To Örebro University

This thesis consists of four papers, and has as its central theme the accumulation of welfare problems and social exclusion. We use Swedish data and all analyses are based on individuals of working age. We perform longitudinal analyses to scrutinize the accumulation of disadvantages over the individual life courses as well as to detect the general trends in social exclusion occurrence in Swedish society during the past two decades.

In Paper I, in an analysis of social exclusion among immigrants in Sweden, we find that immigrants suffer more often from social exclusion than native Swedes do. We also find that even if the accumulation of welfare problems is more common among immigrants than native Swedes, the connections between welfare disadvantages are stronger among Swedes. Furthermore, a logistic regression analysis revealed that time spent in Sweden decreases the risk of social exclusion among immigrants. However, even though we control for several demographic variables, human capital indicators and socio-economic class, the odds for social exclusion are still greater for immigrants than for native Swedes. Some form of discrimination can therefore not be excluded.

Paper II is co-written with Björn Halleröd. This paper involves a longitudinal analysis of the accumulation of closely related welfare disadvantages, showing that the initial deprivation increases over time. Latent growth curve models reveal that a high initial deprivation is related to low socio-economic class and being single. It is also shown that a high initial deprivation decreases the probability of upward class-mobility as well as the probability of deprived singles becoming cohabiting. Moreover, a high initial deprivation increases the risk that couples will experience a household break-up.

In Paper III, we perform a longitudinal analysis of social exclusion in Sweden during the period 1979-2003, in which several logistic regression models for panel data are fitted to our data. We find no support that immigrants have been better integrated into Swedish society over time from the perspective of social exclusion risk. Instead, there are weak signs that integration has become worse. We also find weak signs that the higher social exclusion risk that men have relative to women has decreased during the past two decades. Furthermore, comparing with couples without children, the odds for social exclusion among singles with children have increased and the odds for couples with children have decreased during the period 1979-2003.

Paper IV utilizes latent class factor models to scrutinize the connections between welfare problems and a set of demographic variables, human capital indicators and socio-economic class. We find that welfare problems do cluster. Our results also support several of the findings in the previous paper. Family type, especially being single or living in a relationship, makes a clear difference in the propensity to accumulate welfare problems. Furthermore, immigrants characterize the factors with a high problem accumulation. Additionally, there is no general difference between the sexes in the problem accumulation itself, but experiences of threat or violence and having sleeping problems seem to be more often related to being a woman, whereas the lack of a close friend is most often related to being a man.

To conclude, this thesis reveals several interesting facts concerning the accumulation of welfare problems and social exclusion in Sweden. Considering the implications for policy, the situations of immigrants and single parents need to be underlined. That is, the integration of immigrants should be given more emphasis and measures should be taken to support single parents as well as to promote a discussion on how to make relationships last.

This paper uses longitudinal Swedish data and investigates the accumulation and persistence of welfare problems over time. The data set was compiled in a first wave in 1979, a second wave in 1986–1987, a third wave in 1994–1995, and a fourth wave in 2002–2003 (N = 7,967). First, the results demonstrate that all welfare problems have a tendency to become persistent. For example, the probability of suffering from the lack of a close friend in the fourth wave was 57.1 % if the individual suffered from this welfare problem in the third wave. Second, economic problems constituted the welfare problem with the most associations with other welfare problems. Third, the accumulation of welfare problems significantly differed between different categories of individuals. For example, the accumulation of welfare problems was higher in men compared with women, immigrants compared with native Swedes, single individuals compared with individuals in couples, and poorly educated individuals compared with highly educated individuals. Finally, longitudinal analyses indicated individuals in certain categories have experienced cumulative disadvantages in welfare even during periods when individuals in other categories have experienced a positive trend in welfare. One such example is single parents during the period between 1979 and 1986–1987.

In this paper, we perform a latent class factor analysis of a panel that involves two waves of data from an annual survey of living conditions in Sweden that were gathered in the years 1994-1995 and 2002-2003. We follow the same 3,149 individuals over both waves, describing them by sex, age group, family type, nationality background, education level and socio-economic class. Further, since the welfare disadvantages included in the analysis are chronic unemployment, economic problems, health problems, experiences of threat or violence, crowded housing, lack of a close friend and sleeping problems, we also have data on exactly which welfare problems each individual in the data set suffers from. In the empirical analysis, latent class factor analysis provides us not only with information on which individual characteristics that are important in the accumulation of welfare problems, but also gives us information on which disadvantages in fact are accumulated. First, we find that welfare problems do cluster. Second, the welfare problems that most often appear at a factor level with several disadvantages are experiences of threat or violence and sleeping problems. Finally, being an immigrant and being single are individual characteristics that turn up most often in factor levels with problem accumulation, whereas there is no distinctive difference between the sexes. However, women seem to be more prone to suffer from experiences of threat or violence and sleeping problems, while men are more likely to suffer from lack of a close friend.

We examined the occurrence and coexistence of externalising and internalising problem behaviour among Swedish boys and girls, and investigated whether there are differences in the accumulation of problems among adolescents with different behavioural, demographic and social characteristics. The results are discussed in the context of self-salience schemas. The source material comprised all ninth grade pupils in a province in central Sweden in 2008 (N = 3,095). First, girls were found to be more prone to experience internalising problem behaviour, whereas no sex differences were found regarding externalising problem behaviour. Second, multidimensional scaling maps indicated that, compared with boys, self-esteem and the PsychoSomatic Problem (PSP) scale among girls more closely indicated externalising problem behaviour components, whereas anxiety was situated far from the other externalising problem behaviour indicators. Finally, linear regression analyses indicated family type as the primary explanatory background factor for externalising problem behaviour and economic hardship as the primary explanatory background factor for internalising problem behaviour.

In this paper, we examine the risk of social exclusion among the Swedish population from a longitudinal perspective. In the empirical analysis, a person is considered socially excluded if he or she suffers from at least two of the following six welfare problems: chronic unemployment, economic problems, health problems, experiences of threat or violence, crowded housing and lack of interpersonal relationships. Our three main findings are as follows: There is no evidence that immigrants have been better integrated into Swedish society over time from the perspective of social exclusion risk. Instead, there are weak signs that their situation has become worse. Further, even though men are worse off than women as regards the odds for social exclusion, there are weak signs that their relative situation has improved over time. Finally, compared to couples without children, there is clear evidence that the odds for social exclusion for singles with children have increased over time and that the odds for social exclusion for couples with children have decreased over time. We can, therefore, conclude that among these groups, the inequality has increased over time. To be able to make these conclusions, we have fitted several specifications of a logistic regression model with random effects for panel data to our data set.

This paper examines psycho-social distress among middle-aged and elderly Swedes. We analysed data on 3221 individuals who were 55 to 99 years old. Based on a latent class analysis, we identified four latent classes. Two classes were associated with higher levels of psycho-social problem accumulation. The class with the lowest level of problem accumulation contained the greatest number of individuals, whereas the classes with the highest level of psycho-social distress contained the least number of individuals. The analysis showed that being a man, being married, being a native Swede, or having several hobbies was associated with a low likelihood of belonging to a latent class that was characterised by psycho-social distress. Moreover, being a woman, being between 55 and 65 years of age, or being a widow was associated with a high likelihood of belonging to a latent class that was characterised by the highest levels of problem accumulation.

In this paper, we examine the risk of social exclusion among the Swedish population from a longitudinal perspective. To achieve this aim we use an unbalanced panel involving four waves. The first wave was gathered in 1979, the second during the period 1986-87, the third during the period 1994-95, and the fourth during the period 2002-03, giving a total number of respondents of 7,967. In the empirical analysis, a person is considered socially excluded if he or she suffers from at least two of the following six welfare problems: long-term unemployment, economic problems, health problems, experiences of threat or violence, crowded housing, and lack of interpersonal relationships. As explanatory variables for the risk of social exclusion, the following variables are used: sex, age group, family type, nationality background, education level and socio-economic class. We have fitted several specifications of a logistic regression model with random effects for panel data to our data set. Our main findings are as follows: First, there is no evidence that immigrants have been better integrated into Swedish society over time from the perspective of social exclusion risk. Instead, there are weak signs that integration has become worse. Second, even though men are worse off than women as regards the odds for social exclusion, there are weak signs that their relative situation has improved over time. Finally, there is clear evidence that the odds for exclusion for singles with children have increased over time and that the odds for exclusion for couples with children have decreased over time.

In this paper, we examine social exclusion among immigrants in Sweden. The groups under scrutiny are native Swedes, first generation Swedes, naturalized Swedes, Nordic citizens and non-Nordic citizens. Specifically, because one goal of the welfare state is to break the connections between different welfare problems, we investigate the associations between welfare problems among different immigrant groups as well as among native Swedes. We find that the accumulation of welfare problems is higher among immigrant groups, but that the correlations between welfare problems are strongest among Swedes. Finally, we analyse social exclusion among immigrants using a logistic regression analysis. Because a regression analysis with explanatory variables such as demographic variables, human capital indicators and socioeconomic class cannot explain the difference between immigrants and native Swedes, it appears that discrimination is a probable explanation.

To foster a deeper understanding of the mechanisms behind inequality in society, it is crucial to work with well-defined concepts associated with such mechanisms. The aim of this paper is to define cumulative (dis) advantage and the Matthew effect. We argue that cumulative (dis) advantage is an intra-individual micro-level phenomenon, that the Matthew effect is an inter-individual macro-level phenomenon and that an appropriate measure of the Matthew effect focuses on the mechanism or dynamic process that generates inequality. The Matthew mechanism is, therefore, a better name for the phenomenon, where we provide a novel measure of the mechanism, including a proof-of-principle analysis using disposable personal income data. Finally, because socio-economic theory should be able to explain cumulative (dis) advantage and the Matthew mechanism when they are detected in data, we discuss the types of models that may explain the phenomena. We argue that interactions-based models in the literature traditions of analytical sociology and statistical mechanics serve this purpose.

The paper's first aim is to clarify the differences and relationships between cumulative advantage/disadvantage and the Matthew effect. Its second aim, which is its main contribution, is not only to present a new measure of the Matthew effect, but also to show how to estimate this effect from data and how to make statistical inference. We argue that one should utilize the positivity of the natural logarithm of the largest generalized eigenvalue for a non-linear dynamic process as evidence when claiming that the Matthew effect is present in the dynamic process that generates individuals' socio-economic life-courses. Thus, our measure of the Matthew effect focuses on the dynamic process that generates socio-economic inequality and not on the outcome of this process.

The first aim of this paper is to clarify the differences and relationships between cumulative advantage/disadvantage and the Matthew effect. Its second aim, which is also its main contribution, is not only to present a new measure of the Matthew effect, but also to show how to estimate this effect from data and how to make statistical inference. We argue that one should utilize the positivity of the natural logarithm of the largest generalized eigenvalue for a non-linear dynamic process as evidence when claiming that the Matthew effect is present in the dynamic process that generates individuals’ socio-economic life-courses. Thus, our measure of the Matthew effect focuses on the dynamic process that generates socio-economic inequality and not on the outcome of this process.

We show that the Matthew effect, or Matthew mechanism, was present in the artificial cultural market Music Lab when social influence between individuals was allowed, whereas this was not the case when social influence was not allowed. We also sketch on a class of social network models, derived from social influence theory, that may gener-ate the Matthew effect. Thus, we propose a theoretical framework that may explain why the most popular songs were much more popular, and the least popular songs were much less popular, than when disallowing social influence between individuals.

We show that the Matthew effect, or Matthew mechanism, was present in the artificial cultural market Music Lab in one-fourth of the "worlds" when social influence between individuals was allowed, whereas this effect was not present in the "world" that disallowed social influence between individuals. We also sketch on a class of social network models, derived from social influence theory, that may generate the Matthew effect. Thus, we propose a theoretical framework that may explain why the most popular songs could be much more popular, and the least popular songs could be much less popular, than when disallowing social influence between individuals.

In this chapter, we apply the expectancy-value model of motivation, particularly the family socialization aspect of the model (Eccles (Parsons) et al., 1983; Eccles, 1994, 2007; Wigfield & Eccles, 2002) to address a number of key questions regarding gender differences in adult attainment, in particular educational attainment. When some individuals in the work force of today were children, what kinds of expectations did they have for themselves? What expectations did their parents have for them? Did these expectations vary for girls and boys? Were parents' expectations about their children's future education related to the actual education that these adolescents later attained in midlife? How did the child's academic ability and characteristics of the family figure into this picture? We present original empirical findings, drawing on data collected for a Swedish longitudinal study that spans from childhood to middle adulthood. In line with the expectancy-value model of motivation, the family's socioeconomic status (SES) was identified as an important predictor of several outcomes. Consistent with the model, for both genders, the family's SES and parental educational expectations in middle adolescence predicted middle adult educational attainment. The importance of grades differed by gender in that the mathematics grade was a statistically significant predictor of middle adult educational attainment for males, while for females grades in Swedish were a statistically significant predictor of middle adult educational attainment. In this chapter, we situated these study findings in the wider pertinent scholarly literature and discussed the implications of our results as they might relate to efforts to promote equitable and optimal life chances for the current generation of European girls and boys.

Despite a vast literature investigating the effects of childhood living conditions, there are few studies that investigate the significance of the timing of economic hardship during childhood. This study used the 1987 Finnish Birth Cohort - including all children born in Finland in 1987 (N=59,476) to explore the patterns of childhood economic hardship. Sequence analysis was utilized to extract the patterns of economic hardship, measured as social assistance recipiency (SA). This resulted in five clusters: No SA at all; Occasional SA; Occasional SA, more during early 2000’s; Occasional SA, more during the 1990’s, and Moderate to heavy SA throughout. We then examined the relationship between childhood economic hardship patterns and five different later adolescence and early adulthood outcomes: Cohort members own adulthood SA recipiency, Psychiatric care, Criminal convictions, Early school leaving and Teenage pregnancy (for girls). We found strong relationships between all these outcomes and childhood economic hardship patterns. For example, regression analyses with No SA at all as a reference category showed large OR, especially regarding cohort members own adulthood SA. It seemed also that later childhood economic hardship was more harmful than the earlier economic hardship regarding adulthood SA recipiency.

This paper examines the development of school burnout among Finnish youth aged 16-18, specifically with regard to the following three components: a cynical attitude toward the school, feelings of inadequacy as a student, and exhaustion at school. There is evidence of an increase in all three components over time, but only among students on the academic track. There appear to be differences in burnout levels between those who drop out from school and those who do not, and the risk of an individual with high levels of cynicism or feelings of inadequacy dropping out is clearly higher than among those who score low on these two components. When various explanatory variables are controlled for, cynicism still remains a significant factor explaining drop out, and a low grade point average appears to be a major explanatory variable for school dropout. This study is especially interesting in the context of Finland, known for its equality-striving and high-quality educational system. Finnish youth, compared with youth in many other countries, nevertheless have a low level of in-school well-being. We use the Finnish Educational Transitions data (N = 878) collected in four waves, the first three on an annual basis and the fourth 5 years after the first one.

The central aim of the present paper is to analyse the degree to which welfare problems accumulate over time and to what extent such an accumulation is related to class position and household formation. We utilize longitudinal data from the Swedish Survey of Living Conditions, following a panel that was first interviewed in 1979 and, thereafter, re-interviewed three times (1986–87, 1994–95 and 2002–03). We use structural equation modelling to extract latent deprivation indexes based on seven manifest indicators. Thereafter, the indexes are used as input values in a latent growth curve (LGC) model when estimating an intercept variable and a slope variable. We finally estimate two structural models. In the first model, the relation over time between class position, class mobility and deprivation are estimated, and the second model deals with the relationship between household types, change of household type and deprivation. Several interesting results can be reported. We find a strong relationship between the deprivation indexes over time. The LGC model can also confirm a positive relationship between deprivation intercept and deprivation slope. That is, individuals who score high on the deprivation index from the beginning are increasingly prone to accumulate additional welfare problems over time. The analysis also reveals a clear class gradient as well as effects of class mobility. In addition, it is shown that deprivation affects class mobility, meaning that we can confirm selection effects. Also household constellation and changes of household type are closely connected to deprivation and changes in deprivation over time. The analysis reveals selection effects even in this case.

This report analyses a large-scale survey of living conditions among immigrants in Norway. Data were collected by Statistics Norway in 2016 and covered 12 of the largest immigrant groups in Norway. This aim of the present report is analyse the kinds of welfare problems experienced by immigrants. Which immigrants accumulate most welfare problems? What are the main predictors of welfare problems among immigrants in Norway? How does the accumulation of welfare problems relate to general life satisfaction among immigrants? In addition, we compare the accumulation of welfare problems among immigrants with the total population.

An early draft of this report was presented at the International Forum on Migration Statistics 2018, in Paris in January. A more recent draft was presented internally at a seminar for migration researchers at Oslo Metropolitan University. We would like to thank participants at both events for constructive and valuable feedback. This project was commissioned and financed by the Directorate of Integration and Diversity (IMDi). We would like to thank IMDi for making possible this interesting and challenging project, and especially Anja Wedde Sveen and Eivind Hageberg, both at IMDi, for cooperation underway. Project leader has been Kristian Rose Tronstad of the Norwegian Institute for Urban and Regional Research, NIBR. Tronstad has conducted the statistical analyses and written most of the report. Marit Nygaard at NIBR has contributed on analysis and discussion of the results. Miia Bask at Norwegian Social Research, NOVA, has written about the theory of accumulation of welfare problems in Chapter 3.