With rhyme and reason: Recognizing reasons for disliked practices increases tolerance.

Disapproval of others' beliefs and practices is an inevitable consequence of living with diversity, and the ability to tolerate, or put up with, these differences is crucial to maintain a functional society. Considering reasons to condone what one disapproves of is considered a key aspect of tolerance. Across three national samples (N = 1,708), the current research examines how recognizing arguments to support practices that one disapproves of increases tolerance. Studies 1-2 demonstrate that when participants generate arguments to support Muslim minority practices (Study 1) and Orthodox Protestant minority practices (Study 2), they disapprove of, they show increased tolerance towards such practices in society. In Study 3, the importance of considerations is experimentally extended by demonstrating that perceiving objectionable behaviour as more reasonable increases tolerance. Collectively, these studies demonstrate the importance of engaging in and perceiving reasonable considerations to enhance tolerance of dissenting beliefs or practices.

Background

Disapproval, dislike, and disagreement are inevitable aspects of everyday life, especially in open and diverse societies. People differ across many dimensions and there can be genuine differences in practices, beliefs, and worldviews that are impossible to reconcile as people are committed to their convictions and moral beliefs (Skitka & Morgan, 2014). However, this apparent conflict between competing beliefs and values in society does not have to mean that one wants or tries to negatively interfere in others’ lives; people are capable of tolerating that what they object to, allowing a diverse society to coexist (Verkuyten, Yogeeswaran, & Adelman, 2020). Tolerance implies voluntary forbearance and putting up with differences one disapproves of in a situation in which one (think one) can interfere (Cohen, 2004; King, 2012; Verkuyten & Yogeeswaran, 2017). It is not tolerance if one complies with, is forced or afraid to, or is not in the position or able to act against things one objects to. Rather tolerance involves having relevant reasons for not interfering with, for example, other cultural, religious, and ideological beliefs and different modes of conduct. In such cases, one’s dislike or disapproval of practices or beliefs (e.g., ritual slaughter of animals that is disapproved of by some people in Western Europe) is overpowered by considerations to nonetheless allow those practices or beliefs in society (e.g., religious freedom).

Empirically, the role of thinking about and recognizing reasons for stimulating tolerance has been underexplored. Although some research suggests that considering possible societal consequences might lead to reduced tolerance (Kuklinski, Riggle, Ottati, Schwarz, & Wyer, 1991), others argue that considered thought leads to increased tolerance (Sniderman, Tetlock, Glaser, Green, & Hout, 1989), and can override feelings of dislike (Pennycook, Fugelsang, & Koehler, 2015). Considering relevant reasons for accepting dissenting beliefs or practices is seen as a core aspect of tolerance: ‘tolerance is something we must do for the right reasons. The presence of those reasons matters. … one tolerates what one (believes one) should’ (Cohen, 2004, p. 72; Forst, 2013). This process of recognizing relevant reasons to tolerate something we disapprove of not only involves reasons that people themselves have for being tolerant but also the perceived reasons that others have for their actions.

Using national samples of Dutch majority group members, the current research investigates the critical theoretical assumption within the tolerance literature that recognizing relevant arguments for supporting something we disapprove of increases tolerance of perceived controversial practices. In Study 1, we examine the central prediction that self‐reported reasons for accepting negatively evaluated Muslim minority practices increases tolerance of these practices. Study 2 examines the same prediction with a native religious minority group of orthodox Protestants. In both studies, we consider a range of religious minority practices that personally impact on one’s lives versus those that have a wider impact on society (Adelman, Verkuyten, & Yogeeswaran, 2021). In Study 3, we focus on the perceived reasons of others rather than self‐reported reasons for further testing the central hypothesis. Additionally, in this study, we did not focus on individual differences in considering relevant reasons, but rather used an experimental design to examine whether people are more tolerant of anti‐refugee protest actions depending on whether the protestor’s explanation for their action is considered relevant.

Thus, the key prediction that the recognition of relevant reasons leads to higher tolerance is examined by focusing on an immigrant‐origin and established religious minority group, a range of minority practices and in two contexts, different measures of tolerance, and by focusing on self‐reported and perceived reasons. In this way, the three studies aim to provide a conceptual replication (Crandall & Sherman, 2016; Stroebe, 2019) of the predicted process that will enhance our confidence in the central theoretical proposition of the tolerance process.

The role of reasons in tolerance

In philosophy and the social sciences, there are various approaches and understandings of rationality. However, what most conceptualizations of rationality have in common is the notion of reason: ‘for a belief to be rational is for it to be based on reasons’ (Crane, 2017, p. 147). A belief or action is rational when there are conventions, principles, or values which count in favour of believing or doing it (Scanlon, 2004). People appeal to these reasons when making arguments about what they themselves or others (should) believe or do. Reasons and arguments can be considered good or bad in the sense of being perceived to be a relevant or not relevant consideration in favour of believing or doing something. For example, treating categories of people (Muslims, sexual minorities) differently on the basis of reasons that are considered irrelevant (i.e., prejudicial discrimination) is something else than treating categories of people (e.g., the elderly, the sick) differently for relevant reasons (i.e., differential treatment). Good (vs. bad) reasons can be traced to some understandable or normatively acceptable source, and rationality refers to the process of employing these reasons for forming one’s beliefs. Actions and beliefs that have an unacceptable normative basis are considered unreasonable and irrational, whereas actions and beliefs that are perceived as socially acceptable and having value, are considered reasonable and rational. Obviously, what are considered good reasons can differ between individuals, groups, and situations, but there is a strong first‐person dimension in that the person themselves must believe that there is an acceptable and relevant basis for a particular belief or course of action.

Having relevant reasons for accepting beliefs or practices that one disapproves of is a built‐in component of tolerance: ‘tolerance is only a good thing if it is justified properly’ (Cohen, 2004; Forst, 2013, p. 43). These self‐endorsed normative reasons make tolerance similar to the internal motivation to respond without prejudice, and different from the externally motivated suppression of prejudice based on threat, fear, or the desire to appear unprejudiced in the eyes of others (Verkuyten et al., 2020). People can be expected to be more tolerant if they think that there are relevant reasons to accept what they continue to disapprove of. On the one hand, there is what one thinks is false or wrong, but on the other hand, one must be able and willing to allow others to live the life they want. There need to be relevant reasons to endure the objectionable behaviours or beliefs that trump one’s disapproval. Psychologically, tolerance involves balancing one’s dislike or disapproval with recognizing reasons for forbearance whereby the latter overrides the former (Verkuyten et al., 2020). It is this key prediction that we, for the first time, put to an empirical test in three studies that together explore the role of recognizing reasons for tolerating disapproved of conduct.

In Studies 1 and 2, we examine whether thinking about additional reasons for supporting disapproved minority conduct leads to higher tolerance of that conduct. Study 1 examines this by focusing on Muslim minority practices,1 while Study 2 focuses on religious minority practices of Orthodox Protestants. Muslims and Orthodox Protestants are two numerically small religious groups in the Netherlands (both representing around 5% of the Dutch population). Both struggle to hold on to their religious values and beliefs in the increasingly secular context of the country (Fetzer & Soper, 2003; Ribberink, Achterberg, & Houtman, 2017). The focus on these two groups allows us to assess whether recognizing additional reasons matters for the tolerance of practices of an immigrant‐origin group (Muslims) that often is perceived as threatening national culture and identity (Sniderman & Hagendoorn, 2007), similarly applies to a historically established religious minority group (i.e., Orthodox Protestants) aligned with the majority population.

In Study 3, we use an experimental design for testing whether the perceived rationality of a controversial protest action affects tolerance. People might be more tolerant if they think that relevant reasons are provided for engaging in a disapproved of practice. For example, people may find it more difficult to recognize the rationality of racist bigotry and anti‐immigrant hatred than of procedural justice concerns in relation to affirmative action (Bobocel et al., 1998) or communitarian concerns about continued immigration (Sniderman & Hagendoorn, 2007). People may think that minority favouring policies and immigration are good things, but still recognize that others have understandable normative reasons for being critical and skeptical, which is likely to make them more tolerant of such disagreements. The prejudicial, or irrational, beliefs and actions of a bigot or xenophobe, however, are unlikely to induce greater tolerance. This expected difference is tested in Study 3.

STUDY 1

Study 1 examined whether the self‐recognition of relevant arguments impacts the process of weighing the disapproval against reasons to nonetheless tolerate. Specifically, we aim to examine the importance of thinking about reasons for tolerance by testing the prediction that individuals will be more tolerant when they themselves recognize that there are relevant arguments for supporting practices and events that they are negative about (Verkuyten et al., 2020). Based on ongoing societal debates in the Netherlands, we used different types of Muslim minority practices in this study. This allows us to examine whether the expected role of the recognition of additional reasons for tolerance generalizes across a range of different practices. In Dutch society, some Muslim minority practices are considered by the public more controversial than others and tolerating some of these practices might require more adaptations, whereas other practices might be accommodated relatively easily (Adelman & Verkuyten, 2020). Additionally, we examined the role of relevant reasons for tolerance if such practices personally impact participant’s own lives, or rather have broader societal relevance. People tend to be less tolerant in personalized than societal contexts (Adelman et al., 2021; Capelos & Van Troost, 2012; Chanley, 1994), but that does not necessarily mean that the role of recognizing reasons for tolerance is different across these. Here, we focus on this latter aspect and we do not compare the degree of tolerance between the two contexts since the specific practices in both contexts are not similar. Focusing on a range of Muslim minority practices and across these two contexts allows us to investigate whether the impact of relevant reasons on tolerance replicates conceptually.

Method

Participants

Participants were recruited as part of a survey of ethnic majority Dutch adults collected through the survey company GFK that has a large panel that is representative by gender, age, education, and district of residence in the Netherlands. GFK approached 1,700 panel members and the response rate was 50%, which is common in the Netherlands (Stoop, 2005). This resulted in a sample of 851 participants, which matches the demographic characteristics of the Dutch population (Statistics Netherlands, 2019). The study reported here was embedded within a survey about immigration, diversity, and societal changes. As is common in these national surveys, different researchers were involved and different versions of the questionnaire were used. The questions related to tolerance were presented to a random subsample of 174 participants, ranging in age from 18 to 81 years (M = 46.63, SD = 15.31), 50.6% female, of whom 10.9% had low‐level education, 50.6% had mid‐level education, and 38.5% had high‐level education.

Material

Based on previous research (Adelman et al., 2021; Hirsch, Verkuyten, & Yogeeswaran, 2019) and in order to make the scenarios realistic and ecologically valid, we used eight vignettes about specific Muslim minority practices that have caused public debate in Dutch society. However, in order not to overburden participants, we used a between subjects design: two random subgroups of participants that each were presented with four vignettes in a randomized order. Participants were introduced to these practices with a brief paragraph: ‘The following are a number of specific events or situations about which we like to ask a few questions. We want to ask you to connect as much as possible with the events or situations so that it is as realistic as possible for you’. Next, participants read another brief paragraph introducing them to one of four specific practices or events. For one subgroup, the scenarios had an (imagined) personal impact and for another subgroup the scenarios were about broader societal implications.

For the former subgroup, the scenarios were: ‘Imagine that at the daycare center where your son or daughter goes, the children eat hot food in the afternoon. Because a few children are Muslims, the daycare center considers giving children also halal food (no pork and slaughtered by an Islamic butcher)’, ‘Imagine that at the secondary schools of your children, it is considered, at the request of Muslim parents, to give swimming lessons separately to boys and girls’, ‘Imagine that in the area where you live there are plans to build a mosque. There is a building application with the municipality. If construction continues, the mosque will be located in your neighbourhood’, and ‘Imagine that at work you are being suggested by an Islamic colleague to set up a separate prayer room. Your colleague wants to use this several times a day and there is no such space now’.

The other subgroup was presented with four scenarios that all referred to issues that are debated in Dutch society: ‘Some mosque organizations set up Islamic primary schools in the Netherlands that only Muslim children attend’, ‘Some Muslim organizations want Dutch hospitals to have an arrangement that (if possible) men are treated by men and women by women’, ‘Most Muslims find it important that animals are slaughtered ritually’, and ‘Recently there was a proposal to allow female police officers to wear a headscarf during their work’.

Following the paragraph introducing the specific scenarios, participants were first asked whether they either had a positive or a negative attitude towards the practice or event. A binary scale was used because it serves our purposes in focusing on the evaluative direction of a response without being confounded with intensity, and such a scale is easier to process by respondents and provides a stable and reliable evaluative difference measure (e.g., Dolnicar, Grün, & Leisch, 2011; Dolnicar, & Leisch, 2012). Depending on their answer, they were subsequently asked ‘Despite your positive feelings: do you think that there are also reasonable arguments to not support [the practice]’, or ‘Despite your negative feelings: do you think that there are also reasonable arguments to support [the practice].’ The open‐ended answers categories were ‘Yes, because …’ and ‘No, because …..’ Most participants wrote down short answers or keywords, like ‘freedom of religion’ and for each practice these answers were broadly categorized following the distinction that scholars (Forst, 2013) as well as lay people (Velthuis, Verkuyten, & Smeekes, 2021) make between tolerating on principled (‘freedoms’) or pragmatic (‘living together’) grounds.

Finally, participants were presented with a behavioural intention question measuring tolerance, ‘Suppose you are asked to sign a petition against [the practice]. How likely is it that you would do that? (7‐point scale, reversed scored: ‘Certainly sign it’ to ‘certainly not sign it’).2 This type of measure has been used in previous research (e.g., Adelman et al., 2021; Hirsch et al., 2019; Sleijpen, Verkuyten, & Adelman, 2020) and corresponds with the conceptualization of (in)tolerance as related to being in a position to interfere with beliefs and practices that one disapproves of (Cohen, 2004).

Results

Attitudes towards the practices

As Table 1 shows, the majority of participants were negative towards the practices (59.3% up to 97.7%), except for the separate praying room at work scenario (40.7%). Further, only 20 participants were negative towards all four practices related to the personal domain and 48 participants were negative towards all four societal practices. These descriptive findings demonstrates that the nature of the practice matters, but also that all scenarios raise questions of toleration.

Table 1

Frequencies and row percentages (in brackets) for ‘positive’ or ‘negative’ attitudes towards Muslim practices, and reasonable arguments (yes‐no) to support or not support the practices

	Attitude	Despite own feeling are there reasons to support/ not support the practice? a		Total
		Yes	No
Halal food for all children at day‐care	Negative	22 (43.1%)	29 (56.9%)	51 (59.3%)
	Positive	21 (60.0%)	14 (40.0%)	35 (40.7%)
Separate swimming lessons for boys and girls at school	Negative	12 (14.3%)	72 (85.7%)	84 (97.7%)
	Positive	0 (0%)	2 (100%)	2 (2.3%)
Mosque in neighbourhood	Negative	22 (43.1%)	29 (56.9%)	51 (59.3%)
	Positive	14 (40.0%)	21 (60.0%)	35 (40.7%)
Room to pray for colleague at work	Negative	9 (25.7%)	26 (74.3%)	35 (40.7%)
	Positive	26 (51.0%)	25 (49.0%)	51 (59.3%)
Islamic schools	Negative	25 (30.9%)	56 (69.1%)	81 (92%)
	Positive	4 (57.1%)	3 (42.9%)	7 (8.0%)
Regulation in hospitals that (when possible) people are treated by someone of the same gender	Negative	19 (27.5%)	50 (72.5%)	69 (78.4%)
	Positive	12 (63.2)	7 (36.8%)	19 (21.6%)
Ritual slaughter of animals	Negative	17 (25.0%)	51 (75.0%)	68 (77.3%)
	Positive	6 (30.0%)	14 (70.0%)	20 (22.7%)
Allowing headscarf’s for female police agents	Negative	20 (27.4)	53 (72.6%)	73 (83.0%)
	Positive	5 (33.3%)	10 (66.7%)	15 (17.0%)

Participants were asked if there were reasonable arguments to support the practice, when they had answered that their own feelings were ‘negative’ towards the practice, and if there were reasonable arguments to not support the practice, if they had indicated that their own feelings were ‘positive’.

Table 1 shows that the majority of those with a negative attitude did not think that there were arguments for support (56.9–85.7%). At the same time there was a substantial group of participants who, despite their negative feelings, did recognize that there were additional reasons for support (14.3–43.1%). In their open‐ended answers and depending on the particular practice, these participants mentioned principled reasons of freedom of religion, freedom of education, and personal choice, as well as pragmatic reasons of accommodating others and being able to live together peacefully.

Tolerance

Our central expectation is that considering reasons for supporting what one disapproves of leads to higher tolerance. We first examined for the whole sample the overall level of tolerance (α = .73 for personal scenarios, α = .82 for societal scenarios) by conducting a two‐way analysis of variance with attitude (positive vs. negative) and recognition of a counterargument (yes vs. no) as factors. This resulted in a main effect for attitude, F(1, 173) = 42.18, p < .001,  = .199, and a significant interaction effect between attitude and counterargument, F(1, 173) = 6.84, p = .01,  = .039. The interaction effect indicated that recognizing a counterargument mattered for tolerance only among those with a negative attitude. For these participants and in line with the expectation, tolerance was higher among those who recognized (vs. did not recognize) reasons for supporting what they disapproved of (M = 3.23, SD = 1.23; vs. M = 2.61, SD = 1.18), F(1, 107) = 6.64, p = .019,  = .051. Participants with a positive attitude were overall more accepting (M = 3.94, SD = 1.04, and M = 4.28, SD = 0.90), F(1, 66) = 2.05, p = .16,  = .031.

Tolerance implies disapproval and the disapproval differs considerably between the various practices (Table 1). Therefore, we focused further on the majority of the participants that had a negative attitude towards a particular practice. Participants who recognized reasons for supporting the negative evaluated practice were more tolerant than those who did not think that there are reasons to support what they disapprove of. For the practices that have an impact on one’s personal life (see Figure 1), this was found for ‘halal food at school’, F(1, 49) = 5.64, p = .022,  = .10, ‘Mosque in neighbourhood’, F(1, 49) = 9.73, p = .003,  = .17, and ‘praying room at work’, F(1, 33) = 8.23, p = .007,  = .20, but not for ‘separate swimming lessons’ F(1, 82) = 1.28, p = .262,  = .02, which was overwhelmingly rejected (Table 1).

Figure 1

Tolerance of Muslim practices for participants who are negative towards the practice by reasonable counterargument ‘yes’ or ‘no’. A higher score indicates higher tolerance.

For the practices with a societal impact, tolerance was higher for participants who, despite their negative feelings, recognized reasonable arguments compared to participants who did not think that there were additional reasons (see Figure 1). This was true for all 4 scenarios including for ‘Islamic schools’, F(1, 79) = 17.01, p < .001,  = .18, for ‘hospital gender segregation’, F(1, 67) = 6.79, p = .011,  = .09, for ‘ritual slaughter’, F(1, 66) = 12.04, p = .001,  = .15, and for ‘headscarf female police agent’, F(1, 71) = 11.32, p = .001,  = .14.

Sensitivity power analysis

We conducted sensitivity power analyses for the inferential main effect of tolerance among those with negative attitudes (n = 35–81), accounting for unequal group size. The sensitivity power analysis (t‐test) revealed that, at a desired power of .80 and α = .05, the samples in Study 1 achieved sensitivity to detect at least between medium to large effects ( ≥ .1042–.2375).

Discussion

The findings of the first study provide evidence for the role of recognizing arguments for the process of toleration. Overall, people tend to be more tolerant of practices they disapprove of when they also recognize that there are relevant arguments for supporting that practice. Furthermore, the recognition of arguments opposing one’s attitude did not matter for the overall tolerance among participants with a positive attitude, which suggests that recognition per se does not weaken the relation between attitude and behavioural intention. Additionally, the effect on tolerance among participants with a negative attitude was found for a range of practices that differ in their personal and societal impact and in the degree of required majority group accommodation.3 Some of these practices might even be considered as being too demanding and beyond reasonable accommodation (Verkuyten & Yogeeswaran, 2017). Importantly, however, the range of practices allowed us to determine that the role of recognizing reasons for tolerance was similar across less and more demanding practices and across practices that have an impact on one’s personal life and on the broader society.

STUDY 2

The first study examined people’s tolerance of an immigrant‐origin group in the Netherlands (i.e., Muslims), which raises the question whether the findings are specific to a non‐native target group. Therefore, in Study 2, we further examined the importance of recognizing reasons for tolerance in relation to perceived controversial practices and events of Orthodox Protestants (Sleijpen et al., 2020) which is a native, but similarly small religious minority group as Muslims in the Netherlands (Fetzer & Soper, 2003; Ribberink et al., 2017). In Study 2, we also used a somewhat different measure to assess tolerance in order to examine whether the importance of recognizing additional reasons generalizes to another outcome measure (Adelman et al., 2021). Furthermore, because of the relatively low sensitivity of the small sample size in Study 1, we used a much larger sample in Study 2.

A nationally representative sample of 876 ethnic majority Dutch participants were recruited through the survey company GFK to participate in the study. The response rate was similar to Study 1 and all these participants completed the questions of interest. The participants’ ages ranged from 18 to 85 years (M = 47.04, SD = 15.73) with 51.7% female, and 12.7% having low‐level education, 51.1% mid‐level education, and 36.2% high‐level education and these demographics do again match the characteristics of the Dutch population (Statistics Netherlands, 2019).

The design and manipulation of the study were identical to those in Study 1, with the one difference being that the two‐sets of three scenarios were about Orthodox Protestant practices. For reasons of ecological validity, these scenarios were based on previous research and recent debates in the Netherlands (Sleijpen et al., 2020). For one subsample of participants the three scenarios that were presented in a randomized order were: ‘Recently there was a discussion about Orthodox Protestant schools refusing to hire homosexual teachers’, ‘Last year there was a discussion about Orthodox Protestant parents who refused to vaccinate their children against measles and polio’, and ‘Recently there was a discussion about Orthodox Protestant organizations that refuse women on their board’. For the other subsample of participants, the three scenarios were, ‘Recently there was a debate about an Orthodox Protestant pastor who wanted to give a speech at a university in which he equates abortion with murder’, ‘Recently there was a discussion about Orthodox Protestant schools that refuse admission of children from other faiths’, and ‘Last year Orthodox Protestants published the so‐called Nashville declaration in which people with a sexually “different” way of life are described as sinners who need to change’.

Similar to Study 1, participants were first asked to indicate on a binary scale whether they felt more positive or more negative towards the practice or event. Depending on their answer, they again were subsequently asked to indicate either if there are arguments to not to support [the practice] or to support [the practice]. Finally, participants were presented with a behavioural intention question measuring tolerance, ‘If you were in charge, would you tolerate [the practice]’ (7‐point scale: ‘Certainly not tolerate it’ to ‘certainly tolerate it’). This measure has been used in previous research (Gieling, Thijs, & Verkuyten, 2012; Sleijpen et al., 2020) and again reflects that (in)tolerance involves being in a situation to interfere with beliefs and practices that one disapproves of (Cohen, 2004).

Table 2 shows that the great majority of participants were negative towards the practices (84.2% up to 95.4%; 44% was negative towards all practices). This indicates that all scenarios are relevant toleration cases for most of the participants. Table 2 shows that the majority of those with a negative attitude did not think that there were additional reasons for supporting it. However, there is a substantial minority of participants that despite their negative feelings did recognize that there were also relevant reasons for support (11.3–35.9%). In their open‐ended answers, these participants mentioned predominantly reasons of religious freedom, freedom of education, and freedom of expression, or the importance of minority identity maintenance, and minority rights.

Table 2

Frequencies and row percentages (in brackets) of ‘positive’ or ‘negative’ attitudes towards Orthodox Protestant practices, and reasonable arguments (yes‐no) to support or not the practice

	Attitude	Despite own feeling are there reason to support/not support the practice? a		Total
		Yes	No
Refusing to hire homosexual teachers	Negative	65 (16.3%)	334 (83.7%)	399 (90.9%)
	Positive	20 (50%)	20 (50%)	40 (9.1%)
Not vaccinating children	Negative	70 (17.2%)	338 (82.8%)	408 (92.9%)
	Positive	17 (54.8%)	14 (45.2%)	31 (7.1%)
Not accepting women in management position of association	Negative	56 (13.4%)	361 (86.6%)	417 (95.0%)
	Positive	9 (40.9%)	13 (59.1%)	22 (5.0%)
Speech in which abortion is equated to murder	Negative	138 (35.9%)	246 (64.1%)	384 (87.9%)
	Positive	23 (43.3%)	30 (56.6%)	53 (12.1%)
Refusing children of different religions at Orthodox school	Negative	97 (26.4%)	271 (73.6%)	384 (84.2%)
	Positive	32 (46.4%)	37 (53.6%)	69 (15.8%)
Nashville declaration	Negative	47 (11.3%)	370 (88.7%)	417 (95.4%)
	Positive	7 (35.0%)	13 (65.0%)	20 (4.6%)

Participants were asked if there were reasonable arguments to support the practice, when they had answered that their own feelings were ‘negative’ towards the practice, and if there were reasonable arguments to not support the practice, if they had indicated that their own feelings were ‘positive’.

For testing our central prediction that recognizing arguments for disapproved of practices is associated with tolerance, we again first examined for the whole sample the overall level of tolerance (α = .69 for first set, α = .65 for second set) by conducting a two‐way analysis of variance with attitude (positive vs. negative) and self‐recognized counterargument (yes vs. no) as factors. This resulted in a main effect for attitude, F(1, 875) = 83.42, p < .001,  = .087, and again a significant interaction effect between attitude and counterargument, F(1, 875) = 20.96, p = .01,  = .023. Similar to Study 1, the interaction effect indicated that recognizing a counterargument mattered for tolerance only among those with a negative attitude. For them and in line with the expectation, tolerance was higher among those who recognized (vs. not recognized) reasons for supporting what they disapproved of (M = 3.59, SD = 1.65; vs. M = 2.03, SD = 1.11), F(1, 833) = 127.53, p < .001,  = .133. Participants with a positive attitude were overall more accepting (M = 4.52, SD = 1.69, and M = 4.81, SD = 1.63), F(1, 42) = 0.29, p = .588,  = .007.

Following the conceptualization of tolerance, we then focused further on the majority of the participants that had a negative attitude towards a particular practice. As expected, of these participants those who recognized arguments to support the practice were more tolerant than those who did not think that there are relevant reasons (see Figure 2). This was found for all six practices; for ‘homosexual teachers’, F(1, 397) = 56.76, p < .001,  = .13, for ‘non‐vaccination’, F(1, 406) = 122.82, p < .001,  = .23, for ‘women in management’, F(1, 415) = 141.37, p < .001,  = .25, for ‘abortion speech’, F(1, 382) = 225.94, p < .001,  = .18, for ‘refusing children at school’, F(1, 366) = 125.42, p < .001,  = .26, and for ‘Nashville declaration’, F(1, 415) = 182.27, p < .001  = .31.

Figure 2

Tolerance of Orthodox Protestant practices for participants who are negative towards the practice by reasonable counterargument ‘yes’ or ‘no’. A higher score indicates higher tolerance.

We conducted sensitivity power analyses for the inferential main effect of tolerance among those with negative attitudes (n = 384–417), accounting for unequal group size. The sensitivity power analysis revealed that, at a desired power of .80 and α = .05, the samples in Study 2 achieved sensitivity to detect at least between small to medium effects ( ≥ .0218–.0451).

Study 2 again demonstrates that people overall tend to be more tolerant of practices they disapprove of when they additionally recognize that there are relevant arguments for supporting that practice. Such an overall effect was not found among participants with a positive attitude, which further suggests that recognition per se does not weaken the relation between attitude and behavioural intention. The effect on tolerance was found for a range of practices of a native religious minority group that likely does not draw on the same prejudices that are common against immigrant‐origin Muslim minorities (Ogan, Willnat, Pennington, & Bashir, 2014; Strabac & Listhaug, 2008). In general, people were more negative towards the practices and events in Study 2 than in Study 1, which suggests that these were considered more controversial, despite the enacting group being a native minority group. Further, most people who were negative did not think that there were relevant arguments to nevertheless support it. However, participants who were negative towards the practices while recognizing reasons that people may have for allowing these were indeed, as expected, more tolerant.

STUDY 3

Study 3 goes beyond the first two studies by experimentally testing whether perceiving others to have relevant arguments for a controversial action leads to higher tolerance of that action. Based on previous research (Adelman & Verkuyten, 2020), we focused on protest to the reception of refugees and the perceived reasons that the protester gives for the action. For the experimental manipulation, we used the familiar distinction between irrational (prejudicial) and rational (relevant) bases for negative outgroup attitudes (Billig, 1988). This distinction refers to whether a judgment or action is, or is not, based on clear thought and reason, and is commonly used by lay persons for criticizing as well as justifying people’s negative attitudes towards minority groups (e.g., Kleiner, 1998; Verkuyten, 1998), including refugees (Capdevila & Callaghan, 2008; Figgou & Condor, 2006). Furthermore, for examining the generality of the role of rationality for tolerance, we considered three different concerns that are voiced in debates about the continued arrival of refugees seeking asylum (see Esses, Medianu, & Lawson, 2013; Lynn & Lea, 2003), specifically (1) whether asylum seekers are genuinely in need or opportunistic, (2) the cultural threat to society that continuing immigration would imply, and (3) the economic costs that would be involved in the reception of refugees. Thus, we used a 2 (irrational vs. rational) × 3 (topic of concern) between‐subjects experimental design. We expected the experimental manipulation to affect the perceived reasonableness of the protest action which in turn has an impact on tolerance. Thus, we examined perceived reasonableness as the mediating mechanism between the (ir)rational nature of the protest action and tolerance. Further, we considered the perceived offensiveness of the protest action for assessing whether participants did indeed disapprove of the action and thus whether it involved for them a question of tolerance.

Participants were recruited as part of a national survey of ethnic majority Dutch adults collected through the survey company Kantar, who were representative by gender, age, education, and district of residence in the Netherlands. The experiment was embedded within a survey about immigration, diversity, and tolerance. The response rate was similar to the first two studies. This study consisted of 659 participants who responded to questions of interest. They ranged in age from 18 to 85 years (M = 47.40, SD = 15.70), 51.4% was female, of 26.1% had low‐level education, 33.1% had mid‐level education, and 40.8% had high‐level education.

Following Lindner and Nosek (2009) and based on Adelman and Verkuyten (2020), participants read a brief description of an act of controversial criticism, modelled on a newspaper‐style news article about citizens’ responses to the reception of refugees. The scenario described a situation in which a native Dutch male (Johan Kok) in the city of Gouda in the Netherlands had put up a poster in the front window of his house stating ‘NO, AGAINST REFUGEES’. The focus on a male person was for reasons of relevance because men are more likely than women to be involved in these sorts of protest actions.

The 2 (irrational vs. rational) × 3 (topic of concern) between‐subjects experimental design involved two independent variables that differed between the stories. First, the person putting up the poster gave either an irrational versus rational explanation for their action. These explanations were based on public debates and lay person’s own accounts (e.g., Capdevila & Callaghan, 2008; Verkuyten, 1998) and either involved more prejudiced feelings and beliefs, or argued for considered thought as an operationalization of being rational. Second, the explanation focused on one of three common arguments about the reception of refugees: the nature of refugees (‘I do not like refugees because they are only fortune seekers’ vs. ‘We first should know more about whether we are dealing with real refugees or rather with fortune seekers’), economic concerns (‘Refugees only cost us large sums of tax money’ vs. ‘We should think more about what the costs are of the arrival of refugees’), or cultural concerns (‘Refugees only threaten our culture’ vs. ‘We should think more about what the arrival of refugees might mean for our culture’).

Measures

Reasonableness

Participants were asked to indicate (1 = ‘Totally Disagree’ to 7 = ‘Totally Agree’) whether Kok’s explanation for his action was reasonable and understandable. An average score of the two questions was used in the analysis (r = .83, p < .001; M = 3.73, SD = 0.51).

Tolerance

Participants were presented with two questions (7‐point scales) measuring tolerance, ‘It should be accepted that Kok expresses his opinion is this way’, and ‘The neighbours should tolerate Kok’s action’. These two items were strongly correlated (r = .71, p < .001) and an average score was used (M = 4.45, SD = 1.37).

Offensiveness

To assess whether participants disapproved of the action and therefore made it a question of tolerance, participants were asked two questions (7‐point scales): ‘how offensive do you find the poster’?, and ‘how hurtful do you find the poster’? (r = .88, p < .001; M = 4.34, SD = 1.55).

Multivariate analysis

We conducted a multivariate analysis of the two‐way interaction between our primary experimental variable of rational versus irrational (2) and the specific topic of concern (3) on reasonableness, tolerance, and offensiveness. Neither the main effect of topic of concern, Λ = .992, F(6, 1302) = .87, p = .520,  = .004, nor the interaction, Λ = .983, F(6, 1302) = 1.90, p = .078,  = .009, reached statistical significance. However, a significant multivariate main effect for the rational versus irrational argument emerged, Λ = .976, F(3, 651) = 5.24, p = .001,  = .024, which we explore below.

Reasonableness and tolerance

We first looked at whether perceived reasonableness was affected by the experimental manipulation. We found that participants evaluated Kok’s explanation as being more reasonable in the rational condition (M = 3.97, SD = 1.48) than in the irrational condition (M = 3.50, SD = 1.50), F(1, 653) = 15.67, p < .001,  = .023. Further, rational arguments generated more tolerance (M = 4.57; SD = 1.34) than irrational arguments (M = 4.32; SD = 1.39), F(1, 653) = 5.20, p = .023,  = .008, and lower perceived offensiveness of the message (M = 4.19; SD = 1.55 vs M = 4.49; SD = 1.53), F(1, 653) = 5.43, p = .020,  = .008.

Mediation analysis

We then conducted mediation analysis to investigate the proposed mechanism of rational versus irrational arguments increasing perceived reasonableness of the controversial message, which then increases tolerance of the controversial message. Results indicates that that the total effect of the experimental manipulation indicated higher tolerance in the rational compared to irrational condition, Estimate = .25, SE = .11, 95% CI [0.043, 0.461]. This effect was mediated through perceived reasonableness, as the rational compared to irrational condition was perceived as more reasonable, Estimate = .47, SE = .12, 95% CI [0.242, 0.699], and, in turn, perceived reasonableness predicted greater tolerance, Estimate = .47, SE = .03, 95% CI [0.405, 0.526]. Bootstrapping showed a significant indirect effect of the rational versus irrational condition, Estimate = .22, SE = .06, 95% CI [0.108, 0.340]. The inclusion of the indirect path rendered the direct effect between the experimental manipulation and tolerance no longer statistically significant, Estimate = .03, SE = .09, 95% CI [−0.149, 0.216].

The role of offensiveness

As tolerance implies disapproval whereby people condone what they continue to object to because it is considered, for example, offensive and hurtful, we further investigated the relation between perceived offensiveness of the controversial message and tolerance of the message and messenger. Correlation analysis showed that the more offensive a person found the message, the less tolerant they were towards it, r = −.43, p < .01. However, when testing whether the degree of offensiveness moderates the effect of rational versus irrational messages on tolerance and reasonableness, we only found a strong multivariate main effect of offensiveness, Λ = .616, F(24, 1170) = 13.34, p < .001,  = .215, such that the more offensive the message was perceived to be the less tolerant a person is and the less reasonable. There were no significant interactions, all F < 1.29, all p > .113,  < .040, indicating that the role of rational arguments in promoting perceived reasonableness and tolerance did not depend on degree of offensiveness of the message.

Following our conceptualization of tolerance, we next focused only on those respondents who found the poster offensive (above the neutral midpoint of the scale, N = 323). For this subgroup, the level of offensiveness did not differ between the two experimental conditions, F(1, 321) = 3.28, p = .071,  = .010. Thus, putting up the poster was considered equally offensive in both conditions, which means that perceived offensiveness does not explain any reasonableness or tolerance difference between the two conditions.

Within this subsample, we found identical results to the full sample, such that multivariate analysis showed the main effect of (ir)rational arguments, Λ = .954, F(2, 316) = 7.55, p = .001,  = .046, with rational arguments perceived as being more reasonable, F(1, 317) = 13.28, p < .001,  = .040, and more tolerable, F(1, 317) = 6.21, p = .013,  = .019, than irrational ones.

Similarly, mediation analysis showed that the total effect of the experimental manipulation indicated higher tolerance in the rational condition, B = .38, SE = .15, 95% CI [0.083, 0.669], with the protest behaviour in this condition being perceived as more reasonable, B = .55, SE = .15, 95% CI [0.255, 0.855], and, in turn, more tolerable, B = .33, SE = .05, 95% CI [0.226, 0.428]. Bootstrapping revealed an indirect effect of perceived reasonableness, B = .18, SE = .06, 95% CI [0.075, 0.309], with the direct effect between the (ir)rational manipulation and tolerance no longer statistically significant, B = .19, SE = .14, 95% CI [−0.088, 0.478].

Sensitivity power analysis for the main effect of the (ir)rational manipulation revealed that, at a desired power of .80 and α = .05, our full sample of 659 participants achieved sufficient sensitivity to detect at least a small effect ( ≥ .0118).

Study 3 demonstrated that people tend to be more tolerant of a controversial action, if the explanation for this action is perceived to be more rationally based. This effect was found independently of the perceived offensiveness of the action, which suggests that perceived rationality caused acceptance more generally and not only tolerance. The effect was found with an experimental manipulation in which participants simply read a short text in an online questionnaire and with the use of a rational condition that only argued for the need for considered thought about the impact of refugees on society. The fact that our manipulation affected perceived reasonableness, which in turn was related to acceptance and tolerance specifically thus suggests that even short explanations of controversial actions can affect the behaviour to these practices.

GENERAL DISCUSSION

Meaningful differences in beliefs, convictions, and worldviews are inevitable and make tolerance necessary in diverse societies. And toleration as the willingness to put up with things one feels negatively about, makes diversity possible (Walzer, 1997). Tolerance involves perceiving relevant reasons (e.g., not out of fear, or compulsion) for not interfering with disapproved conduct of others (Cohen, 2004; Forst, 2013). We tested this central theoretical proposition for the first time by systematically investigating whether the recognition of relevant arguments for negatively evaluated practices leads to higher tolerance of those practices.

In three studies with national samples, we used different operationalizations and variable tests of this proposition as a matter of conceptual replication (Crandall & Sherman, 2016). Specifically, we first focused on individual differences in self‐recognized reasons for allowing Muslim (Study 1) and Orthodox Protestant (Study 2) religious minority groups to engage in a broad range of practices in different contexts to examine its impact on tolerance. Then, we manipulated the rationality of offensive protest actions (Study 3) to examine the impact of this framing on perceived reasonableness and tolerance, and we found that considered thought as a rational basis for a controversial action was considered more reasonable, and therefore, tolerated more, independently of the perceived offensiveness of the action. This latter finding indicates that in Study 3 perceived reasonableness was not only associated with tolerance which implies disapproval, but with higher acceptance more generally. A possible reason is that on average the protest action was not considered very offensive, whereas in Studies 1 and 2, a majority of participants was clearly negative of the dissenting practices. Overall, the findings across the three studies support the notion that tolerance is higher when people recognize relevant arguments to support what they are negative towards. Depending on the specific practice, these arguments have to do with the importance of principles of free speech, freedom of education, religious freedoms, and minority rights, as well as pragmatic considerations of coexistence and peaceful cohabitation (Kirchner, Freitag, & Rapp, 2011). The use of more principled and pragmatic reasons for tolerance is in line with theoretical discussions (Forst, 2013) and with lay people’s endorsement of two forms of tolerance (principal‐based respect, and pragmatic‐based coexistence) in relation to different minority groups (Velthuis et al., 2021). Future research could examine further the types of arguments that people use in thinking about whether, why, and when disapproved of conduct should or should not be tolerated (e.g., Verkuyten & Kollar, 2021).

The pattern of results demonstrates that people are capable of tolerating a range of practices that they dislike, disapprove of, or disagree with and which might affect them personally or the wider society. Importantly, the range of practices, the different social contexts, the different minority groups (immigrant and non‐immigrant origin) and the focus on self‐reported and perceived reasons allowed us to determine that the recognition of relevant reasons is associated with higher tolerance. For example, we presented participants with scenarios that differ in the degree of required majority group accommodation and some of these practices are perceived as more demanding than others. Although the degree of tolerance differed, the role of recognizing relevant reasons for tolerance was similar across less and more demanding minority practices. This conceptual replication supports the theoretical proposition that values and principles can overpower negative feelings and beliefs and thereby enhances our confidence in the process underlying tolerance (Verkuyten et al., 2020). Yet, it is important to note that a majority of participants were intolerant and did not think that there were relevant arguments to nevertheless accept the dissenting practice.

Limitations

In the first two studies, we relied on data from self‐selected participants who disapproved of conduct but nonetheless recognized relevant arguments to support the disapproved practices. This was done because some people may simply see no reason to tolerate what they morally disapprove of, while others may not disapprove of the conduct in the first place, which makes it not a matter of toleration (Cohen, 2004). However, in Study 1, there were relatively few participants who were negative but nevertheless recognized relevant arguments. One reason for this might be that we used a binary attitude measure. In the current research, this measure has the advantage of providing a stable and reliable evaluative distinction (e.g., Dolnicar et al., 2011; Dolnicar, & Leisch, 2012), but a possible downside is that people are forced in one direction or the other.

Furthermore, in the first two studies, we considered individual differences in attitudes and self‐recognized reasons, while Study 3 used an experimental framing manipulation. Such a design could also be used in future research for examining experimentally whether the act of engaging in, for example, deliberative thinking about various reasons to accept or not accept disapproved of practices affect people’s tolerance. Additionally, future research could also examine whether tolerance depends on individual difference in, for example, open‐mindedness and rational decision‐making style (Stanovich, 2011) or dialectical thinking (Spencer‐Rogers, Williams, & Peng, 2010) and need for cognitive closure (Webster & Kruglanski, 1994).

Finally, future research could examine the importance of the situational context for making it less or more likely to consider additional reasons for tolerance. For example, whereas secure and stable situations might increase tolerance, threatening and uncertain situations are likely to reduce reflective thinking and lower tolerance (Capelos & Van Troost, 2012; Haas, & Cunningham, 2014). The latter situations can trigger feelings of fear or anxiety that are associated with reduced capacities to use cognitive abilities and influence how information and arguments are being processed and evaluated (MacLeod & Mathews, 2012).

Conclusion

We have provided a first systematic empirical test of the theoretical proposition that tolerance is based on relevant arguments (Cohen, 2004; Forst, 2013). There are norms, principles, and values that provide justified reasons for condoning what one continues to disapprove of. What is considered a relevant argument will depend on social, cultural, and historical circumstances and can differ between individuals. However, this does not mean that there are no general moral principles and no shared social conventions about what is and what is not relevant and acceptable. Furthermore, rationality has a strong first‐person perspective because individuals themselves must recognize and appreciate the relevant reasons. We focused on this perspective and demonstrated that recognizing relevant arguments can increase tolerance. This indicates that individuals are able to think about the complexity of living in a diverse world in which disapproval, dislike, and disagreement are inevitable, but allowing others to live the life that they want is necessary. Those who hold an objection to dissenting practices but see good reasons for not transferring this in a rejection are the ones engaged in toleration for a peaceful plural society.

Conflicts of interest

All authors declare no conflict of interest.

Author Contribution

Maykel Verkuyten: Conceptualization (equal); Funding acquisition (equal); Methodology (equal); Writing – original draft (equal). Anniek Schlette: Data curation (equal); Formal analysis (equal). Levi Adelman: Conceptualization (equal); Methodology (equal). Kumar Yogeeswaran: Conceptualization (equal); Methodology (equal); Writing – original draft (equal).