2.1 Adolescents’ changing rejection of equal opportunities

According to realistic conflict theory, intergroup antagonism can be explained as a result of competition between social groups in society over scarce resources [2, 24–26]. Because of intergroup competition over scarce resources, incompatible group interests arise which are argued to enhance solidarity within a social group and induce antagonism between social groups in society (see [1] for more information). The likelihood to experience intergroup competition is assumed to be particularly prevalent among those whose socio-economic position resembles the position of out-group members, i.e. individuals who find themselves in the less favourable socio-economic strata of society [27, 28].

The dynamic version of realistic conflict theory does not assume that individuals have “permanent hostile attitude[s] toward the outgroup based on innate dispositions, authoritarian personalities, or slow social experiences” [8, p. 109], but proposes more flexibility and versatility of intergroup attitudes. This means focus is placed on differences in intergroup attitudes within individuals instead of differences between individuals [7, 11]. Hence, the general proposition of the dynamic perspective on intergroup competition argues that rejection of equal opportunities for foreigners increases when individuals’ socio-economic status deteriorates to a level that they are more strongly in competition with out-group members.

We add insights from the psychological development literature to this dynamic version of realistic conflict theory, stating that attitudes are especially prone to change during adolescence and young adulthood. Though there is no clear consensus about precisely which years people are most likely to change in their attitudes, this study follows other studies that argue that political socialisation still takes place up to the mid-twenties [15, 19]. According to the ‘impressionable years’-hypothesis, adolescence is pivotal in the formation of attitudes [19, 29]. Later in life, these attitudes become ‘crystallised’, indicating that individuals are then less vulnerable to attitudinal change. Such ‘crystallisation’ of intergroup attitudes is confirmed by for instance Henry and Sears [17], who demonstrate that most within-individual change in symbolic racism is observed during people’s early adulthood. Since the literature refers to different ages to describe the impressionable years, we include the population between 14 and 25 in this study, to which we refer as the adolescent and young adult population. We therefore expect that the dynamic version of realistic conflict theory is in particular applicable to adolescents and young adults. Therefore, our specific proposition is that primarily among this group rejection of equal opportunities is enhanced by a transition towards a position in which competition with ethnic out-groups is more likely.

A key transition during adolescence that is likely to increase experienced competition with ethnic outgroups is entering the labour market [20]. One of the main scarce resources that creates incompatible interests between social groups in society is jobs [1, 24]. Competition over jobs and economic privileges may create fear or anxiety towards ethnic outgroups, as they can undermine the natural superiority of the dominant group by claiming these scarce resources [30]. Hence, as adolescents complete their education and enter the labour market, they are expected to find themselves in a situation of competition with ethnic outgroups over jobs. Due to this higher level of competition, adolescents might become less supportive of equal opportunities for foreigners. Considering that especially those who are not able to find a job are vulnerable to experience competition from foreigners [31], we expect that entering the labour market but not being able to find a job increases the likelihood to reject equal opportunities even more. This results in the following hypothesis:

1. H1: (a) As adolescents enter the labour market, the likelihood to reject equal opportunities for foreigners increases, (b) and this is in particular the case when adolescents are not able to find a job.

Before completing education, some adolescents may already become acquainted or have a first encounter with the labour market. This holds primarily for adolescents who follow vocational education with high labour market relevance. Adolescents who receive vocational or professional training with special focus on labour market competences and who do work placement are more likely to get a sense of intergroup competition with ethnic outgroups already at an early stage of their career, which is expected to evoke an increase in anti-immigrant attitudes such as rejection of equal opportunities [20]. Previous research, furthermore, demonstrated that adolescents in vocational education are more likely to experience an increase in ethnocentrism over time compared to their counterparts in general education [32]. We therefore expect that:

1. H2: As adolescents start with vocational training, the likelihood to reject equal opportunities for foreigners increases.

Not only objective, tangible transitions in adolescents’ life can evoke a change in rejection of equal opportunities. Previous research [e.g. 12, 33] has demonstrated that perceptions of a worsened socio-economic position might also spark rejection of equal opportunities, independent of the influence of objective transitions in life. Blalock’s [2] distinction between actual and perceived competition helps us to understand the influence that independent, subjective changes can have on the change in rejection of equal opportunities. Though actual competition may be absent, individuals who perceive a deterioration in their socio-economic position can still have a higher level of perceived competition, which might in turn spark rejection of equal opportunities for foreigners [12, 34]. Hence, we expect that:

1. H3: As adolescents’ subjective economic dissatisfaction increases, the likelihood to reject equal opportunities for foreigners increases.

2.2 The impact of the economic household context

Childrens’ attitudes can be influenced by their parents in multiple ways. This might hold in particular for adolescents and young adults who still live with their parents and who are not economically independent yet. In this study, we focus on direct intergenerational transmission of attitudes within a family (see below) versus indirect transmission via parents’ (economic) position. The latter is called ‘positional parental socialisation’ [35], and underlines the impact of parents’ and childrens’ shared environement and common social status. Adolescents are exposed to divergent household experiences, because of differences in the social position of their parents and the household they live in. Hence, it is argued that indirect parental transmission occurs via the provision of the setting in which children develop attitudes [36].

Considering that the household context is the main economic context in which adolescents come of age, we argue that adolescents are not only influenced by transitions and changes in their own economic situation but also by changes in the economic situation of the household that they live in. Previous research demonstrates that the structural and cultural position of the household matters to adolescents’ rejection of equal opportunities [37]. Growing up in relatively favourable economic living conditions is associated with lower levels of perceived ethnic threat and, subsequently, more favourable intergroup attitudes [38, 39]. Disadvantageous living conditions, on the other hand, can cause perceptions of disstress among members of the household [40], which may affect parents’ and adolescents’ level of perceived or experienced competition and, subsequently, their intergroup attitudes. When combining the notion of ‘positional parental socialisation’ with the theoretical assumptions of realistic conflict theory, we expect that economic characteristics of the household have an impact on adolescents’ attitude towards equal opportunities for foreigners.

From a dynamic perspective, we argue that a deterioration of the economic situation of the household implies less favourable living conditions, which can subsequently evoke higher levels of experienced or perceived competition. These higher levels of competition experienced from foreigners can in turn increase rejection of equal opportunities among all members of the household. As adolescents are especially prone to attitude change, we expect these negative changes in the household’s living conditions to increase rejection of equal opportunities among adolescents in particular.

One of the key characteristics that ensures an economically more secure position is a household’s income. Compared to their less privileged counterparts, households with a higher income are to a higher extent insulated from economic pressures induced by foreign competitors, for instance on the labour market [41, 42]. When experiencing a decrease in income, the economically secure and priviliged position of the household becomes pressured. Consequently, these pressures might evoke experiences of ethnic threat and increase rejection of equal opportunities. Based on studies accentuating the effect of the shared economic position [35, 36], we expect that a decrease in income has an effect on all members of a household, adolescents included. Therefore, we expect that:

1. H4: If the household income decreases, adolescents’ likelihood to reject equal opportunities for foreigners increases.

Another economic change that could have a large impact on the economically secure position of the household concerns a transition to unemployment of (one of) the household’s main earner(s). Being confronted with unemployment is often unforeseen and involuntarily, and can disrupt a households’ accustomed way of living [8, 40]. A transition to unemployment could lead to an increase in experienced ethnic competition, as one of the household members suddenly has to compete with outgroups on the labour market over jobs at offer [43]. The confrontation with unemployment and increased level of competition among the unemployed household member(s), can place the household and all of its members in an economically stressful position. These expectations lead to the following hypothesis:

1. H5: If the household is confronted with unemployment, adolescents’ likelihood to reject equal opportunities for foreigners increases.

Again, we expect that not only objective and tangible transitions in the household can lead to a change in rejection of equal opportunities, we also argue that perceptions of a worsened socio-economic position in itself can evoke rejection of equal opportunities [33]. If members of a household, parents in particular, perceive an increase in subjective economic dissatisfaction, a feeling of economic insecurity can be generated and discussed within the household. Though actual competition may be absent, household members could still perceive higher levels of competition as a consequence [2]. Because such subjective economic dissatisfaction creates a less favourable living environment in a household, we expect that:

1. H6: If the household’s subjective economic dissatisfaction increases, adolescents’ likelihood to reject equal opportunities for foreigners increases.

Not all households may be equally affected by changes in the economic situation. In general, foreigners in Switzerland have a less advantaged socio-economic status as compared to the native population [44]. Therefore, we expect that especially those who live in a household that is already positioned in the relatively lower socio-economic strata of society are more likely to find themselves in a situation of competition over scarce resources with foreigners as a consequence of a deterioration in their household’s socio-economic status [2, 45]. We hence expect that if the household’s economic position resembles the position of foreigners to a greater extent, an experienced or perceived change in the economic consequences has a greater impact on the likelihood to become less supportive of equal opportunities for foreigners. This results in the following hypothesis:

1. H7: The effects of changes in a household’s economic situations on the change in adolescents’ likelihood to reject equal opportunities for foreigners (H4-H6) are stronger for those living in a household with a more disadvantageous socio-economic position.

In research studying the impact of the household context on adolescents’ intergroup attitudes, parents’ role as direct socialising agents is emphasised. Already in 1954, Allport [46] argued that children’s intergroup attitudes are shaped by their parents’ intergroup attitudes. Parents are argued to create a learning environment for their children, in which their children are taught basic social norms and values [47]. Children are subsequently suggested to imitate their parents via modelling by adapting to their parents’ attitudes, resulting in intergenerational transmission of intergroup attitudes [23, 48]. Though evidence for the existence of parental socialisation of intergroup attitudes is mixed, the meta-analysis of Degner and Dalege [49] shows that there is a medium-sized average association between intergroup attitudes of parents and their offspring, and that this effect is particularly visible among adolescents.

As we set out to test adolescents’ rejection of equal opportunities dynamically, we expect that a change in parents’ rejection of equal opportunities is associated with a change in adolescents’ rejection of equal opportunities. Whereas we presume that attitudes are less vulnerable to change after adolescence [19], we argue that the changes in parents’ rejection of equal opportunities that do occur may evoke a change in adolescents’ rejection of equal opportunities. Longitudinal studies on direct parental socialisation indeed found support for linked changes in parents’ intergroup attitudes to changes in their children’s intergroup attitudes [23, 39, 50]. We therefore hypothesise that:

1. H8: If parents’ likelihood to reject equal opportunities for foreigners increases, adolescents’ likelihood to reject equal opportunities for foreigners increases as well.

3.1 Data

In this study, we tested our hypotheses for the Swiss case. Switzerland is a country with a relatively high share of immigrants, as about 26% of its population is foreign-born. A large part of this foreign-born population has a European background (OECD, 2021). After World War II, mostly unqualified manual workers migrated to Switzerland, followed by their families. Since the 1990s, immigration policies have become more restrictive for these labour migrants, while favouring higher educated European immigrants. Comparable to other west-European countries, immigration and integration have become highly politicized topics in the political and public debate, reflected in the success of the populist radical-right Swiss People’s Party [51, 52]. By contrast, Switzerland is not an EU-member and cannot be categorised easily nor clearly into one of the typical welfare regimes in the EU because of its rather fragmented welfare state [53].

The Swiss Household Panel, collected by FORS center, is used to test our hypotheses. More information on the data collection procedure and consent concerning participant recruitment can be found in the Swiss Household Panel User Guide [54] or at www.forscenter.ch. Because we use data that is publicly available, we needed no further approval of an institutional review board for the data collection. This longitudinal panel is based on a stratified, random sample of private households in Switzerland. Data were collected by means of three different questionnaires: a household grid questionnaire to assess the household composition; a household questionnaire, filled in by the reference person of the household; and an individual questionnaire, administered to all members of the household aged 14 years and older. Interviews were initially conducted telephonically only, but since 2010 face-to-face interviews and web-based questionnaires were available for those reluctant to participate (1.03% of the observations). Response rates were relatively high; about sixty percent of the approached people responded to the initial questionnaire and about eighty percent of the respondents participated in follow-up surveys. Those who attrite were somewhat more often younger people, males, and lower-educated people [54–56]. Refreshment samples were added to restore the representativity of the panel. For more information on the fieldwork approach, response rates and attrition, and data documentation, see Voorpostel et al. [54].

The data used in this study covers the 1999–2017 period. Data gathered in 2010, 2012, 2013, 2015, and 2016 were not taken into account in this study, as our dependent variable was not included in these waves. Because of our focus on adolescents and young adults, we only included observations of respondents aged 14 to 25. Respondents that we could select as having a non-Swiss nationality were not incorporated in this study. Analyses were performed on two samples: to test the first three hypotheses we included all adolescents; to test the impact of the household context we only included adolescents who lived with (one of) their parents. For the latter group, we matched for each separate survey year adolescents’ individual data to data of the household questionnaire and (if available) individual data of both the mother and father. After deleting observations with missing values on the dependent or independent variables and after deleting respondents who participated only once, our final sample consisted of 9,530 observations from 2,353 respondents. On average, respondents participated 4.05 times. Of this final sample, 2,125 respondents lived with (one of) their parents, of whom we have 8,948 observations.

3.2 Dependent variable

We used the question ‘Are you in favour of Switzerland offering foreigners the same opportunities as those offered to Swiss citizens, or in favour of Switzerland offering Swiss citizens better opportunities?’ to measure whether adolescents reject equal opportunities for foreigners. According to Pecoraro and Ruedin [51], this measurement of rejection of equal opportunities for foreigners can be understood as an unfavourable attitude towards foreigners more generally. Those who answered ‘in favour of better opportunities for Swiss citizens’ clearly surpass and reject the option to provide equal opportunities to foreigners and hence were ascribed a score of 1, whereas those who answered ‘in favour of equality of opportunities’ were ascribed a score of 0. A score of 0 was also ascribed to those who answered ‘neither’ (6.03% of the observations). Robustness checks in which the ‘neither’ category was dropped or when these respondents were ascribed a score of 1 instead of 0 did not yield substantially different conclusions (see”S1 Table”). The descriptive statistics (see “Table 1”) demonstrate that 28.10% of the observations in our sample were in favour of offering Swiss citizens better opportunities than foreigners.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0296883.t001

3.3 Independent variables

To get an insight in whether rejection of equal opportunities changes as adolescents enter the labour market, we first created a variable that captures adolescents’ main activity. Each year, respondents were asked if they were currently studying at a school. Those who indicated that they were not, were considered as having left school and having entered the labour market. For these respondents, we checked what their actual occupation was. The eleven categories from which respondents could choose were reduced to three categories: employed, unemployed, and non-employed. The fourth category encompassed those who are in school. An additional fifth category was created for those with a missing value on this variable.

Those respondents who did not finish education yet were asked what their current type of education was. Respondents could choose between 17 categories that were recoded according to a coding scheme of Bergman et al. [57] to the six educational levels in Switzerland: primary education, secondary education without Matura (i.e. Swiss high school diploma), secondary education with Matura, secondary vocational education, tertiary vocational education, and university. Those who indicated to be in secondary or tertiary vocational education were, given the strong role of learning practical labour market competences and the importance of internships in these types of education, considered to be in vocational training.

We measured adolescents’ subjective economic dissatisfaction with the following question: ‘Overall how satisfied are you with your financial situation?’. Respondents could answer on an eleven-point scale ranging from 0 ("not at all satisfied”) to 10 ("completely satisfied"). We recoded the scale so a higher score denotes a higher level of dissatisfaction.

The household income was measured by the equivalised Swiss Conference for Social Welfare (SKOS)-scale, which takes into account the size and the composition of the household. For each separate year, we constructed deciles. For each of these deciles a dummy-variable was created, as well as an additional dummy-variable for respondents with a missing score.

Parents’ working status was used to indicate whether a household was confronted with unemployment. If either adolescents’ father or mother indicated to be unemployed, respondents got ascribed a score of 1. If both parents in the household indicated to be employed and/or that they were not part of the labour force, respondents got ascribed a score of 0.

The item used to measure the household’s subjective economic dissatisfaction is similar to the one measuring adolescent’s subjective economic dissatisfaction. Here, however, the reference person of the household was asked how satisfied (s)he is with the financial situation of the household. A higher score on the eleven-point scale indicated a higher level of dissatisfaction.

Parents’ rejection of equal opportunities was measured in the same way as adolescents’ rejection of equal opportunities. For both the mother and the father, a variable was created where those in favour of better opportunities for Swiss citizens were distinguished from those who did not. Additional dummy-variables were created for respondents who lived with (one of) their parents, but whose mother or father had a missing score on this variable.

3.4 Control variables

In our analyses, we controlled for the household composition. Here, we distinguished between adolescents living with two parents, adolescents living with one parent, adolescents living alone, adolescents living with a partner and/or child, and other household types (e.g. households with other related family members). We also control for the year of the survey.

3.5 Methods

To get an insight into the individual change in rejection of equal opportunities for foreigners, we plot predicted (population averaged) margins. We also present the percentage individuals who experienced a within-individual change in their rejection of equal opportunities and the proportion of individuals who remained stable between succeeding waves.

To test our dynamic hypotheses, we make use of logistic fixed effects models [58]. Fixed effects models enable us to investigate within-person variation in rejection of equal opportunities, as these models examine whether a change in individuals’ rejection of equal opportunities is related to changes in their economic situation. Because of its focus on within-individual variation, all unobserved time-invariant heterogeneity between individuals can be partialed out in fixed effects models, which contributes to the reliability of the estimated effects. Since fixed effects models rely on within-person variation only when estimating the effects of changes in the economic situation on changes in attitudes towards equal opportunities, the analyses are based on respondents whose attitude towards equal opportunities for foreigners changes over time. To minimise threats of identification, we make use of a large sample of as much as 14 waves. Though observations of transitions to unemployment were still rather infrequent, our strategy of pooling as much as 14 waves lowered the risk of power issues as we aimed at incorporating as much transitions as possible on both the dependent and independent variable(s) that we could analyse. Given the dichotomous character of our dependent variable, we use logistic fixed effects models that estimate the likelihood of having a negative attitude towards equal opportunities for foreigners as determined by its logistic cumulative distribution.

In the first logistic fixed effects model, we estimate the likelihood to reject equal opportunities for all adolescents aged 14 to 25. As we are primarily interested in the way in which attitudes towards equal opportunities are affected by transitions to the labour market or to vocational training during adolescence, we apply a modelling strategy in which we differentiate between origin and destination categories (for more information on this modelling strategy, see [59]). That is, we are primarily interested in for instance the transition from being in school to being employed, and not the other way around. We therefore transformed the variables measuring adolescents’ main activity and educational attainment into several variables capturing the transition of interest. In the second logistic fixed effects model, we estimate the impact of changes in the household context. As these changes can only be estimated for those who live with (one of) their parents, we excluded respondents who do not live with (one of) their parents from this model. A test for collinearity showed that there was no multicollinearity between the variables in our models.

Additionally, we performed our analyses separately for adolescents living in households that could be argued to be more likely to end up in a situation of ethnic competition as a result of a deterioration of the economic situation: adolescents who live in a household with a lower rather than higher income or whose parents have a lower rather than higher educational attainment. Adolescents were denoted as living in a household with a lower income when the mean income decile of the household was below the fifth decile, representing the respondents in the 40% lowest household incomes. Whether parents were considered having a lower educational attainment was based on the number of years needed to complete the highest level of educational achieved. Using education in years enabled us to calculate a more fine-grained mean score of both parents for each respondent. In case information of only one of the parents was available, we used the score of just that parent. Parents were denoted as having a lower educational attainment when they had completed their highest level of education achieved in 12 years or less, which can be translated to an ISCED (International Standard Classification of Education 2011) level of 3, i.e. an educational attainment of upper secondary education or lower [60]. Though this cut-off point might seem relatively high, it represents about 40% of the respondents as well. We performed multiple robustness checks where we examined whether our outcomes were affected by our choice for a specific cut-off point. In most cases, changing the cut-off point did not yield different substantial conclusions. If the cut-off points that we chose did affect our conclusions, we discuss the outcomes of these robustness checks in the results section.

4.1 Descriptive results

In “Fig 1” adolescents’ overall trend in rejection of equal opportunities among the members of the panel is presented. Rejection of equal opportunities was highest in 1999. Between 1999 and 2000, there is a decrease in rejection of equal opportunities for foreigners in general. From 2000 onwards, rejection of equal opportunities was found to be rather stable, as we could not observe significant changes in rejection of equal opportunities between succeeding years.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Margins plot on rejection of equal opportunities among adolescents.

Source: Swiss Household Panel (SHP) 1999–2017, N = 9,530 observations of 2,353 respondents.

https://doi.org/10.1371/journal.pone.0296883.g001

To get a better insight in the magnitude of change in rejection of equal opportunities among adolescents, we made an overview of the share of adolescents that experienced a within-individual change in rejection of equal opportunities for foreigners, as presented in “Table 2”. When comparing adolescents’ responses between succeeding waves, we see about 75% to 85% of the respondents remained stable in their rejection of equal opportunities. The individual change in rejection of equal opportunities was highest between 2000 and 2001; 23.39% changed in their rejection of equal opportunities during this period. Additional descriptive analyses furthermore showed that 66.52% of the respondents remained stable in their attitudes during all waves of participation. All in all, change in rejection of equal opportunities among adolescents thus is rather small.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Percentage of adolescents that experiences a within-individual change in rejection of equal opportunities for foreigners between succeeding waves.

https://doi.org/10.1371/journal.pone.0296883.t002

4.2 Multivariate analyses on changes in adolescents’ economic situation

The results of the logistic fixed effects models on the change in rejection of equal opportunities for foreigners are presented in “Table 3”. In the first model the effects of the individual characteristics are estimated for all adolescents. In this model we estimated the effect of the transition from school to employment or to unemployment on the likelihood to reject equal opportunities. Transitions from primary school to the labour market have been set to missing because of irregular patterns in the panel data. Our results indicate that the transition from school to employment induced a significant decrease in rejection of equal opportunities. This effect holds when controlling for household characteristics in the second model. For those who made the transition from school to unemployment we did not find a significant impact on the likelihood to reject equal opportunities. A closer look at this effect and its standard error demonstrates that the confidence interval of this effect is rather large (0.179 ± (1.96 x 0.447) = [-0.697; 1.055]). In comparison to the other estimates, the relatively high standard error and, accordingly, the rather wide confidence interval of the estimate for the transition to unemployment may indicate a relatively low power to test this effect. Indeed, only 46 respondents (0.48%) make the transition to unemployment. Based on these analyses, we therefore cannot support or refute our hypotheses on the transition to unemployment. For now, we can only state that finding a job decreases the likelihood to reject equal opportunities. Still, these findings are not in line with our first hypothesis that entering the labour market induced rejection of equal opportunities for foreigners among adolescents. In”S1 Text”, a more elaborate analysis can be found in which we estimated the effect of transitions from different educational levels to employment on changes in attitudes towards equal opportunities for foreigners. These additional analyses show that none of the specific transitions induce an increase in rejection of equal opportunities.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Logistic fixed effects analysis on the likelihood to reject equal opportunities for foreigners.

https://doi.org/10.1371/journal.pone.0296883.t003

In the first model of “Table 3”, we also estimated the effect of the transition to vocational education on the likelihood to reject equal opportunities for foreigners. We found that those who made the transition to secondary vocational education did not significantly change in their rejection of equal opportunities. We did find a significant effect of the transition to tertiary vocational education, but this effect was opposed to what we expected. That is, we found that entering tertiary vocational education led to a decreased likelihood to reject equal opportunities instead of an increased likelihood. Additional analyses (see”S1 Text”) show that this negative significant effect is caused primarily by those who make the transition from secondary vocational training to tertiary vocational training, and not by those who make the transition from secondary education with Maturity. Considering that those who make the transition from secondary vocational training to tertiary vocational training already have had experience with vocational training and entering an apprenticeship, we could interpret this negative effect of tertiary vocational training on the change in rejection of equal opportunities as caused by entering a higher level of education and not per se as a result of entering an apprenticeship. All in all, we do not find support for our second hypothesis that states that entering an apprenticeship induces rejection of equal opportunities for foreigners.

Model 1 also includes the effect of adolescents’ financial dissatisfaction on negative attitudes towards equal opportunities for foreigners. Though we hypothesised that an increase in adolescent’s financial dissatisfaction would lead to an increased likelihood to support better opportunities for Swiss people, we found no significant effect. We performed an additional analysis to test whether results are different during times of economic recession, and found a significant effect of financial dissatisfaction on the likelihood to develop a negative attitude towards foreigners during the economic crisis (see”S2 Table”). Thus, those who became more dissatisfied with their economic situation during the recession, became more likely to reject equal opportunities for foreigners. Our third hypothesis is therefore not fully rejected.

4.3 Multivariate analyses on the impact of the household context

To test the impact of the household context, we focus only on adolescents who live with (one of) their parents. In the second model of “Table 3”, we removed those who do not live with their parents from our sample and included various household characteristics. A decrease in household income was expected to increase the likelihood to reject equal opportunities for foreigners. Mostly, we found effects that were statistically not significant. Compared to a change to the first income decile, we found that a change to the fourth and sixth income decile increased the likelihood to reject equal opportunities. Based on our hypothesis, we would, however, expect that the likelihood to reject equal opportunities would decrease as one’s household experiences a change towards a higher income decile compared to a change towards the lowest decile. Changing the reference category furthermore demonstrated that there was a positive effect of changing to the fourth category compared to changing to the second category, but no other significant effects of income were found when changing the reference categories. Taken together, these findings point at a lack of support for our fourth hypothesis.

Parental unemployment was also included in the second model to test whether a household’s confrontation with unemployment led to an increase in rejection of equal opportunities for foreigners among adolescents. The effect of parental unemployment was not significant. Again, the standard error was relatively high in comparison to all other effects. Moreover, our models rely here as well on a relatively low number of respondents that lived in a household that was confronted with a transition to unemployment (70 transitions to unemployment among (one of the) parents). This may indicate a lack of power, and therefore we cannot support or refute our fifth hypothesis. Moreover, our sixth hypothesis on the effect of changes in the household’s financial dissatisfaction is not supported. A change in a household’s financial dissatisfaction did not lead to an increased likelihood among adolescents to reject equal opportunities for foreigners.

To make sure that the absence of significant effects of the changes in economic household characteristics is not due to the inclusion of parents’ rejection of equal opportunities in this model, we also estimated a model in which we left out mothers’ and fathers’ attitudes towards equal opportunities for foreigners (not shown). This model does not show different outcomes as regards the effects of the economic household characteristics.

Additionally, we checked whether specific groups–adolescents who live in a household with a relatively lower income or whose parents have a relatively lower educational attainment–are to greater extent affected by a deterioration of the economic situation (see “S1 Appendix”). We expected that these specific groups were more likely to develop rejection of equal opportunities for foreigners as a result of a deteriorating income, a confrontation with unemployment, or an increase in financial dissatisfaction of the household. Also in these subgroups, changes in the household’s income, employment situation, or financial dissatisfaction did not evoke a significant change in adolescent’s likelihood to reject equal opportunities for foreigners. Robustness analyses showed that changing the cut-off points as to who is considered to live in household with a relatively lower income or with relatively lower educated parents did not yield different conclusions. Therefore, we reject hypothesis 7.

In the second model of “Table 3”, changes in mother’s and father’s rejection of equal opportunities are included as well. We found that a change in both mothers’ and fathers’ rejection of equal opportunities leads to a significant change in adolescents’ rejection of equal opportunities. As parents changed towards a negative stance towards equal opportunities, the likelihood for adolescents to reject equal opportunities increased concurrently.

Lastly, we performed a Bayesian analysis to reduce possible concerns regarding our rather small sample size for some categories [61]. Given the small number of adolescents and parents that made a transition to unemployment, we were particularly interested in the outcomes of this Bayesian analysis regarding our first and fifth hypothesis. The conclusions drawn from the Bayesian analysis (available upon request) are largely in line with the conclusions from our frequentist analysis as presented in “Table 3”. However, regarding adolescents’ transition to unemployment, we found a significant positive effect on the likelihood to reject equal opportunities for those living with their parents. Though we remain cautious considering our small N, this effect alludes to a change towards a greater likelihood to reject equal opportunities for foreigners due to unemployment, as was expected in Hypothesis 5. For parental unemployment, we found an effect in the opposite direction to what we expected. The results of this Bayesian analysis also point at support for our hypothesis regarding financial dissatisfaction within the household. Though we prefer to either support or refute hypotheses based on more conservative and stringent analyses, these Bayesian results provide us with additional insights from which we may derive that the null-hypothesis that there is no effect should not immediately accepted for the hypotheses that we found no support for in our initial analyses.