The francophone vote fragmented first

The francophone vote fragmented, from 3.5 to 3.9 effective parties; the non-francophone vote stayed far more concentratedEffective number of parties at each election: official result and reported vote by mother tongue, with 95% confidence intervals
Line chart of the effective number of parties at each Quebec election from 1998 to 2022: the official result, francophone respondents and non-francophone respondents. Among francophones it rises from 2.8 in 1998 to 3.9 in 2022; among non-francophones it stays between 1.3 and 2.5 until 2018 and reaches 2.9 in 2022, with a wide interval. Values in the table view.Line chart of the effective number of parties at each Quebec election from 1998 to 2022: the official result, francophone respondents and non-francophone respondents. Among francophones it rises from 2.8 in 1998 to 3.9 in 2022; among non-francophones it stays between 1.3 and 2.5 until 2018 and reaches 2.9 in 2022, with a wide interval. Values in the table view.

Source: Quebec Election Studies (one per election) and, for 1998, the CROP and CREATEC polls, which interviewed francophones only, each by its own definition (mother tongue for CREATEC, the language spoken most often at home for CROP); official results of Élections Québec. The effective number of parties (Laakso and Taagepera 1979) is 1 divided by the sum of the squared vote shares; the smaller parties are grouped into a single category, other parties. Bands: 95% confidence intervals. Hollow dots (1998, 2008): unweighted, no validated weight. The x axis is in years, with a break between 1998 and 2007.

Table view
Election Study Group Effective number [95% CI] n Official, all voters Weighting
1998 1998 polls Francophones 2.78 [2.67, 2.88] 1126 2.58 unweighted: no validated weight
2007 QES 2007 Francophones 3.54 [3.39, 3.70] 1525 3.47 weighted
2007 QES 2007 Non-francophones 2.53 [1.94, 3.12] 163 3.47 weighted
2007 2007 panel Francophones 3.45 [3.34, 3.56] 1337 3.47 unweighted: no validated weight
2007 2007 panel Non-francophones 1.78 [1.49, 2.06] 155 3.47 unweighted: no validated weight
2008 QES 2008 Francophones 3.12 [2.96, 3.29] 771 3.03 unweighted: no validated weight
2008 QES 2008 Non-francophones 1.58 [1.32, 1.84] 127 3.03 unweighted: no validated weight
2012 QES 2012 Francophones 3.12 [2.93, 3.31] 1094 3.60 weighted
2012 QES 2012 Non-francophones 2.14 [1.75, 2.53] 180 3.60 weighted
2012 2012 panel Francophones 3.42 [3.20, 3.64] 576 3.60 unweighted: no validated weight
2012 2012 panel Non-francophones 2.31 [1.59, 3.03] 57 3.60 unweighted: no validated weight
2014 QES 2014 Francophones 3.68 [3.51, 3.84] 981 3.38 weighted
2014 QES 2014 Non-francophones 1.34 [1.15, 1.54] 201 3.38 weighted
2018 QES 2018 Francophones 3.62 [3.45, 3.78] 1656 3.88 weighted
2018 QES 2018 Non-francophones 1.85 [1.63, 2.06] 359 3.88 weighted
2018 2018 panel Francophones 3.48 [3.11, 3.85] 560 3.88 weighted
2018 2018 panel Non-francophones 2.02 [1.64, 2.40] 143 3.88 weighted
2022 QES 2022 Francophones 3.92 [3.64, 4.20] 883 3.99 weighted
2022 QES 2022 Non-francophones 2.93 [1.94, 3.92] 121 3.99 weighted

The two survey lines do not follow the same path. Among francophones the effective number of parties rose over the period, though not steadily. It reached a first peak in 2007 (3.5), the year of the ADQ’s breakthrough, fell back to 3.1 in 2012, and has been no lower than 3.6 since 2014. Among non-francophones it was 2.5 in 2007, when the PLQ took 58% of their vote, and from 2008 to 2018 it stayed between 1.3 and 2.1, a vote largely dominated by the PLQ. It reached 2.9 in 2022, when the PLQ fell to 54% of their vote. That said, only 121 non-francophone voters answered that year, and the interval runs from 1.9 to 3.9, wide enough to include the 2007 value. The rise is clear against the 2008 to 2018 elections; its size is not.

The survey lines run close to the official one, but not on it. Surveys over-report some parties and under-report others (see Survey vs official results). In 2022, for instance, 33% of respondents reported a vote for the CAQ, against 41% of the ballots, so the survey makes the 2022 vote look more fragmented than it was (4.7 effective parties among all respondents, 4.0 officially). The survey lines therefore say more about the gap between the two groups, and how it changed, than about the level of either.

Party by party, the language divide held

The CAQ's wins are francophone wins: in 2018 it took 41% of francophones and 12% of other votersShare of the reported vote among francophones and non-francophones, by party, with 95% confidence intervals and the official result
Five small line charts, one per party (PLQ, PQ, QS, ADQ, CAQ), of the share of the reported vote among francophones and among non-francophones at each election from 1998 to 2022, with the official result as a tick. The PLQ takes 72% of the non-francophone vote in 2018 against 12% of the francophone vote; the CAQ takes 41% of francophones and 12% of the others. Values in the table view.Five small line charts, one per party (PLQ, PQ, QS, ADQ, CAQ), of the share of the reported vote among francophones and among non-francophones at each election from 1998 to 2022, with the official result as a tick. The PLQ takes 72% of the non-francophone vote in 2018 against 12% of the francophone vote; the CAQ takes 41% of francophones and 12% of the others. Values in the table view.

Source: Quebec Election Studies (one per election) and, for 1998, the CROP and CREATEC polls (francophones only). Tick: official share of valid votes, all voters. Bars: 95% confidence intervals. Hollow dots (1998, 2008): unweighted, no validated weight. Groups of fewer than 30 respondents are not drawn. The ADQ merged into the CAQ in 2012; the two are kept apart here. The PCQ and the smaller parties are in the table view.

Table view
Party Election Study Group Share of reported vote [95% CI] n Official, all voters Weighting
PLQ 1998 1998 polls Francophones 34.1 [31.4, 36.9] 1126 43.6 unweighted: no validated weight
PLQ 2007 QES 2007 Francophones 20.1 [17.9, 22.6] 1525 33.1 weighted
PLQ 2007 QES 2007 Non-francophones 58.4 [48.1, 68.1] 163 33.1 weighted
PLQ 2007 2007 panel Francophones 23.5 [21.3, 25.8] 1337 33.1 unweighted: no validated weight
PLQ 2007 2007 panel Non-francophones 73.5 [66.1, 79.9] 155 33.1 unweighted: no validated weight
PLQ 2008 QES 2008 Francophones 32.7 [29.5, 36.1] 771 42.1 unweighted: no validated weight
PLQ 2008 QES 2008 Non-francophones 78.7 [70.8, 85.0] 127 42.1 unweighted: no validated weight
PLQ 2012 QES 2012 Francophones 15.5 [13.2, 18.0] 1094 31.2 weighted
PLQ 2012 QES 2012 Non-francophones 65.5 [57.3, 72.8] 180 31.2 weighted
PLQ 2012 2012 panel Francophones 22.7 [19.5, 26.3] 576 31.2 unweighted: no validated weight
PLQ 2012 2012 panel Non-francophones 63.2 [50.0, 74.6] 57 31.2 unweighted: no validated weight
PLQ 2014 QES 2014 Francophones 24.2 [21.3, 27.4] 981 41.5 weighted
PLQ 2014 QES 2014 Non-francophones 86.0 [78.0, 91.4] 201 41.5 weighted
PLQ 2018 QES 2018 Francophones 12.2 [10.7, 13.9] 1656 24.8 weighted
PLQ 2018 QES 2018 Non-francophones 71.8 [66.6, 76.6] 359 24.8 weighted
PLQ 2018 2018 panel Francophones 16.1 [13.0, 19.7] 560 24.8 weighted
PLQ 2018 2018 panel Non-francophones 67.8 [59.0, 75.5] 143 24.8 weighted
PLQ 2022 QES 2022 Francophones 6.1 [4.4, 8.4] 883 14.4 weighted
PLQ 2022 QES 2022 Non-francophones 54.0 [40.5, 67.0] 121 14.4 weighted
PQ 1998 1998 polls Francophones 45.9 [43.0, 48.8] 1126 42.9 unweighted: no validated weight
PQ 2007 QES 2007 Francophones 34.1 [31.3, 37.1] 1525 28.3 weighted
PQ 2007 QES 2007 Non-francophones 15.0 [8.3, 25.4] 163 28.3 weighted
PQ 2007 2007 panel Francophones 33.4 [31.0, 36.0] 1337 28.3 unweighted: no validated weight
PQ 2007 2007 panel Non-francophones 7.7 [4.4, 13.1] 155 28.3 unweighted: no validated weight
PQ 2008 QES 2008 Francophones 42.3 [38.8, 45.8] 771 35.2 unweighted: no validated weight
PQ 2008 QES 2008 Non-francophones 7.9 [4.3, 14.0] 127 35.2 unweighted: no validated weight
PQ 2012 QES 2012 Francophones 46.4 [43.1, 49.8] 1094 31.9 weighted
PQ 2012 QES 2012 Non-francophones 6.0 [3.4, 10.2] 180 31.9 weighted
PQ 2012 2012 panel Francophones 41.3 [37.4, 45.4] 576 31.9 unweighted: no validated weight
PQ 2012 2012 panel Non-francophones 8.8 [3.7, 19.4] 57 31.9 unweighted: no validated weight
PQ 2014 QES 2014 Francophones 35.7 [32.3, 39.3] 981 25.4 weighted
PQ 2014 QES 2014 Non-francophones 4.1 [1.4, 11.6] 201 25.4 weighted
PQ 2018 QES 2018 Francophones 23.4 [21.3, 25.7] 1656 17.1 weighted
PQ 2018 QES 2018 Non-francophones 2.6 [1.4, 5.1] 359 17.1 weighted
PQ 2018 2018 panel Francophones 18.0 [14.6, 22.1] 560 17.1 weighted
PQ 2018 2018 panel Non-francophones 2.9 [1.1, 7.5] 143 17.1 weighted
PQ 2022 QES 2022 Francophones 19.6 [17.0, 22.6] 883 14.6 weighted
PQ 2022 QES 2022 Non-francophones 1.7 [0.5, 6.1] 121 14.6 weighted
ADQ 1998 1998 polls Francophones 18.1 [16.0, 20.5] 1126 11.8 unweighted: no validated weight
ADQ 2007 QES 2007 Francophones 34.4 [31.7, 37.4] 1525 30.8 weighted
ADQ 2007 QES 2007 Non-francophones 12.1 [7.0, 20.1] 163 30.8 weighted
ADQ 2007 2007 panel Francophones 34.6 [32.1, 37.1] 1337 30.8 unweighted: no validated weight
ADQ 2007 2007 panel Non-francophones 11.6 [7.4, 17.7] 155 30.8 unweighted: no validated weight
ADQ 2008 QES 2008 Francophones 17.8 [15.2, 20.6] 771 16.4 unweighted: no validated weight
ADQ 2008 QES 2008 Non-francophones 5.5 [2.6, 11.1] 127 16.4 unweighted: no validated weight
QS 2007 QES 2007 Francophones 5.3 [4.0, 7.0] 1525 3.6 weighted
QS 2007 QES 2007 Non-francophones 1.7 [0.4, 7.1] 163 3.6 weighted
QS 2007 2007 panel Francophones 4.0 [3.0, 5.2] 1337 3.6 unweighted: no validated weight
QS 2007 2007 panel Non-francophones 2.6 [1.0, 6.7] 155 3.6 unweighted: no validated weight
QS 2008 QES 2008 Francophones 4.7 [3.4, 6.4] 771 3.8 unweighted: no validated weight
QS 2008 QES 2008 Non-francophones 1.6 [0.4, 6.1] 127 3.8 unweighted: no validated weight
QS 2012 QES 2012 Francophones 6.4 [5.0, 8.1] 1094 6.0 weighted
QS 2012 QES 2012 Non-francophones 6.9 [3.6, 12.8] 180 6.0 weighted
QS 2012 2012 panel Francophones 6.9 [5.1, 9.3] 576 6.0 unweighted: no validated weight
QS 2012 2012 panel Non-francophones 8.8 [3.7, 19.4] 57 6.0 unweighted: no validated weight
QS 2014 QES 2014 Francophones 9.2 [7.5, 11.2] 981 7.6 weighted
QS 2014 QES 2014 Non-francophones 4.7 [2.1, 10.4] 201 7.6 weighted
QS 2018 QES 2018 Francophones 18.9 [16.9, 20.9] 1656 16.1 weighted
QS 2018 QES 2018 Non-francophones 4.9 [3.0, 7.9] 359 16.1 weighted
QS 2018 2018 panel Francophones 14.7 [11.4, 18.9] 560 16.1 weighted
QS 2018 2018 panel Non-francophones 5.3 [2.8, 9.9] 143 16.1 weighted
QS 2022 QES 2022 Francophones 18.8 [16.2, 21.7] 883 15.4 weighted
QS 2022 QES 2022 Non-francophones 11.5 [5.9, 21.3] 121 15.4 weighted
CAQ 2012 QES 2012 Francophones 27.4 [24.5, 30.4] 1094 27.1 weighted
CAQ 2012 QES 2012 Non-francophones 16.8 [11.3, 24.2] 180 27.1 weighted
CAQ 2012 2012 panel Francophones 25.2 [21.8, 28.9] 576 27.1 unweighted: no validated weight
CAQ 2012 2012 panel Non-francophones 10.5 [4.8, 21.5] 57 27.1 unweighted: no validated weight
CAQ 2014 QES 2014 Francophones 27.7 [24.5, 31.1] 981 23.1 weighted
CAQ 2014 QES 2014 Non-francophones 4.2 [1.7, 10.2] 201 23.1 weighted
CAQ 2018 QES 2018 Francophones 41.1 [38.6, 43.7] 1656 37.4 weighted
CAQ 2018 QES 2018 Non-francophones 12.5 [9.2, 16.7] 359 37.4 weighted
CAQ 2018 2018 panel Francophones 45.1 [40.2, 50.2] 560 37.4 weighted
CAQ 2018 2018 panel Non-francophones 16.2 [10.5, 24.2] 143 37.4 weighted
CAQ 2022 QES 2022 Francophones 39.5 [35.9, 43.1] 883 41.0 weighted
CAQ 2022 QES 2022 Non-francophones 12.2 [4.3, 30.2] 121 41.0 weighted
PCQ 2022 QES 2022 Francophones 14.7 [12.2, 17.6] 883 12.9 weighted
PCQ 2022 QES 2022 Non-francophones 9.4 [5.1, 16.7] 121 12.9 weighted
Other 1998 1998 polls Francophones 1.9 [1.2, 2.8] 1126 1.8 unweighted: no validated weight
Other 2007 QES 2007 Francophones 5.9 [4.6, 7.6] 1525 4.1 weighted
Other 2007 QES 2007 Non-francophones 12.8 [7.3, 21.4] 163 4.1 weighted
Other 2007 2007 panel Francophones 4.6 [3.6, 5.8] 1337 4.1 unweighted: no validated weight
Other 2007 2007 panel Non-francophones 4.5 [2.2, 9.2] 155 4.1 unweighted: no validated weight
Other 2008 QES 2008 Francophones 2.6 [1.7, 4.0] 771 2.6 unweighted: no validated weight
Other 2008 QES 2008 Non-francophones 6.3 [3.2, 12.1] 127 2.6 unweighted: no validated weight
Other 2012 QES 2012 Francophones 4.3 [3.3, 5.7] 1094 3.6 weighted
Other 2012 QES 2012 Non-francophones 4.9 [2.5, 9.4] 180 3.6 weighted
Other 2012 2012 panel Francophones 3.8 [2.5, 5.7] 576 3.6 unweighted: no validated weight
Other 2012 2012 panel Non-francophones 8.8 [3.7, 19.4] 57 3.6 unweighted: no validated weight
Other 2014 QES 2014 Francophones 3.2 [2.2, 4.7] 981 2.0 weighted
Other 2014 QES 2014 Non-francophones 0.9 [0.3, 2.4] 201 2.0 weighted
Other 2018 QES 2018 Francophones 4.3 [3.3, 5.6] 1656 3.1 weighted
Other 2018 QES 2018 Non-francophones 8.2 [5.6, 11.8] 359 3.1 weighted
Other 2018 2018 panel Francophones 6.0 [4.0, 9.1] 560 3.1 weighted
Other 2018 2018 panel Non-francophones 7.7 [4.3, 13.5] 143 3.1 weighted
Other 2022 QES 2022 Francophones 1.3 [0.7, 2.5] 883 1.7 weighted
Other 2022 QES 2022 Non-francophones 11.2 [5.1, 22.7] 121 1.7 weighted

The PLQ remained the party of non-francophones. It took 58% of their vote in 2007, 86% in 2014 and still 54% in 2022, its lowest share over the period. Among francophones, however, its share fell from 34% in 1998 to 6% in 2022, well below its official result of 14% that year. The PQ never had many non-francophone voters (15% in 2007) and kept almost none (2% in 2022).

The CAQ’s victories, for their part, are francophone victories. In 2018 it won 41% of the francophone vote and 12% of the others, and in 2022 39% and 12%. Like the ADQ before it, it barely reaches non-francophones. The two parties that came closest to crossing the language divide in 2022 are the two whose breakthrough is most recent. Québec solidaire, which grew among francophones first (19% in 2018), took 19% of francophones and 12% of the others; the Parti conservateur du Québec (PCQ) took 15% and 9%. With so few non-francophone respondents these gaps are uncertain. Using the same 2022 study, Bélanger et al. (2025) find that, unlike most other parties, the PCQ drew as much from non-francophones as from francophones once other factors are held constant; the raw shares shown here do not hold those factors constant. The new parties thus divided the francophone vote first. Among non-francophones the break came only in 2022, and it rests on 121 respondents.

Sovereignty still divides the francophone vote, but no longer channels it

Each camp's old party lost its hold: from 2012 to 2022 the PQ fell from 75% to 38% of francophone sovereigntistsReported vote of francophones, by the side they would take in a referendum on sovereignty. Among federalists, the PLQ went from 34% in 2012 to 10% in 2022
Two panels of stacked horizontal bars, one bar per election from 1998 to 2022: how francophones who would vote Yes and those who would vote No in a referendum split their reported vote among the parties. Among those who would vote Yes, the PQ takes 75% in 2012 and 38% in 2022; among those who would vote No, the PLQ takes 34% in 2012 and 10% in 2022, and the CAQ 43% in 2022. Values in the table view.Two panels of stacked horizontal bars, one bar per election from 1998 to 2022: how francophones who would vote Yes and those who would vote No in a referendum split their reported vote among the parties. Among those who would vote Yes, the PQ takes 75% in 2012 and 38% in 2022; among those who would vote No, the PLQ takes 34% in 2012 and 10% in 2022, and the CAQ 43% in 2022. Values in the table view.

Source: Quebec Election Studies (one per election) and, for 1998, the CROP poll; francophone respondents who reported a vote and chose a side. From 1998 to 2008 the question is the one of the 1995 referendum, which offered a partnership with Canada, with the undecided asked which way they lean; from 2012 on, it asks about an independent country. The Yes side is about as large among these francophones in 2008 (52%) as in 2012 (52%), but the two questions may not draw the same voters, so compare the rows within each period. Labels show shares of 10% or more; the 95% confidence intervals and the Durand panels are in the table view. 1998 and 2008: unweighted, no validated weight.

Table view
Election Study Question Side Party Share of reported vote [95% CI] n Weighting
1998 1998 polls The 1995 question (partnership) Would vote Yes (sovereigntists) PLQ 1.4 [0.4, 5.5] 141 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote Yes (sovereigntists) PQ 83.7 [76.6, 88.9] 141 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote Yes (sovereigntists) ADQ 14.2 [9.3, 21.0] 141 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote Yes (sovereigntists) Other 0.7 [0.1, 4.9] 141 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote No (federalists) PLQ 52.1 [44.5, 59.5] 169 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote No (federalists) PQ 24.9 [18.9, 31.9] 169 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote No (federalists) ADQ 20.7 [15.3, 27.5] 169 unweighted: no validated weight
1998 1998 polls The 1995 question (partnership) Would vote No (federalists) Other 2.4 [0.9, 6.1] 169 unweighted: no validated weight
2007 QES 2007 The 1995 question (partnership) Would vote Yes (sovereigntists) PLQ 2.8 [1.7, 4.7] 715 weighted
2007 QES 2007 The 1995 question (partnership) Would vote Yes (sovereigntists) PQ 58.7 [54.3, 63.0] 715 weighted
2007 QES 2007 The 1995 question (partnership) Would vote Yes (sovereigntists) ADQ 25.7 [22.0, 29.7] 715 weighted
2007 QES 2007 The 1995 question (partnership) Would vote Yes (sovereigntists) QS 7.8 [5.7, 10.7] 715 weighted
2007 QES 2007 The 1995 question (partnership) Would vote Yes (sovereigntists) Other 5.0 [3.4, 7.1] 715 weighted
2007 QES 2007 The 1995 question (partnership) Would vote No (federalists) PLQ 40.0 [35.8, 44.2] 741 weighted
2007 QES 2007 The 1995 question (partnership) Would vote No (federalists) PQ 8.7 [6.6, 11.4] 741 weighted
2007 QES 2007 The 1995 question (partnership) Would vote No (federalists) ADQ 42.4 [38.2, 46.7] 741 weighted
2007 QES 2007 The 1995 question (partnership) Would vote No (federalists) QS 2.5 [1.3, 4.5] 741 weighted
2007 QES 2007 The 1995 question (partnership) Would vote No (federalists) Other 6.4 [4.5, 9.1] 741 weighted
2007 2007 panel The 1995 question (partnership) Would vote Yes (sovereigntists) PLQ 5.0 [3.4, 7.2] 541 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote Yes (sovereigntists) PQ 56.0 [51.8, 60.1] 541 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote Yes (sovereigntists) ADQ 28.1 [24.5, 32.0] 541 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote Yes (sovereigntists) QS 6.3 [4.5, 8.7] 541 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote Yes (sovereigntists) Other 4.6 [3.1, 6.7] 541 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote No (federalists) PLQ 44.9 [40.6, 49.2] 517 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote No (federalists) PQ 9.9 [7.6, 12.8] 517 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote No (federalists) ADQ 39.1 [35.0, 43.3] 517 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote No (federalists) QS 1.4 [0.6, 2.8] 517 unweighted: no validated weight
2007 2007 panel The 1995 question (partnership) Would vote No (federalists) Other 4.8 [3.3, 7.1] 517 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote Yes (sovereigntists) PLQ 9.6 [7.0, 13.0] 376 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote Yes (sovereigntists) PQ 73.7 [69.0, 77.9] 376 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote Yes (sovereigntists) ADQ 6.9 [4.7, 10.0] 376 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote Yes (sovereigntists) QS 7.2 [5.0, 10.3] 376 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote Yes (sovereigntists) Other 2.7 [1.4, 4.9] 376 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote No (federalists) PLQ 60.5 [55.2, 65.6] 342 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote No (federalists) PQ 9.4 [6.7, 12.9] 342 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote No (federalists) ADQ 27.2 [22.7, 32.2] 342 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote No (federalists) QS 1.5 [0.6, 3.5] 342 unweighted: no validated weight
2008 QES 2008 The 1995 question (partnership) Would vote No (federalists) Other 1.5 [0.6, 3.5] 342 unweighted: no validated weight
2012 QES 2012 An independent country Would vote Yes (sovereigntists) PLQ 0.5 [0.1, 1.9] 514 weighted
2012 QES 2012 An independent country Would vote Yes (sovereigntists) PQ 74.8 [70.5, 78.7] 514 weighted
2012 QES 2012 An independent country Would vote Yes (sovereigntists) QS 8.9 [6.5, 12.0] 514 weighted
2012 QES 2012 An independent country Would vote Yes (sovereigntists) CAQ 10.5 [7.9, 13.8] 514 weighted
2012 QES 2012 An independent country Would vote Yes (sovereigntists) Other 5.3 [3.7, 7.5] 514 weighted
2012 QES 2012 An independent country Would vote No (federalists) PLQ 33.7 [29.1, 38.6] 475 weighted
2012 QES 2012 An independent country Would vote No (federalists) PQ 13.0 [10.0, 16.7] 475 weighted
2012 QES 2012 An independent country Would vote No (federalists) QS 4.1 [2.6, 6.4] 475 weighted
2012 QES 2012 An independent country Would vote No (federalists) CAQ 45.7 [40.8, 50.7] 475 weighted
2012 QES 2012 An independent country Would vote No (federalists) Other 3.5 [2.2, 5.6] 475 weighted
2012 2012 panel A sovereign country Would vote Yes (sovereigntists) PLQ 1.4 [0.5, 4.3] 214 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote Yes (sovereigntists) PQ 72.4 [66.1, 78.0] 214 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote Yes (sovereigntists) QS 12.1 [8.4, 17.3] 214 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote Yes (sovereigntists) CAQ 9.3 [6.1, 14.0] 214 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote Yes (sovereigntists) Other 4.7 [2.5, 8.5] 214 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote No (federalists) PLQ 39.7 [34.4, 45.2] 315 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote No (federalists) PQ 19.4 [15.4, 24.1] 315 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote No (federalists) QS 2.9 [1.5, 5.4] 315 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote No (federalists) CAQ 34.6 [29.6, 40.0] 315 unweighted: no validated weight
2012 2012 panel A sovereign country Would vote No (federalists) Other 3.5 [1.9, 6.2] 315 unweighted: no validated weight
2014 QES 2014 An independent country Would vote Yes (sovereigntists) PLQ 1.9 [0.9, 4.2] 397 weighted
2014 QES 2014 An independent country Would vote Yes (sovereigntists) PQ 68.7 [63.4, 73.6] 397 weighted
2014 QES 2014 An independent country Would vote Yes (sovereigntists) QS 13.0 [10.1, 16.6] 397 weighted
2014 QES 2014 An independent country Would vote Yes (sovereigntists) CAQ 12.6 [9.3, 16.9] 397 weighted
2014 QES 2014 An independent country Would vote Yes (sovereigntists) Other 3.7 [2.0, 6.8] 397 weighted
2014 QES 2014 An independent country Would vote No (federalists) PLQ 46.4 [41.3, 51.5] 480 weighted
2014 QES 2014 An independent country Would vote No (federalists) PQ 7.6 [5.2, 10.9] 480 weighted
2014 QES 2014 An independent country Would vote No (federalists) QS 4.6 [2.9, 7.1] 480 weighted
2014 QES 2014 An independent country Would vote No (federalists) CAQ 38.9 [34.0, 44.1] 480 weighted
2014 QES 2014 An independent country Would vote No (federalists) Other 2.6 [1.5, 4.4] 480 weighted
2018 QES 2018 An independent country Would vote Yes (sovereigntists) PLQ 1.0 [0.5, 2.0] 639 weighted
2018 QES 2018 An independent country Would vote Yes (sovereigntists) PQ 45.7 [41.6, 49.9] 639 weighted
2018 QES 2018 An independent country Would vote Yes (sovereigntists) QS 23.5 [20.2, 27.2] 639 weighted
2018 QES 2018 An independent country Would vote Yes (sovereigntists) CAQ 28.6 [24.9, 32.6] 639 weighted
2018 QES 2018 An independent country Would vote Yes (sovereigntists) Other 1.2 [0.5, 2.6] 639 weighted
2018 QES 2018 An independent country Would vote No (federalists) PLQ 24.2 [21.2, 27.5] 773 weighted
2018 QES 2018 An independent country Would vote No (federalists) PQ 5.8 [4.2, 7.8] 773 weighted
2018 QES 2018 An independent country Would vote No (federalists) QS 13.7 [11.3, 16.6] 773 weighted
2018 QES 2018 An independent country Would vote No (federalists) CAQ 48.7 [44.9, 52.5] 773 weighted
2018 QES 2018 An independent country Would vote No (federalists) Other 7.6 [5.7, 10.1] 773 weighted
2018 2018 panel Favourable to independence Would vote Yes (sovereigntists) PLQ 3.5 [1.6, 7.2] 200 weighted
2018 2018 panel Favourable to independence Would vote Yes (sovereigntists) PQ 34.5 [27.3, 42.4] 200 weighted
2018 2018 panel Favourable to independence Would vote Yes (sovereigntists) QS 19.6 [14.2, 26.6] 200 weighted
2018 2018 panel Favourable to independence Would vote Yes (sovereigntists) CAQ 39.2 [31.2, 47.8] 200 weighted
2018 2018 panel Favourable to independence Would vote Yes (sovereigntists) Other 3.3 [1.3, 8.0] 200 weighted
2018 2018 panel Favourable to independence Would vote No (federalists) PLQ 25.4 [20.3, 31.3] 315 weighted
2018 2018 panel Favourable to independence Would vote No (federalists) PQ 6.3 [3.6, 10.6] 315 weighted
2018 2018 panel Favourable to independence Would vote No (federalists) QS 12.8 [8.4, 19.2] 315 weighted
2018 2018 panel Favourable to independence Would vote No (federalists) CAQ 46.9 [40.1, 53.7] 315 weighted
2018 2018 panel Favourable to independence Would vote No (federalists) Other 8.6 [5.3, 13.6] 315 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) PLQ 1.3 [0.5, 3.3] 335 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) PQ 37.9 [32.4, 43.7] 335 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) QS 17.8 [13.8, 22.7] 335 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) CAQ 32.6 [27.1, 38.6] 335 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) PCQ 9.8 [6.7, 14.2] 335 weighted
2022 QES 2022 An independent country Would vote Yes (sovereigntists) Other 0.6 [0.1, 2.6] 335 weighted
2022 QES 2022 An independent country Would vote No (federalists) PLQ 10.4 [7.3, 14.7] 411 weighted
2022 QES 2022 An independent country Would vote No (federalists) PQ 6.2 [4.1, 9.2] 411 weighted
2022 QES 2022 An independent country Would vote No (federalists) QS 17.4 [13.9, 21.6] 411 weighted
2022 QES 2022 An independent country Would vote No (federalists) CAQ 43.4 [38.1, 48.8] 411 weighted
2022 QES 2022 An independent country Would vote No (federalists) PCQ 21.0 [16.9, 25.8] 411 weighted
2022 QES 2022 An independent country Would vote No (federalists) Other 1.5 [0.6, 3.8] 411 weighted

Two things stand out. First, the wall between the two old camps holds. From 2012 on, the PLQ never took more than 2% of the francophone sovereigntist vote, and the PQ never more than 13% of the federalist vote. Second, within each camp the old party lost its hold, and not at the same pace. Federalist francophones were the first to leave, though not in a straight line. In 2007 the ADQ already took about as many of them as the PLQ (42% against 40%), and after a brief return to the PLQ in 2008 (61%), the CAQ took 46% of them in 2012. From 2012 to 2022 the PLQ’s share of federalist francophones fell from 34% to 10%, behind the CAQ, the PCQ and even QS. Sovereigntist francophones stayed with the PQ longer. Its share of their vote went from 75% in 2012 to 38% in 2022, and most of that loss came in 2018, when Québec solidaire and the CAQ divided them. As a result, the share of francophones who voted for the party of their camp fell from 55% in 2012 to 22% in 2022 (the table view gives every election).

The party that gained most is the one that sits on neither side of the question. In 2022 the CAQ drew 43% of the federalist francophone vote and 33% of the sovereigntist one. The sovereignty question has therefore not disappeared from Quebec elections; what disappeared is the two-party system that translated it into votes. This echoes Daoust and Gareau-Paquette (2024), who find that in federal elections support for independence remains a key predictor of the vote, its weight stable since 2006 after a fall from the 2000 and 2004 elections. Here, by contrast, what changed is how that division maps onto parties.

What this means

Overall, the Quebec party system did not fragment evenly. The new parties divided the francophone vote, and they did so within each sovereignty camp: the wall between the camps held while the PQ and the PLQ lost their hold on their own side. Two limits deserve mention. First, these are reported votes in separate surveys: they show how groups of voters split at each election, not whether the same people changed parties (the Transitions page follows the same respondents through a campaign). Second, the non-francophone samples are small, between 121 and 359 voters per election, so their estimates move a lot from one election to the next, and the 2022 break among them remains to be confirmed. The next question is whether the francophone vote will settle again around new poles, or whether generational replacement will keep it divided (the sovereignty page follows the birth cohorts).

How the figures are made

Every figure on this page uses one harmonized variable, vote_choice, restricted to the vote reported after the election, with mother tongue and the referendum question:

h <- qes_harmonize(
  studies = c("qes1998", "qes2007", "qes2007_panel", "qes2008", "qes_crop_2007_2010",
              "qes2012", "qes2012_panel", "qes2014", "qes2018", "qes2018_panel", "qes2022"),
  targets = c("vote_choice", "lang_mother", "sov_support"),
  types = list(vote_choice = "recall"),
  missing = "reasons", quiet = TRUE
)

Within one study, a weighted share and its confidence interval take one qes_design() and one call to the survey package:

library(survey)
d18 <- qes_design(h[h$study == "qes2018", ], weight = "weight_post")
d18 <- subset(d18, !is.na(vote_choice) & !is.na(lang_mother))
svyby(~I(vote_choice == "CAQ"), ~lang_mother, d18, svyciprop,
      method = "logit", vartype = "ci")
#>         lang_mother I(vote_choice == "CAQ")       ci_l      ci_u
#> French       French               0.4114613 0.38603333 0.4373710
#> English     English               0.1101960 0.07564846 0.1578270
#> Other         Other               0.1717929 0.10011622 0.2788822

The figures repeat this for every study. Each estimate is made within one study and never pooled across studies.

Going further

  • Bélanger, Éric, and Richard Nadeau. 2009. Le comportement électoral des Québécois. Montréal: Presses de l’Université de Montréal. https://doi.org/10.4000/books.pum.9772
  • Bélanger, Éric, Jean-François Daoust, Valérie-Anne Mahéo, and Richard Nadeau. 2022. Le nouvel électeur québécois. Montréal: Presses de l’Université de Montréal. https://doi.org/10.1515/9782760645929
  • Bélanger, Éric, Philippe Mongrain, Thomas Gareau-Paquette, and Valérie-Anne Mahéo. 2025. “A Party that Went Viral? The Drivers of Support for the Parti Conservateur du Québec in the 2022 Election.” Canadian Journal of Political Science 58 (2): 277-296. https://doi.org/10.1017/S0008423924000829
  • Daoust, Jean-François, and Thomas Gareau-Paquette. 2024. “Is Quebec Independence Still Key in Making Sense of Canadian Elections? A Longitudinal Analysis (2000-2021).” Regional & Federal Studies 34 (5): 781-806. https://doi.org/10.1080/13597566.2023.2233422
  • Laakso, Markku, and Rein Taagepera. 1979. “‘Effective’ Number of Parties: A Measure with Application to West Europe.” Comparative Political Studies 12 (1): 3-27. https://doi.org/10.1177/001041407901200101
  • Mahéo, Valérie-Anne, and Éric Bélanger. 2018. “Is the Parti Québécois Bound to Disappear? A Study of the Current Generational Dynamics of Electoral Behaviour in Quebec.” Canadian Journal of Political Science 51 (2): 335-356. https://doi.org/10.1017/S0008423917001147

About the data

  • Studies. One Quebec Election Study per election (2007, 2008, 2012, 2014, 2018, 2022) and, for 1998, the CROP and CREATEC polls, which surveyed francophones only (CREATEC by mother tongue, CROP by the language of use); qesR has no mother-tongue column for 1998, and this page counts every 1998 respondent as francophone. The table views also include the Durand panel surveys of 2007, 2012 and 2018, which interviewed the same respondents during the campaign and after the election; they are not drawn. The CROP polls of 2007 to 2010 asked no reported vote and are not used.
  • Vote. The party respondents said they voted for, asked after the election. Those who did not vote, spoiled their ballot, did not know or refused are left out. A party that a study did not list, or that did not run, is not drawn. The ADQ and the CAQ are kept apart, although they merged in 2012.
  • Language. Mother tongue, French against every other answer. In 2022 the question asked for all the languages learned in childhood, and the 146 respondents who named more than one are left out of the language split, as are 115 respondents in 2014 whose answer could not be classed.
  • Sovereignty. The side a respondent would take in a referendum, among those who chose one. From 1998 to 2008 the studies asked the 1995 referendum question, and those who hesitated were asked which way they leaned; from 2012 on, the question is about an independent country. In 1998 only the CROP poll asked it. The Durand panels in the table views use their own wordings (see the Question column).
  • Weights. Each study’s post-election weight, where it has a validated one (the weights table gives each study’s); the others, 1998 and 2008 among the drawn studies, are unweighted and drawn as hollow dots. The 1998 respondents had first answered pre-election polls and were interviewed again after the election, and that second round over-selected undecided voters and those who would not say how they would vote.