French Politics

, Volume 14, Issue 2, pp 272–285 | Cite as

Untangling myths and facts: Who supported the Québec Charter of Values?

Data, Measures and Methods


In 2013–2014, the Parti Québécois’ (PQ) Québec Charter of Values was hotly debated in the province of Quebec and the rest of Canada. For some Canadians, it was seen as a calculated move from the PQ to win a majority in the next provincial election, highlighting the stereotypical Quebec citizen’s aversion towards immigrants. For others, it was an important policy that confirmed the secular status of Quebec. The media played a large role in depicting the average Charter supporter. The goal of this study is to empirically test the claims regarding the sociodemographic profile of the Charter’s supporters. Using opinion poll data, we demonstrate that most of these assumptions were false, and that the nature of the support evolved during the debate.


Quebec political behaviour contact theory secularism policy 


Between September 2013 and the end of the most recent Quebec election in April 2014, the opportunity to legislate secularism in the public sector took a prominent place in Quebec’s political debate. Commonly called the Charter of Values, the Bill 60 presented by the Parti Québecois aimed to banish religious signs for most civil servants. It also served to define what constitutes an accommodation in the Québec Charter of Human Rights and Freedoms, therefore setting out the conditions for granting such accommodations.

Complying with the restriction on wearing religious objects that overtly indicate a religious affiliation for public servants and educators led to a much polarized debate. Many questions were raised regarding the public approval rating of this particular aspect of Bill 60, and discussions about its political impact and the potential electorate profiles in favour or opposed to it also took place. The debate that followed the presentation of the Charter took place between those who considered it as a legitimate way to confirm the secular nature of the state and those who considered it to be anti-immigrants. Most media elites in the province were among the later and because of that, throughout the whole debate, they depicted voters who supported this Charter as being old, not very educated and living outside Montréal. In this article, we challenge this view of the support for the Charter.

We have two main goals. Our first aim is to empirically test how accurate was the depiction made of the supporters by the medias. Second, we challenge the idea that this profile remained stable throughout the debate and assess how it evolved within this period.

This article is divided into two main sections. First, we will discuss the political context and the content of the Charter in more details. In the second part, our data and analysis will be presented. Here we will identify Charter supporters and how their views have evolved during these 5 months. These results will be discussed and we will speculate on how these data will impact the Parti Québécois’ (PQ) electorate in the long term.

The PQ and the Charter of Values

Since the 1970s, Quebec’s party system includes, in addition to the Left/Right axis, a very strong independent/federalist axis. Two parties dominate this dimension: the PQ and the Quebec Liberal Party (QLP). The former’s main objective is to achieve Québec independence, while the latter advocates the status quo regarding this question (for more on this, see Richez and Bodet, 2012). As a corollary of their primary objective, the PQ felt entrusted with the task of protecting Quebec’s distinct francophone culture. In the past, they carried out this task without falling into extremism and by remaining a credible opposition or governing party. While some of the bills they voted on initially resulted in heated debates, some of these laws are now seen as the cornerstone of the protection of Quebec culture.

Despite their differences, both parties agreed for many years on the immigrant integration model the province should adopt. The policy of all Quebec government, QLP and PQ alike, has been invested in interculturalism; that is, newcomers to Quebec are invited to contribute to the evolution of a common identity while at the same time keeping their distinctiveness (Bouchard, 2012). Because interculturalism differs from assimilation, it implies the land of welcome has to accommodate newcomers on some of their religious beliefs. In 2006, a large debate emerged about what could be recognized as an accommodation and set out the conditions for granting such an accommodation. To respond to a public outcry created by a few accommodations that were considered to be unreasonable, a public inquiry named the Bouchard-Taylor commission was put in place by the Liberal government at the time. Their recommendations to settle the unease were never put in place, leaving the public’s discomfort unanswered.

Upon being elected in September 2012, the PQ (having a minority government) had difficulties demonstrating a coherent vision and often acted defensively for the first months of their term. As seen in Figure 1, according to CROP, the PQ was in the low 20 per cent of support in April, May and June 2013. This dip in the polls also concurs with the arrival of a new Liberal leader.
Figure 1

Support for parties from April 2013 to January 2014 (CROP).

In August 2013, less than a year after their election, the PQ presented the Quebec Charter of Values. The support for the policy was first measured in September 2013. As seen in Figure 2, the Charter of Values captured more support in the electorate than in the PQ at the beginning. In fact, nearly 50 per cent of Quebeckers were already in favour of its principles in September. In the months following, that support grew to nearly 60 per cent. Among Francophone voters, from whom the PQ expected more constituencies, the level of support was even higher. It was reasonable to think that this could help Pauline Marois gain back political momentum. Many aspects of the proposal regarding accommodations were consensual, even among parties. The main contentious aspects were mostly related to religious signs and the scope of limitation of rights.
Figure 2

Support for the Charter of Values from September 2013 to January 2014 (CROP).

Leaked to the press in August 2013, the Charter of Values was seen by some as part of a broader strategy to take control of the political agenda. In that regard, it is important to identify how the debate was initiated. Therefore:
  • 20 August 2013: The content of the Charter was published in the most widely circulated newspaper in Québec.

  • 10 September 2013: The Minister responsible for the policy officially presented the government’s propositions and launched an information campaign.

  • 1 October 2013: End of Web consultation (26 000 comments).

  • 7 November 2013: The Minister introduced the Charter affirming the values of State secularism and religious neutrality and of equality between men and women, and providing a framework for accommodation requests (Bill 60) at the National Assembly.

  • 14 January 2014: General consultation and public hearings on Bill 60 officially began at the National Assembly.

  • 20 February 2014: The last day of general consultation and public hearings on Bill 60.

  • 5 March 2014: Bill 60 was not adopted as the National Assembly dissolved.

The main elements from Bill 60 were as follows:
  1. 1

    ‘to specify, in the Charter of human rights and freedoms, that the fundamental rights and freedoms guaranteed by that Charter are to be exercised in a manner consistent with the values of equality between women and men and the primacy of the French language as well as the separation of religions and State and the religious neutrality and secular nature of the State, while making allowance for the emblematic and toponymic elements of Québec’s cultural heritage that testify to its history’.

  2. 2

    ‘public bodies must, in the pursuit of their mission, remain neutral in religious matters and reflect the secular nature of the State’.

  3. 3

    ‘obligations are set out for personnel members of public bodies in the exercise of their functions, including a duty to remain neutral and exercise reserve in religious matters by, among other things, complying with the restriction on wearing religious objects that overtly indicate a religious affiliation’.

  4. 4

    ‘personnel members of a public body must exercise their functions with their face uncovered, and persons to whom they provide services must also have their face uncovered when receiving such services’.

  5. 5

    to ‘defines, in the Charter of human rights and freedoms, what constitutes an accommodation resulting from the application of the Charter, and sets out the conditions for granting such an accommodation’.


In autumn 2013, the Charter of Values was an issue vastly debated in the media. From the Bill’s elements outlined above, item 3 was the most contentious. The press was mostly negative, while the opposition parties were divided on the subject. The provincial Liberals were firmly opposed to it. The Coalition Avenir Québec (CAQ), the third party at the National Assembly, was open for discussion and offered a compromise; their position was near the Bouchard-Taylor Commission report, which was to limit the restriction on wearing religious objects that overtly indicate a religious affiliation to people with an authority role. While some well-known figures did publicly support the Charter, like an ex Canada Supreme Court Justice, most of Quebec’s intelligentsia was against the policy. Despite the possible agreement with the CAQ, the government decided not to compromise on the issue. The parliamentary commission mandated to study the bill received more than 200 memoirs. The hearings were stopped when an election was called, making the Charter of Values one of the key issues in the election. Despite very high public support of the Charter, the leader of the PQ lost her electoral bid. In the following section, we will present the data and the analyses of the support for the Charter.

Who Supported the Charter?

A lot of media space was devoted to this issue during the debate, and most columnists stood against the proposition of the Marois government. However, most surveys showed that Quebec’s electorate was divided on the issue of the Charter of Values. With most of the opinion leaders being against, who could support this policy? In many papers, supporters of the Charter were depicted as minimally educated (Benoit, 2013; Boivin, 2013; Pelletier, 2014), elderly (David, 2014) and living outside of Montreal (Boivin, 2013; Corbeil, 2013; Guilbault, 2013). There was also a presumption that this supporter profile did not evolve throughout the 5 months of our analysis.

Is this portrait credible? The answer to this question depends on the point of view we adopt towards this policy. The first point of view, defended by the government, sees the Charter as an affirmation of secularism of the state and a way to not favour any religion over the other. The second point of view sees this policy as favouring the Catholic religion over other faiths, and as being discriminatory against immigrants.

The former offers no theoretical basis to this view, while the latter does. Literature indicates older and less educated people have a more negative attitude towards immigrants and immigration (O’Rourke, 2006). Less educated people also have a more negative attitude toward immigration and immigrants than their more educated counterpart (Espenshade and Hempstead, 1996; Scheve and Slaughter, 2001). The Montreal aspect of the profile is clearly related to the contact theory. The arguments raised indicated that voters from Montreal were against the Charter because they were in contact with immigrants and therefore knew they had nothing to fear from them. It is true that the contact theory partly supports the idea that social connections with immigrants improve the native citizen’s attitude towards them (Allport, 1954; Pettigrew, 1997; Hayes and Dowds, 2006). Nevertheless, this is a narrow interpretation of this theory because it also says that contact with immigrants induces positive effects only in certain settings. In fact, in some contexts, such contacts can have no effect or negative effects (Jackman and Crane, 1986; Taylor, 1998; Sibley et al, 2013). That being said, the question remains: Does this last point of view explain the support for the Charter of Values? The remainder of this section will be used to test these four hypotheses.

Hypothesis 1:

  • Older voters are more likely to support the Charter of Values than younger voters.

Hypothesis 2:

  • More educated voters are more likely to support the Charter of Values than less educated voters.

Hypothesis 3:

  • Voters residing outside Montreal are more likely to support the Charter of Values than residents of Montreal.

Hypothesis 4:

  • The profile of the supporters of the Charter did not change between September 2013 and January 2014.

These hypotheses were derived from the second point of view discussed above, since our goal is to empirically test the conceptions present in the francophone mass media. The analysis that will follow uses data from monthly surveys conducted between April 2013 and January 2014 (except July) by CROP, a well-established survey firm. Each sample includes 1000 respondents collected with a web panel and weighted to be representative of the population.1 We will use a logistic regression model with a dependant variable coded 1 if the respondent is for the Charter and 0 if he is against it. The model includes three variables of interest. Age will be split into three categories (18–34, 35–54, and 55 and more) and the respondents’ residence is a dummy variable coded 1 if living in Montreal, and 0 if residing outside of Montreal. Education varies by 12 categories that we scaled between 0 and 1. We also include socio-demographic control variables that are standard in Quebec’s electoral statistical models (Bastien et al, 2013).

These control variables include income (a 7-category variable scaled between 1 and 0), a dummy variable for language (0 for francophone and 1 for others), and a dummy variable for gender (coded 1 for female and 0 for male). We also included an interaction between age and education. This interaction was added because it is not clear if the effect is the same for highly educated and lower educated older individuals as the presumed effects for these two variables are different. An older individual could have a different attitude towards the Charter depending on his level of education.

The last variable worth mentioning is the vote intention. This last one is trickier because adding it in the model could induce an endogeneity problem. Indeed, it is not clear at this point if voters choose the PQ because of their positive opinion on the Charter or if they have a positive opinion of the Charter because they vote for the PQ. In fact, the voter opinion on this issue depends on his perception of the problem that the government tried to solve with this policy, and whether the policy does address the problem. The literature on electoral behaviour shows that partisanship induces biases when voters have to perceive and judge the political world (Bartels, 2002; Gaines et al, 2007). As this issue was highly polarized, it is likely that partisanship influenced the opinion on the Charter. Unfortunately, we do not have access to the partisan identification of the respondents, but because this variable is highly correlated to voter choice, we will use the voting intention as a proxy in our model. Table 1 presents the results for the models of September, October, November, December and January.
Table 1

Support to the Charter – Logistic model











































55 and more












35–54 × education












55 and more × education


































































R 2






*P<0.01; **P<0.05.

t statistics in parentheses.

As we can see, the only variable that holds the same effect over the five models is for the non-francophone individuals who are massively against the Charter. The education variable varies widely with time. The first three models do not show any relationship, but the last two show that the probability of voters to support the Charter decreases when voters are more educated. Therefore it is not clear if education really has an effect on the support for the Charter. As for age, there is a difference in all five models between the 18–34 year olds and one of the older groups. It is not always the same group in all models, but the 35–54 year old and the 55 year old and more age groups are significant in at least three of the five models each. Everything else being equal, there seems to be a significant relationship between support for the Charter and age. The results we get for the Montreal variable are surprising. All things being equal in four of the five models, there is a significant relationship between support for the Charter and residing in Montreal. However, the direction of the effect is not as expected, as these result shows voters from Montreal have a higher probability of supporting the Charter than voters living outside the city.

These first results show that the profile of the supporters of the Charter is not as straightforward as we could have expected. However, logistic coefficients cannot be used to interpret the magnitude of the effects. Moreover, interactions with this category of estimator are very difficult to interpret. It is difficult in this context to tell if the analysis supports or invalidates the hypothesis. To facilitate the interpretation of the results from Table 1, we will use the model we created to make several predictions of the dependant variable for different values of our variable of interest. For each set of predictions, only one variable of interest at a time will be examined, while all other variables will be constant. This way, it will be easier to assess the effect of each of these variables on Charter support while taking the interaction into account. To make these predictions, we will use the software Clarify (King et al, 2000; King et al, 2003). This software uses a Monte Carle simulation to simulate confidence bounds on the predictions of the model, making the results and the conclusions more reliable. For all the predictions that will follow, except for the variable of interest, all dummy variables will be set to zero, and all continuous variables set to their means. As for the age variable, it will be set to the reference category (18–34).

The first variable of interest we simulated was age. Sets of predictions were simulated for each age category, with all the other variables constant. Figure 3 shows the effect of age on the predictions of the probability of supporting the Charter.
Figure 3

Prediction simulations (Clarify) – Age.

The September predictions show the voters in the 35–54-years-old category have a higher probability to support the Charter than voters from the 18–34-years-old category. There is no significant difference in the oldest category of voter. In October, voters aged between 18 and 34 had a higher probability of supporting the Charter than every other age group. In November, we observe the reverse effect: younger voters had a lower probability of supporting the Charter than their older counterparts. In the past 2 months, there is no significant change in the probability of supporting the Charter between the voters from the different age categories. It is difficult to point out a specific pattern in the first 3 months as the results show such variability. The pattern stabilizes in December and January; however, there is clearly an evolution in the profile of the supporter of the Charter.

The second variable we will simulate predictions with is education. As this variable is continuous and bounded between 0.1 and 1, we will simulate prediction with three different values of education (0.1, 0.5 and 1), while keeping every other variable constant. The results from the simulations are presented in Figure 4.
Figure 4

Prediction simulations (Clarify) – Education.

Unlike the expectations we had for this variable, the predictions for the first 2 months show that the more educated voters have higher probability of supporting the Charter than the less educated ones. The effect wears off in November. In December, the predictions support the hypothesis and the higher educated voters had less chance than their lower educated counterparts. In January, however, the effect wears off again. As for the age variable, there seems to be differences in the support for the Charter between the first and last month.

The last variable of interest we will simulate is the place of residence. Predictions will be simulated for voters from Montreal and outside Montreal, again, keeping all other variables constant. The results from this simulation are presented in Figure 5.
Figure 5

Prediction simulations (Clarify) – Montreal.

Results from Figure 6 show voters from Montreal have higher probability of supporting the Charter in September and January, while there are no significant differences in the months between. Even if the coefficient was significant in the results shown in Table 1 for four of the 5 months, the magnitude of the effect was not strong enough to overcome the error margins of the simulation. It would therefore be tricky to say voters from Montreal were more likely to support the Charter, but they definitely do not support it less.
Figure 6

Prediction simulations (Clarify) – Profiles.

Finally, a last set of predictions will be done using two profiles of voters. The voter ‘A’ will be from outside Montreal, lower educated and old. The profile ‘B’ will be from Montreal, highly educated and young. Based on our hypothesis, we expect the voter ‘A’ to have higher probability of supporting the Charter than the profile ‘B’. Results from these simulations are presented in Figure 6.

Again, these results show clearly an evolution in the support for the Charter over this period of 5 months. In September, voter ‘B’ has a significantly higher probability of supporting the Charter than voter ‘A’. There is no significant difference in October. In November and December, voter ‘B’ has significantly lower chances of supporter the Charter. Finally, in January, there is no significant difference between the predictions for the two profiles.


In this article, we seek to validate four hypotheses that could help untangle the myths and facts related to the support for the Charter of Values.

Hypothesis 1:

  • Older voters are more likely to support the Charter of Values than younger voters.

The results from the raw coefficient in Table 1 indicated that this variable had a significant impact on the support for the Charter. However, the prediction simulations show a slightly different picture. Only two of the 5 months ended up showing significant difference between age groups that support the hypothesis. Furthermore, only one month (November) showed a clear cleavage between younger and older voters. Even if the coefficients were significant, the magnitude of this effect was too weak to overcome the effect of other variables, explaining why many simulations ended up showing no difference. As there is no clear effect, it is therefore impossible to reject the null hypothesis in this case.

Hypothesis 2:

  • More educated voters are more likely to support the Charter of Values than less educated voters.

The answer to the question of the link between education and the Charter varied according to the time frame covered in our analysis. In the first months following the presentation of the Charter, more educated voters were more likely to support the Charter than lower educated voters. In the following months, the effect disappeared. Again, no clear pattern emerged from any of the models, making it impossible to reject the null hypothesis.

Hypothesis 3:

  • Voters residing outside Montreal are more likely to support the Charter of Values than residents of Montreal.

If the first two hypotheses received somewhat mixed answers, this one has a clear answer. The coefficient presented in Table 1 showed clearly that living in Montreal did not decrease voter probability of supporting the Charter, but actually increased it. The results from the simulations show the magnitude of the effects were not that strong. Nevertheless, these results clearly invalidate this hypothesis. The lower support received for Charter in Montreal cannot be attributed to living there.

Hypothesis 4:

  • The profile of the supporters of the Charter did not change between September 2013 and January 2014.

The results from the first three simulations show that there is a difference in support for the Charter from the first two months to the following months. However, the last simulation using the two voters’ profile makes it even clearer. The cut point after the first two months is even more visible. It is therefore impossible to reject the null hypothesis.

This last simulation could tell us something about the points of view adopted towards the Charter during these 5 months. If observed individually, the variables of interest did not demonstrate clear patterns. However, the simulation using the multivariate profiles showed more clearly that these variables did had an effect. Based on the profile we created, the results from the first two months indicate that the Charter was seen by many as a tool to secure the secular nature of the state. In the following months, it changed in favour of a vision more oriented towards immigrants.

Even in this last simulation, in two of the five months, the difference is not significant. It shows that even if these variables have an effect, other variables are even more explicative of Charter support. The results from Table 1 show that this issue was mainly defined by two cleavages, language and vote intention. These two variables are powerful predictors of support for the Charter. The language cleavage explains the lower support the Charter received in Montreal, as the proportion of francophones is lower there compared to the rest of the province.

This language cleavage could have been a good strategy to mobilize the francophone electorate. But the Charter did not have the effect expected by Pauline Marois and her strategists. As a result, PQ’s branding took a hard hit on election night and the challenge to appeal new generations into their ranks became a reality. However, this does not mean that the identity issue that the Charter tried to solve is over. If the liberals do not act, another crisis could pop up at any time. As for the charter’s supporters profile depicted during the debate, our analysis showed it was not as straighforward as it was presented. Moreover, our analysis demonstrated that it changed at some point around November. This reminds us that political debates related to identity issues can be very passionate and can lead to false assumptions.


  1. 1.

    The specific questions for each of the variables are available in Appendix A.



The authors would like to thank CROP for the data, and the Fonds de recherche du Québec - Société et culture for its financial support.


  1. Allport, G. (1954) The Nature of Prejudice. New York: Anchor.Google Scholar
  2. Bartels, L.M. (2002) Beyond the running tally: Partisan bias in political perceptions. Political Behavior 24 (2): 117–150.CrossRefGoogle Scholar
  3. Bastien, F., Bélanger, É. and Gélineau, F. (dir) (2013) Les québécois aux urnes. Montréal: Les Presses de l’Université de Montréal.Google Scholar
  4. Benoit, S. (2013) La stratégie de la peur. La Presse 24 September: A23.Google Scholar
  5. Boivin, S. (2013) Des gains nets pour la Charte. Le Soleil 12 November: 3.Google Scholar
  6. Bouchard, G. (2012) L’interculturalisme. Montréal: Boréal.Google Scholar
  7. Corbeil, M. (2013) La charte et le PQ comparés au Tea Party. Le Soleil 13 November: 16.Google Scholar
  8. David, M. (2014) Les vieux péquistes. Le Devoir 23 January: A3.Google Scholar
  9. Espenshade, T.J. and Hempstead, K. (1996) Contemporary American attitudes toward U.S. immigration. International Migration Review 30 (2): 535–570.CrossRefGoogle Scholar
  10. Gaines, B.J., Kuklinski, J.H., Quirk, P.J., Peyton, B. and Verkuilen, J. (2007) Same facts, different interpretations: Partisan motivation and opinion on Iraq. The Journal of Politics 69 (4): 957–974.CrossRefGoogle Scholar
  11. Guilbault, J. (2013) Lettre d’un souverainiste mou au PQ. La Presse 23 October: A29.Google Scholar
  12. Hayes, B.C. and Dowds, L. (2006) Social contact, cultural marginality or economic self-interest? Attitudes towards immigrants in Northern Ireland. Journal of Ethnics and Migration Studies 32 (3): 455–476.CrossRefGoogle Scholar
  13. Jackman, M.R. and Crane, M. (1986) Some of my best friends are black … : Interracial friendship and whites’ racial attitudes. Public Opinion Quarterly 50 (4): 459–486.CrossRefGoogle Scholar
  14. King, G., Tomz, M. and Wittenberg, J. (2000) Making the most of statistical analyses: Improving interpretation and presentation. American Journal of Political Science 44 (2): 347–361.CrossRefGoogle Scholar
  15. King, G., Tomz, M. and Wittenberg, J. (2003) CLARIFY: Software for interpreting and presenting statistical results. 2.1 2.1. Stanford University, University of Wisconsin, and Harvard University. 5 January. Available at
  16. O’Rourke, K.H. (2006) The determinants of individual attitudes towards immigration. European Journal of Political Economy 22 (4): 838–861.CrossRefGoogle Scholar
  17. Pelletier, F. (2014) Appeler un chat un chien. Le Devoir 15 January: A7.Google Scholar
  18. Pettigrew, T.F. (1997) Generalized intergroup contact effects on prejudice. Personality and Social Psychology Bulletin 23 (2): 173–185.CrossRefGoogle Scholar
  19. Richez, E. and Bodet, A.M. (2012) Fear and disappointment: Explaining the persistence of support for Quebec secession. Journal of Elections, Public Opinions and Parties 22 (1): 77–93.CrossRefGoogle Scholar
  20. Scheve, K.F. and Slaughter, M.J. (2001) Labor market competition and individual preferences over immigration policy. Review of Economics and Statistics 83 (1): 133–145.CrossRefGoogle Scholar
  21. Sibley, C.G. et al (2013) A dual process model of attitudes towards immigration: Person x residential area effects in a national sample. Political Psychology 34 (4): 553–574.CrossRefGoogle Scholar
  22. Taylor, M.C. (1998) How white attitudes vary with the racial composition of local populations: Numbers count. American Sociological Review 63 (4): 512–535.CrossRefGoogle Scholar

Copyright information

© Palgrave Macmillan, a division of Macmillan Publishers Ltd 2016

Authors and Affiliations

  1. 1.Department of Political ScienceFaculty of Social Sciences, Université LavalQuébecCanada

Personalised recommendations