1. Introduction
Liberal states’ handling of ethnic minority rights has been a pressing issue in social sciences for decades, and the approach of multiculturalism—the accounting for cultural diversity (Kymlicka Reference Kymlicka1995)—has received much attention. A vast research corpus investigates the nexus between public attitudes and multiculturalism in general. For example, works from many disciplines cover support for multiculturalism as a political ideology or (un-)desired societal composition in various settings (e.g., Arends-Tóth and van Vijver Reference Arends-Tóth and de van Vijver2003; Link and Oldendick Reference Link and Oldendick1996; Nagayoshi Reference Nagayoshi2011; Verkuyten and Brug Reference Verkuyten and Brug2004). Others investigate if multiculturalist policies shape citizens’ attitudes toward migrants and immigration issues (e.g., Bartram and Jarochova Reference Bartram and Jarochova2022; Citrin, Levy, and Wright Reference Citrin, Levy and Wright2014; Hooghe and Vroome Reference Hooghe and de Vroome2015). Very little attention is spent, however, on public attitudes toward ethnic minority rights policies per se (for noteworthy exceptions see Goodman and Alarian Reference Goodman and Alarian2021; Stolle et al. Reference Stolle, Harell, Soroka and Behnke2016).
My paper contributes to this research by focusing on public attitudes toward policies addressing Indigenous rights issues. Most of the world’s Indigenous peoples live in political systems that were externally imposed on them. Through processes of colonization, they were deprived of fundamental collective rights, such as rights to self-rule, own and use traditional lands, speak Indigenous languages, and preserve and develop their cultural heritage.
In political theory, recognizing Indigenous rights is seen as a measure of rectifying such injustices (e.g., Moore Reference Moore2003; Mörkenstam Reference Mörkenstam2015). In practice, however, states often fail to adopt policies that create a comprehensive Indigenous rights regime (c.f. Lightfoot Reference Lightfoot2012). Existing explanations argue that domestic actors’—for example, governments or parliaments—conventional understandings of nationhood and state sovereignty conflict with Indigenous rights (e.g., Mörkenstam Reference Mörkenstam2019). Yet, they do not factor in the role of public attitudes. In democratic societies, public support is of great importance in policy-making. The shortcomings of current Indigenous policies thus raise the question of whether there is a lack of public support and demand for a genuine recognition of Indigenous rights. Which Indigenous policies receive support from the public, that is, the Indigenous and the ethnic majority population,Footnote 1 and which do not? And do preferences differ between these groups?
To investigate these questions, I focus on the case of the Sámi people in Norway and Sweden.Footnote 2 The countries differ considerably in the Sámi policies they pursue today. This allows for comparing public preferences toward a variation of real-life policies. Moreover, like many other Indigenous peoples, the Sámi suffered injustices such as losses of their traditional lands, culture and languages for centuries (Lantto Reference Lantto2010; Reference Lantto2014; Minde Reference Minde, Jentoft, Minde and Nilsen2003a; Trosterud Reference Trosterud2008). Yet, unlike many other Indigenous peoples, the Sámi in Norway and Sweden today live in affluent welfare states with high material standards of living that do not differ substantially between the majority and Indigenous population (Yasar et al. Reference Yasar, Bergmann, Lloyd-Smith, Schmid, Kupisch and Holzinger2023). This, in turn, creates a least-likely scenario of finding group differences in policy preferences based on material disparities. Finally, findings that the Sámi are, in comparison to the majority population, more likely to experience discrimination (Hansen et al. Reference Hansen, Minton, Friborg and Sørlie2016; Yasar et al. Reference Yasar, Bergmann, Lloyd-Smith, Schmid, Kupisch and Holzinger2023) suggest that other inequalities still exist and remain to be rectified.
2. Indigenous Rights and Sámi Policies in Norway and Sweden
A substantial part of political science research on Sámi rights—that is, Indigenous rights the Sámi hold as a people—revolves around the Sámi Parliaments.Footnote 3 Descriptions of the designs of these popularly elected self-government institutions point out their bearing on Sámi self-determination (e.g., Broderstad Reference Broderstad2011; Falch, Selle, and Strømsnes Reference Falch, Selle and Strømsnes2016; Lawrence and Mörkenstam Reference Lawrence and Mörkenstam2016). Comparatively, the Norwegian institution is better equipped to advance Sámi rights owing to its greater autonomy, influence in the decision-making process and scope of responsibilities (e.g., Henriksen Reference Henriksen2008; Josefsen, Mörkenstam, and Saglie Reference Josefsen, Mörkenstam and Saglie2015). The Swedish Sámi Parliament’s responsibilities are assigned by the national government and, in large part, concern administrative tasks related to reindeer husbandry (Josefsen, Mörkenstam, and Saglie Reference Josefsen, Mörkenstam and Saglie2015; Lawrence and Mörkenstam Reference Lawrence and Mörkenstam2016).
The focus on reindeer husbandry, the most prominent form of traditional Sámi livelihood, runs like a common thread through Swedish Sámi policies. For a long time, the official policy goal was “to protect [reindeer herding] Sámi from the detrimental influences of Swedish society” (Lantto Reference Lantto2014, 54) through segregation, while all those Sámi not practicing reindeer husbandry were to be assimilated. Even though these discriminatory policies were abandoned in the 1970s, the consequences of dividing the Sámi into two categories are still visible in contemporary policies and politics (see also Kvist Reference Kvist1994). For example, Sámi territorial rights, such as the right to use land for traditional livelihoods, only apply to members of reindeer herding corporations (Strömgren Reference Strömgren, Allard and Funderud Skogvang2017). The main political cleavage within the Sámi Parliament “separates parties representing the reindeer herders from those representing other Sámi interests” (Saglie, Mörkenstam, and Bergh Reference Saglie, Mörkenstam and Bergh2020).
Norway’s assimilationist norwegianization policy lasted well into the 20th century, too. However, it did not differentiate between reindeer herding and other Sámi but instead aimed at assimilating all Sámi—and other ethnic minorities—into the Norwegian mainstream society (e.g., Minde Reference Minde2003b). The so-called Alta affair in 1979 was an important watershed in Norway’s Sámi policies. Protests against the damming of the Alta River in the traditional Sámi settlement area developed into a broad discussion about Sámi self-determination and the recognition of their rights in general (Minde Reference Minde, Jentoft, Minde and Nilsen2003a). As a result, Norway was one of the first countries to ratify the legally binding International Labour Organization’s Indigenous and Tribal Peoples Convention (1989), which Sweden, for instance, has not signed to this day.
As becomes apparent in the course of this article, current Norwegian Sámi policies are more advanced in realizing Sámi’s rights to land and natural resources than Swedish and apply to the entire Norwegian Sámi population (see also Allard Reference Allard2011). Likewise, they also facilitate Sámi’s rights to use and maintain their Indigenous languages more assertively (see also Lloyd-Smith et al. Reference Lloyd-Smith, Bergmann, Hund and Kupisch2023). Combined with the above-mentioned greater self-governance capacity of the Norwegian Sámi Parliament, a consistent picture emerges illustrating the comparatively higher level of Sámi rights recognition in Norway.
Knowledge about public attitudes toward these issues, however, is scarce. Opinion research in this area mainly employs surveys among the Sámi Parliaments’ electorates.Footnote 4 Findings show that self-determination is an essential political cleavage within the electorate in Norway but not in Sweden (Bergh and Saglie Reference Bergh, Saglie and Berg-Nordlie2016; Saglie, Mörkenstam, and Bergh Reference Saglie, Mörkenstam and Bergh2020). Because of the Swedish Sámi Parliament’s weak self-determination power, its voters largely agree on the demand for increased responsibility (Mörkenstam, Nilsson, and Dahlberg Reference Mörkenstam, Nilsson, Dahlberg, Koivurova, Broderstad, Cambou, Dorough and Stammler2021).
Thus, even though empirical evidence suggests that the Sámi have preferences for increased recognition of their rights, the preferences these works study are limited to the Sámi Parliaments in substance and the group of registered voters in scope. Also viewed from a wider angle, the vast research corpus on public opinion and minority issues worldwide only sporadically features attitudes toward Indigenous rights. Most of the pertinent works focus on the majority population. They examine the role anti-Indigenous racism (e.g., Beauvais Reference Beauvais2022) or individual social-psychological factors play in attitudes toward Indigenous policies (e.g., Hartley, McGarty, and Donaghue Reference Hartley, McGarty and Donaghue2013; Milojev, Sengupta, and Sibley Reference Milojev, Sengupta and Sibley2014; Pehrson, González, and Brown Reference Pehrson, González and Brown2011) or look at broad issues like constitutional recognition (e.g., Levy and McAllister Reference Levy and McAllister2022). On the whole, prior research did not yet comprehensively analyze—neither in the case of the Sámi nor of any other Indigenous peoples—attitudes by both Indigenous and majority populations toward the recognition of Indigenous rights.
3. Theoretical considerations
The fundamental theoretical consideration induced by the research questions is how ethnic identity matters for such attitudes. In what ways might the preferences of an Indigenous person differ from that of a non-Indigenous person when it comes to Indigenous policies? The obvious answer from a rational resource-maximizing perspective is self-interest (c.f. Kim Reference Kim2014). If Indigenous and non-Indigenous individuals expect different utilities produced by the outcomes of Indigenous rights recognition, then they are likely to have different preferences toward it.
For example, it is not hard to imagine that an Indigenous person would be in favor of increased recognition of Indigenous territorial rights because being better able to practice traditional livelihoods has a high utility for them. A person with an ethnic majority identity, in turn, might oppose such a policy because it potentially restrains other forms of land use—be they economical or recreational—that have higher utilities for them. Self-interest and utility functions of the non-Indigenous population might be less evident in areas like self-governance or linguistic rights, though, where material zero-sum competitions are less apparent.
3.1. Social identity
Here, a social identity perspective (Tajfel Reference Tajfel1974; Tajfel and Turner Reference Tajfel, Turner, Austin and Worchel1979) can help theorize policy preferences. It implies that individuals pursue positive social identities and draw comparisons of their identity group’s position with other groups in society. Based on this, established social psychology theories such as social dominance theory (e.g., Pratto, Sidanius, and Levin Reference Pratto, Sidanius and Levin2006; Sidanius and Pratto Reference Sidanius and Pratto1999) or intergroup threat theory (e.g., Stephan, Ybarra, and Morrison Reference Stephan, Ybarra, Morrison and Todd2009) view comparisons of relative group status as pivotal in intergroup relations. Groups with higher status—not only in terms of collective economic resources but also political power and cultural value, that is, the wider society’s appraisal of the group’s cultural norms and practices—seek to perpetuate the social order. Consequently, they should be skeptical of policies that potentially increase the relative status of lower-ranking groups.
Yet, recognizing Indigenous rights implies advancing Indigenous people’s political power, economic opportunities, and social and cultural evolvement. In this vein, the rectification of historical injustices raises not only Indigenous people’s absolute status in relation to the state but also its relative standing vis-à-vis the ethnic majority group. Accordingly, the latter’s aversion toward Indigenous policies should grow with the degree to which these policies realize Indigenous rights.
3.2. Hypotheses
Based on these considerations, the basal hypothesis of this paper expects the following:
H1: The majority population’s support for policies conducive to realizing Sámi rights is lower than the Sámi population’s support.
This does not necessarily mean, however, that the groups’ preferences are completely diametrical. Because social hierarchies are often intractable, system justification theory (e.g., Jost, Banaji, and Nosek Reference Jost, Banaji and Nosek2004), for example, argues that social and psychological needs motivate members of disadvantaged groups to justify the existing order, at least to some extent. Furthermore, tests of intergroup contact theory consistently find that more contact between members of in- and outgroups decreases prejudice (Pettigrew and Tropp Reference Pettigrew and Tropp2006). Contact with outgroup members can hence reduce perceptions of status threat (Visintin et al. Reference Visintin, Green, Falomir-Pichastor and Berent2020).Footnote 5 Thus, if there is frequent contact between an ethnic majority and an Indigenous people, the former might be more open to an increase in the recognition of the latter’s rights. In the case of the Sámi in Norway and Sweden, it makes sense to assume a high level of contact with the respective ethnic majority populations. These groups have been coexisting in the traditional Sámi settlement areas for centuries (Kent Reference Kent2018). As will become evident further below, the regions considered in the empirical part of this paper are, in particular, characterized by a comparatively high share of Sámi residents, making frequent encounters between the groups very probable. In sum, I expect that middle-ground policies receive more support among both groups than extreme policies. That is, the Sámi population should prefer policies that seek to establish status equality over policies that would even put them above the majority by reversing the status hierarchy:
H2: The more a Sámi policy provides for status equality, the more support this policy receives among the Sámi population.
In a similar vein, the majority should be more supportive of policies that enable some degree of Sámi rights realization as compared to very inhibiting policies:
H3: The majority population’s support for policies precluding the realization of Sámi rights is lower than its support for policies that allow for a moderate level of Sámi rights realization.
Eventually, historical institutionalism and policy feedback theories argue that an existing policy context shapes people’s attitudes toward social and political issues (c.f. Béland, Campbell, and Weaver Reference Béland, Campbell and Weaver2022; Campbell Reference Campbell2012; Mettler and Soss Reference Mettler and Soss2004). As the previous chapter pointed out, the recognition of Sámi rights is currently more advanced in Norway than in Sweden. Against this backdrop, it seems likely that public opinion about Sámi policies differs between the two countries. Support for increased Sámi rights recognition might well be higher in Norway because the current policies’ “interpretive effects” (Pierson Reference Pierson1993, 610) signal to the public that a—comparatively—more substantial recognition of Sámi rights is in its interest and constitutes “the desirable state of affairs” (Svallfors Reference Svallfors2010, 120):
H4: Support for policies conducive to realizing Sámi rights is higher in Norway than in Sweden.
4. Data and method
For empirically testing these hypotheses, I use original survey data collected in 20 northern municipalities of Norway and Sweden during spring and summer 2021.Footnote 6 These municipalities were selected based on their share of residents registered with the Sámi Parliaments’ electoral rolls. In the absence of official Sámi census data (c.f. Axelsson and Storm Mienna Reference Axelsson, Storm Mienna, Walter, Kukutai, Russo Carroll and Rodriguez-Lonebear2021; Young and Bjerregaard Reference Young and Bjerregaard2019), this served as a proxy indicating which municipalities potentially have the highest shares of Sámi residents.
As a consequence, the survey’s scope is geographically restricted. At the same time, however, it is highly representative of the municipalities’ population precisely because of this restriction. The manageable numbers of municipality residents did not necessitate drawing samples. Instead, we attempted to contact all 17,096 Norwegian and 22,073 Swedish adult residents whose telephone numbers were available. In each case, about 14% of them participated in a brief telephone interview. About 22% of the respondents in Norway and 12% of the Swedish participants stated that they identify as Sámi.
Owing to the survey’s overall volume, all interview respondents were invited to fill out a subsequent, more extensive questionnaire. A total of 502 respondents from Norway and 867 from Sweden did so online. They constitute 21% (Norway) and 29% (Sweden) of the telephone interview participants.
4.1. Policy vignettes
These respondents’ answers are the observational units for this paper, as the online questionnaire contained, among other elements, three factorial items to measure preferences for Sámi policies. They cover the essential aspects of Sámi rights—language and education, self-governance, and territorial rights (c.f. Anaya Reference Anaya2011; Tauli-Corpuz Reference Tauli-Corpuz2016). The overall design consists of vignettes that describe an issue related to Sámi rights in one of these areas featuring common and salient topics.Footnote 7
Respondents saw two profiles, each entailing two policy features as potential strategies for dealing with the matter. This multidimensionality is the vignettes’ main advantage over ordinary attitudinal survey questions. They allow respondents to jointly evaluate both features, which would otherwise require two separate survey questions. Moreover, employing conjoint analysis on vignette responses also enables exploiting between-feature links (Bansak et al. Reference Bansak, Hainmueller, Hopkins and Yamamoto2021).
The levels of the features, that is, the policies’ content, were randomly chosen from a set of four possible levels each.Footnote 8 For all features, two levels were fictitious policies, whereas the remaining two were based on existing Swedish and Norwegian policies. The first level constitutes the policy alternative most precluding for Sámi rights. While the second level resembles a moderate step toward Sámi rights facilitation (the Swedish policy), it is not as affirmative as the third (the Norwegian policy). Finally, the fourth level constitutes a policy that would enhance the Sámi’s status to the greatest extent, up to the point of exceeding the majorities.
Eventually, respondents had to choose the profile they felt was more beneficial for “the population as a whole.” Naturally, this does not imply that one solution necessarily needs to provide benefits for everyone in society. Instead, the forced-choice question requires respondents to trade off potential benefits for one part of society against possible losses for other parts and choose the option that, in their view, comprises the better cost–benefit ratio. The focus on the benefits for the overall population makes the question a neutral instrument to detect potential group-level differences in policy. It does not prime respondents to evaluate the cost–benefit ratios against the background of their own ethnic identity. To be sure, asking about the general population’s benefits is also a primer toward national identity. However, prior research does not find significant differences between the Sámi and majority populations’ levels of national identity in either country (Gerdner Reference Gerdner2021; Selle, Semb, and Strømsnes Reference Selle, Semb and Strømsnes2013). I thus assume that members of both groups feel similar levels of attachment to their countryFootnote 9 and should hence be equally motivated to choose what they believe is the best solution for that country.
4.1.1. Language and education
The first vignette compares different Sámi language and education policies. It concerns the Sámi languages’ official status and the extent to which education in Sámi is available. The actual policies of Norway and Sweden on these issues vary in several aspects.Footnote 10
Whereas Sweden recognizes Sámi as one of several national minority languages (c.f. Law on National Minorities and Minority Languages 2009), its official status in Norway is that of a language of “equal worth” to Norwegian (The Sámi Act 1987, ch. 1, para. 1–5). Furthermore, both countries have designated areas where additional Sámi language rights apply, like the right to use Sámi when communicating with public institutions. However, the Norwegian provisions are more comprehensive. They pertain to more parts of public life and oblige public authorities to make all their publications available in Sámi.Footnote 11
Likewise, the extent to which Sámi education is provided differs. In Sweden, a handful of sameskolorna [Sámi schools] follow a syllabus focused on bilingualism. This syllabus covers the first 6x years of schooling and allocates 800 hours of instruction to teaching in Sámi and 910 in Swedish (Belancic and Lindgren Reference Belancic and Lindgren2020).
In Norway, by contrast, every child has the right to receive education in Sámi throughout primary and secondary education (Sollid Reference Sollid, TA and Sollid2022). Even though this right is implemented differently within and outside the Sámi language area, it enables receiving Sámi education until year 10, including the possibility of first language teaching, that is, Sámi being the primary language of instruction across subjects (Vangsnes Reference Vangsnes2022). Since Norway is home to the only higher education institution where Sámi is the primary teaching and working language—the Sámi University of Applied Sciences in Kautokeino—it is possible to receive Sámi tuition on all education levels.Footnote 12 These policy differences are the basis for the Swedish (II) and Norwegian (III) feature levels, as listed in Table 1. The other two levels are fictitious, taking the real-life policies to the extreme—either in a way detracting from (I) or privileging (IV) Sámi status.
Levels of language and education policy features
4.1.2. Self-governance
The next vignette presented two Sámi representative institutions with varying levels of decision-making power and scopes of responsibility. Evidently, it measures self-governance policy preferences with respect to the Sámi Parliaments. However, it does not mention the term Sámi Parliament to avoid priming respondents too much with their preexisting attitudes toward their country’s real Sámi Parliament.
As mentioned, the Norwegian Sámi Parliament has more political power and a broader mandate than its Swedish counterpart. For example, since 2005, a comprehensive consultation agreement exists with the Norwegian government. It codifies that all state authorities must consult the Sámi Parliament before making legislative, regulative, or administrative decisions that may affect Sámi interests. Moreover, it states that “[t]he Sami Parliament can also independently identify matters which in its view should be subject to consultations” (Procedures for Consultations between State Authorities and The Sami Parliament, 2005, ch. 6, para. 4). However, there is no actual veto right. When consultations do not result in an agreement, the government has the final say.
At the time the survey was conducted, no comparable consultation process existed in Sweden.Footnote 13 Instead, legislation such as the Minerals Act (1991) only arranged for duties to inform the Sámi Parliament in matters affecting reindeer herding. Furthermore, the Swedish government can determine the Sámi Parliament’s areas of responsibility, and the latter has little say in what these should and should not include. Its Norwegian counterpart is more independent in deciding where its responsibilities should lie (Josefsen, Mörkenstam, and Saglie Reference Josefsen, Mörkenstam and Saglie2015). Table 2 lists the feature levels that reflect these differences between Norway and Sweden and the levels contrived to represent the extrapolations on both sides.
Levels of self-governance policy features
4.1.3. Territorial rights
The third vignette centers on conflicting land use interests. It depicts a mining company applying for an extraction permit in an area where a local Sámi community practices reindeer herding. Albeit a hypothetical case, it strongly references actual conflicts commonly occurring in Norway’s and Sweden’s northern regions.Footnote 14 Reindeer husbandry requires large land areas for animals to roam and graze. Infrastructure and industry projects impacting the landscape and natural environment in reindeer herding areas are thus often seen as a grave impediment.
Central to such conflicts of interest are land ownership and usage rights questions. Norway and Sweden do not per se recognize the Sámi claim of a usufruct land ownership right to territories used since time immemorial. Instead, in Sweden, Sámi land-rights claims generally must be brought before courts, which has so far proven to entail high costs and slim prospects of success for the Sámi side (Torp Reference Torp2013).Footnote 15 In Norway, a significant policy change occurred when, through the introduction of the Finnmark Act (2005), the ownership of all previously state-owned land in Finnmark—Norway’s northernmost mainland region—was transferred to the newly founded Finnmark Estate. Its governing board consists in equal parts of appointees from the Sámi Parliament and county administration. In addition, the act established a commission that investigates land use and ownership rights cases and a special tribunal for settling related disputes.
Besides the issue of who decides about contested territorial rights claims, Swedish and Norwegian policies also differ regarding how these decisions are made. In both countries, assessments of the impact on Sámi cultural practices are required before concessions for other land use activities can be granted. In Sweden, these assessments focus on potential economic repercussions for reindeer herding; commercial land use interests and Sámi reindeer herding are treated “as competing economic interests” (Raitio, Allard, and Lawrence Reference Raitio, Allard and Lawrence2020, 6). In contrast, Norwegian legislation stipulates that decisions should take both the material and immaterial value of land for all Sámi cultural practices—including and exceeding reindeer husbandry—into account when weighing up conflicting interests (e.g., the Nature Diversity Act 2009; the Minerals Act 2009). Accordingly, as listed in Table 3, the vignette about attitudes toward territorial rights policies entails features about who makes decisions and on what basis.
Levels of territorial rights policy feature
4.2. Group membership and identity
Before moving on to the empirical assessment of whether groups’ preferences toward these vignettes differ, elaborating on how group membership is determined is necessary. There is no single hard-and-fast definition of the Sámi population (Pettersen and Brustad Reference Pettersen and Brustad2013). In the context of this paper, it makes sense to focus on self-identification. The theory’s social identity perspective presupposes that individuals identify with an ingroup to draw comparisons with an outgroup. That is, all those who consider themselves to be Sámi are assumed to assess Sámi policies in light of their potential threat or benefit for the Sámi’s relative status. Majority-population members, in turn, are those who do not self-identify as Sámi and for whom the Sámi population, hence, constitutes an outgroup. I operationalize group membership based on respondents’ answers to the survey question, “Overall, do you consider yourself to be Sámi?”
However, it would be a fallacy to count everyone who answered “no” toward the majority population. There are, of course, other ethnic minorities in Norway and Sweden apart from the Sámi. Their perceptions of status threats posed by Sámi policies might differ from that of the ethnic majority, too. For the analysis, I thus consider those survey respondents as members of the majority population who do not self-identify as Sámi and report having an ethnic Norwegian and Swedish background, respectively.Footnote 16
4.3. Conjoint analysis
I use conjoint analysis to compute how much these groups support the policies presented in the vignettes. Since the seminal paper by Hainmueller, Hopkins, and Yamamoto (Reference Hainmueller, Hopkins and Yamamoto2014), conjoint analysis has become a prevalent method for factorial design studies of political preferences. Usually, the quantity of interest is the average marginal component effect (AMCE), which is the average effect a given feature level has—relative to a baseline level—on the probability of respondents choosing a profile. Calculating AMCEs is thus a suitable way to formally test hypotheses H2 and H3, which are about how much support a group of respondents shows for a particular policy compared to policies that are more—or less—conducive to realizing Sámi rights.
In the case of H1 and H4, however, the preferences of two groups of survey respondents—members of the Sámi and the majority population (H1) and Norwegians and Swedes (H4), respectively—need to be compared. Yet, following the argumentation of Leeper, Hobolt, and Tilley (Reference Leeper, Hobolt and Tilley2020), contrasting potential differences in AMCEs across subgroups would not constitute an effective comparison of differences in preferences.Footnote 17 Instead, preference differences between subgroups should be described with subgroup marginal means (MMs). These represent the “level of favorability toward profiles with a particular feature level, ignoring all other features” (Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020, 210).
5. Results
Table 4 lists the numbers of online survey respondents according to the group-membership operationalization above. Estimating Sámi population numbers and shares is very complex, and results can vary when applying different operationalizations (e.g., Pettersen and Brustad Reference Pettersen and Brustad2013). Nonetheless, Table 4 shows that the data fit well with the prevalent general assumptions about Sámi population distributions. The share of self-identifying Sámi respondents in Norway is close to the figures the SAMINOR study arrives at (Pettersen and Brustad Reference Pettersen and Brustad2013). Furthermore, the numbers are very much in line with the supposition that there are considerably more Sámi in Norway than in Sweden (Young and Bjerregaard Reference Young and Bjerregaard2019).
Online survey respondents by ethnicity and country
Note: Column percentages in parentheses; a total of 25 Norwegian and 104 Swedish respondents could not be assigned to either of the groups.
5.1. Marginal means
Looking at the MMs produced by the vignettes, Figures 1, 2 and 3 indicate that policy preferences differ between these two groups.Footnote 18 Their left panels show the majority population’s MMs, and the middle panels show the Sámi population’s MMs. In a forced-choice setting with two alternatives like the one at hand, MMs readily express the probability with which respondents chose a profile that includes the given feature level (Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020).
Subgroup comparison – Language and education.
Subgroup comparison – Self-governance.
Subgroup comparison – Territorial rights.
Thus, in Figures 1, 2, and 3, a MM to the right of the dark gray line positioned at 0.5 means that this level increases the probability of the policy profile being chosen, while a MM smaller than 0.5 reduces it. The panels on the right eventually show the group differences. Here, the light gray line positioned at 0.0 is essential: A difference value below it signifies that such levels receive more support among the Sámi than among the majority population. Positive difference values mean that majority support is larger.
In the case of the language and education policy vignette (Figure 1), the language status is the more pivotal feature for explaining respondents’ choices. Most of the education feature levels’ MMs are close to 0.5, indicating a weak influence on the likelihood of a profile being chosen. The only exception is Sámi respondents’ rejection of Sámi being merely an optional subject at school. With language status, the figure shows that the policy least conducive to Sámi rights (level I) produces higher MMs among the majority than the Sámi population. In contrast, more conducive policies (levels III and IV) produce larger MMs among the Sámi respondents. This is quite in line with hypothesis H1. The MMs of level II, that is, Sámi having the status of a minority language, are very similar between both groups.
The same is true for the self-governance policy vignette (Figure 2). Level II of the feature on the Sámi Parliament’s influence has similar MMs among majority and Sámi respondents. Yet, also the MMs of level IV—the right to object—do not differ between the groups. Instead, clear differences are apparent with level III—more support from Sámi respondents for a consultation obligation—and level I—more support from majority respondents for no entitlement to be involved in the decision-making process. Likewise, in the case of the second feature, that is, the Sámi Parliament’s scope of responsibility, the most apparent group differences in MMs occur in levels III and I. Overall, the findings thus support hypothesis H1, since majority respondents are less likely to choose policies conducive to the realization of Sámi rights.
Eventually, Figure 3 shows that preferences toward how land use conflicts should be decided neatly follow the pattern expected by H1. The majority is far less supportive of prioritizing reindeer husbandry (level IV) or protecting Sámi culture from considerable obstructions (level III). Instead, its support for prioritizing mineral extraction (level I) or basing the decision on an economic value comparison (level II) is considerably higher. Regarding who should decide in such cases, only level IV—a commission in which Sámi representatives hold the majority—exhibits a distinct difference because the majority’s MM is relatively small. As a matter of fact, the Sámi seem to be quite indifferent about this feature. Their MMs are close to 0.5, suggesting that who decides on land use conflicts is less relevant to them than the principles of the decision.
5.2. Average marginal component effects
Table 5 shows the AMCEs among the Sámi population for all three policy areas. In each case, level III constitutes the base level, as it is the supposedly most status-equalizing policy in each feature. There is some support for H2 as there is never a policy for which the Sámi population has significantly stronger preferences than the level-III policy. Unsurprisingly, the least preferred levels are the fictitious policies obstructing Sámi rights the most (level I). Only in the case of the territorial rights vignette’s first feature does the level-I AMCE not reach statistical significance.
Sámi population policy preferences (AMCEs)
Note: Standard errors are in parentheses clustered on respondent level; number of observations is the number of respondents times the number of profiles.
† p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.
Yet, the picture of whether the Sámi prefer levels based on Norwegian (level III) over Swedish (level II) policies is very vague. None of the respective AMCEs are statistically significant, indicating no preference in either direction. In addition, no clear trend is discernible on whether Sámi respondents prefer status-equalizing policies over policies that would potentially reverse the status hierarchy (level IV). The AMCE for Sámi being the only official language is significantly negative, as is the AMCE for the Sámi Parliament having the responsibilities of a provincial government. For other issues, the AMCEs of level-IV policies are not statistically significant.
Finally, Table 6 reports the results of testing whether the majority population prefers level-II over level-I policies. For this purpose, the latter now constitute the base levels. There is support for H3, as in most cases, the AMCE of level II is positive and statistically significant—yet sometimes only marginally. In two cases, though, level-II AMCEs lack statistical significance. The majority is indifferent about whether courts or the national government should decide on conflicting land use interests. Likewise, it has no preference regarding whether the Sámi Parliament’s responsibility should be defined by the government or confined to reindeer herding issues.
Majority population policy preferences (AMCEs)
Note: Standard errors are in parentheses clustered on respondent level; number of observations is the number of respondents times the number of profiles.
† p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.
Beyond that, Table 6 shows that the AMCEs of level-IV policies are significantly negative except for the second feature of the language and education and the first feature of the self-governance vignette. This suggests that in most instances, the majority prefers policies detracting from Sámi status over policies privileging Sámi status.
5.3. Cross-country differences
So far, the analysis pooled together data from both countries. H4, however, regards potential cross-country differences and expects that respondents from Norway are generally more supportive of stronger Sámi rights recognition. To test this claim, I again use MMs but base the subgroup comparison on the country level. As Table 4 shows, the share of Sámi respondents is considerably higher in Norway. On the one hand, it thus seems sensible to distinguish between Sámi and majority respondents simultaneously. Otherwise, any country differences might be due to differences in the weight of Sámi responses within the entirety of answers from each country.
On the other hand, Table 4 also shows that the absolute numbers of Sámi respondents within each country are rather low. This constitutes a serious statistical power issue for cross-country comparisons of Sámi’s preferences. In the following, I therefore concentrate on the comparisons of majority respondents, for which the numbers of observations per country are pretty sound. In Appendix D, I additionally provide the MMs of the cross-country comparison among Sámi respondents.
The results are again presented visually in Figures 4, 5, and 6. It becomes apparent that across all three areas, policy preferences are surprisingly similar between Norway and Sweden. Only the level-I education policy (Figure 4) produces statistically significantly different MMs among Norwegian and Swedish majority respondents. The latter are less likely to support Sámi being only an optional subject at school. Yet, importantly, there is no indication that the existing policy context shapes respondents’ preferences. Support for level-III policies is not higher in Norway, and support for level-II policies is not higher in Sweden.
Country comparison – Language and education – Majority respondents.
Country comparison – Self-governance – Majority respondents.
Country comparison – Territorial rights – Majority respondents.
Furthermore, when running formal nested-model tests of preference heterogeneity, none of the test statistics reach statistical significance.Footnote 19 That is, in none of the vignettes do majority respondents’ preferences, taken as a whole, differ significantly between Norway and Sweden. Overall, the analysis thus broadly refutes H4 and suggests that, in this case, policy feedback does not play a significant role in preference formation.
6. Discussion
At this point, it is necessary to put the findings into context. In one sense, they have a broad scope because there are no discernible trends of differing preferences between countries. The countries’ existing policy contexts do not seem to shape policy preferences. Instead, the findings suggest that preferences among the Sámi and majority populations align more with the group members’ respective social identities, which appear to be transnational.
In another sense, however, the scope of the analysis is narrow. As it was conducted in both countries exclusively in municipalities with a high share of Sámi residents, the findings cannot be readily expanded to the countries’ entire populations. Nevertheless, focusing on the selected municipalities makes sense. Many Sámi policies, too, are geographically limited in their scope and apply in the selected municipalities but not, for example, in southern metropolitan areas like Oslo or Stockholm. Lacking reliable information about how salient Indigenous rights issues actually are for the Norwegian and Swedish general public—and especially the ethnic majorities—it is sensible to concentrate for the time being on areas where Sámi policies actually apply and are hence probably most salient.
7. Conclusion
This paper studied public attitudes toward Indigenous policies that vary in how they facilitate Indigenous rights. It asked which policies the ethnic majority and the Indigenous population prefer. Investigating the case of the Sámi people in northern Norway and Sweden, it found preferences differing between the groups. Respondents from the ethnic majority have less positive attitudes toward policies conducive to Sámi rights but, more often than not, prefer policies that realize Sámi rights to some degree over policies that preclude them. Regarding Sámi respondents’ policy preferences, the findings are less clear-cut. Some support the hypotheses, while others suggest that Sámi respondents are often indifferent.
The ethnic majority, in particular, seems to employ a zero-sum status competition lens when evaluating policies of territorial rights, self-governance, and language and education. Regarding the latter, the results align with multiculturalism theory considerations about the status of minority languages. “For the minority group, official multilingualism is desired in part because it is a symbol of, and a step toward, acceptance that it is a multination state, a partnership of two or more nations within a single state. Yet this is precisely what members of the dominant group typically wish to avoid. For accepting that a regional language group is also a ‘nation’ has potentially far-reaching consequences” (Patten and Kymlicka Reference Patten, Kymlicka, Kymlicka and Patten2007, 5, italics adopted from the original).
The finding of distinctly conflicting group preferences regarding territorial rights, which has the most direct link to material matters, is especially noteworthy, since previous research finds no difference in material status between the Sámi and majority groups (Yasar et al. Reference Yasar, Bergmann, Lloyd-Smith, Schmid, Kupisch and Holzinger2023). Material issues thus also matter in the absence of economic inequality. This is also highly relevant to the situations of other Indigenous peoples around the world, which are most often characterized by grave economic inequalities. The Sámi case demonstrates that disparity in groups’ attitudes toward Indigenous policies will not be resolved by establishing solely material equality.
From a practical perspective, the findings suggest that policymakers cannot count on widespread support from ethnic majorities when proposing policies that sizably facilitate Indigenous rights. This could help explain why, despite provisions in international law and judicial decisions, there often seems to be a reluctance to introduce assertive policy measures. For example, at the time of writing, the Norwegian government has still not taken substantial action more than two years after the Supreme Court ruled that the construction of a large-scale wind farm on the Fosen Peninsula violated local Sámi’s human rights.Footnote 20
A way forward to enable more rectification of injustices against Indigenous peoples would need to target the boundaries of group identities to create a more common and inclusive identity (c.f. Gaertner and Dovidio Reference Gaertner and Dovidio2014). Naturally, this should not be understood in assimilationist terms. Instead, it connotes the acceptance of distinct cultures being equal intrinsic elements of the overall society. By way of example, public opinion research from the Chilean case shows that majority population support for Indigenous rights is higher when people see the Indigenous population as an inherent part of the country’s history and identity (Pehrson, González, and Brown Reference Pehrson, González and Brown2011).
Admittedly, long-established group identities will not be shaken up at the drop of a hat. Instead, laborious and persistent efforts are required to amend the public image of the interrelation between the nation and the Sámi. For example, school curricula, media, and politics need to represent Sámi culture and history as an integral component of the national culture and history. Nevertheless, making these efforts seems essential when striving for a proper rectification of injustices.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/rep.2023.38
Data availability statement
The replication data of this study are available in the Harvard Dataverse at https://doi.org/10.7910/DVN/9LX8VT.
Competing interests
None.
Open access
