IMPROVING HEALTH IN THE NETHERLANDS. THE ROLE OF SOCIO-ECONOMIC STATUS, PERCEIVED SELF-RELIANCE AND EXPECTATIONS FROM THE GOVERNMENT.
Master thesis, Human Resource Management
University of Groningen, Faculty of Economics and Business
February 3, 2019
Lonne Milou Stegeman S1900633 Johan de Wittstraat 90
9716 CJ Groningen +31 (0)6 227 88 311 l.m.stegeman@student.rug.nl
Supervisor dr. Susanne Täuber
s.tauber@rug.nl
Second Assessor dr. Laetitia Mulder
l.b.mulder@rug.nl
Acknowledgement: I would like to thank dr. Susanne Täuber for her helpful feedback, support and above all, her contagious enthusiasm. I thank my friends and family who believed in me and supported me while writing this thesis, and all 151 respondents who participated in this study.
ABSTRACT
Socio-economic status (SES) is an important predictor for health outcomes. This research investigated the role of perceived self-reliance and expectations from the government in explaining lower levels of trust in institutions among citizens with a low SES, as trust in institutions antecedes health outcomes. It was proposed that SES influences trust in institutions through perceived self-reliance and health-related expectations from the government. Citizens with a low SES were expected to report less perceived self-reliance, which in turn was hypothesized to be associated with higher expectations from the government. Subsequently, as the responsibility of health shifted towards citizens as a consequence of the participation society, higher expectations from the government with health-related issues were expected to relate to more mistrust in institutions as these expectations were often not met.
One-hundred-fifty-one respondents recruited in low and high SES areas filled in a hard-copy survey. Results show that contrary to predictions, no indirect effect of SES on trust in institutions through perceived self-reliance and expectations from the government was evident. Rather, separate indirect effects through perceived self-reliance on trust in institutions and through expectations from the government were found, indicating that interventions aiming to improve the health of citizens with a low SES need to focus on perceived self-reliance and expectations from the government separately.
Results indicated that citizens with a lower SES were more likely to report lower perceived self-reliance scores, and higher expectations from the government. The positive relationship between SES and perceived self-reliance was reinforced for respondents with a migration background. Theoretical and practical implications of the findings are discussed.
INTRODUCTION
Health is and has been high on the agenda of today’s leading forums of the world’s major economies, as it is crucial for sustainable development and life on our planet (Group of Twenty [G20]
Argentina, 2018; Italian Group of Seven [G7], 2017). The three main focus areas of the G20 regarding health are: strengthening health systems in terms of efficiency, malnutrition and obesity, and pandemics (G20 Argentina, 2018). Health is a dynamic concept and refers to the individual ability to adapt to and self-manage the physical, emotional and social challenges of life (Huber et al., 2011).
Being healthy is important because it contributes to the length and quality of one’s life (Chetty et al., 2016; Post, Zwakhals, & Polder, 2010). Health is also needed to be able to work, develop oneself, enjoy free-time and participate in society. Poor health can interfere with personal aspirations to give meaning to life as it hinders citizens from participating socially, economically or in society. Of the Dutch population, 79% thinks that being healthy is important, just as love, friendship and family (Synovate, as cited in Van Ewijk, Van der Horst, & Besseling, 2013).
Health is not only important on the individual level. Society, too, benefits from healthy citizens because public health is related to prosperity (Arora, 2001; Bloom, Canning, & Sevilla, 2004; Bloom,
& Malaney, 1998; Bloom, & Williamson, 1998; Fogel, 1994; Suhrcke, Rocco, & McKee, 2007). In short, this means that health is wealth. One needs to be healthy in order to be able to work, which is needed to provide an income. Without income, it is very difficult to become wealthy. However, this relationship is reciprocal. Wealth is also health. A decrease in public wealth can cause a decrease in income, which can subsequently cause a decrease in the living conditions and lifestyle, which are determinants of health (Klinkmann, & Vienken, 2008). Collectively, poor health can lead to higher costs for society and decreased productivity. The Dutch government, too, recognizes the importance of health and acknowledges it as a societal value in its many reports and policy documents (Ministerie van Volksgezondheid, Welzijn en Sport, 2007; Post et al., 2010; Rutte, & Samsom, 2012).
However, despite its importance, health is unequally distributed. Not only across countries around the world (Hosseinpoor, Bergen, Schlotheuber, Victoria, Boerma, & Barros, 2016), but also within (Western) countries. In the Netherlands, large health differences between population groups exist (Organisation for Economic Cooperation and Development [OECD], 2014; Pickett, & Wilkinson, 2015;
Volksgezondheid Toekomst Verkenning [VTV], 2018). Overall, Dutch citizens with a low socio- economic status (SES) and/or a migration background are less healthy than indigenous people with a high SES (Brussaard, Van Erp-Baart, Brants, Hulshof, & Löwik, 2001; De Greef, 2009). For instance, chronic diseases such as diabetes occur more often among low SES groups (Statistics Netherlands [CBS]
Statline, 2014). Also, high SES groups have a seven-year-longer life expectancy than low SES groups and an 18 year-longer healthy-life expectancy (VTV, 2018). Aforementioned studies indicate that these groups are less healthy, in absolute terms.
However, in subjective terms, these groups also perceive themselves as less healthy. In the Netherlands, 50.9% of the people with a low education level perceive their health as good, whereas 86.8% of the people with a high education level perceive their health as good. Of the people in the lowest income group, 69.6% perceive their health as good versus 86.8% of the highest income group. And of the first-generation non-western migrants, 62.1% perceive their health as good versus 80.7% of the indigenous people (CBS Statline, 2018a).
Previous studies showed that trust is needed for participation in health initiatives and that mistrust is hindering health improvements (Wang, Schlesinger, Wang, & Hsiao, 2009; Giordano, &
Lindström, 2015). However, in the Netherlands, trust differs by SES (CBS Statline, 2018b) in such a way that citizens with a low SES report less trust in the government in comparison to citizens with a high SES.
Despite numerous health initiatives aimed at specific habits of these specific groups, the health disparities still remain today (OECD, 2014; Pickett, & Wilkinson, 2015; VTV, 2018). Several effectivity studies corroborate that initiatives aiming at smoking, use of alcohol, obesity, diabetes and depression, are not (yet) effective (Hoogervorst, 2006; National Institute for Public Health and the Environment [RIVM], 2010; 2012). The large health disparities are at odds with the goals of the Dutch government to reduce health differences between population groups (Hoogervorst, 2006). In 2013, the Dutch government introduced the participation society as a substitute for the welfare state. This meant that
“everyone who is able will be asked to take responsibility for their own lives and immediate
Rucell, 2015). At that time, the government noticed that citizens became more assertive and independent and that there was a growing need for the freedom to arrange one’s own life. The former system could not fulfill this need and often even restrained citizens from taking initiative in the health domain, due to its hierarchical set-up. At the same time, the budget deficit needed to be reduced (Rutte, 2014). The new system provided an opportunity to increase the autonomy of citizens and reduce the need for funding of the government. At first sight: a win-win solution.
Yet, research indicates that some citizens may not be self-reliant enough to cope with the increased responsibilities regarding health (Wetenschappelijke Raad voor het Regeringsbeleid [WRR], 2017). Focus-group research showed for instance that people who belong to certain population groups, in particular groups with low SES and migration background, experience a lack of knowledge about the health domain and therefore prefer more help from the government regarding health (De Ruyter, 2017).
This is however contradictive to the expectations of the Dutch government towards citizens, since more self-reliance is expected from individual citizens (Speech from the throne, 2013). Considering these divergent expectations, it comes as no surprise that some of these citizens become demotivated (De Ruyter, 2017) and mistrust the government (CBS Statline, 2018b).
I therefore suggest that the expectations of the government towards citizens regarding health on the one hand, and of the citizens with a low SES and migration background towards the government regarding health on the other hand, do not coincide. Subsequently, I suggest that citizens can become disappointed in the government for not living up to their expectations, which potentially evokes mistrust towards the government and other institutions. As trust is needed for participation in health initiatives, mistrust is blocking health improvements (Wang et al., 2009; Giordano, & Lindström, 2015).
To summarize, this research aims to find out why health differs by SES by looking into trust in institutions. More specific, the relationship between SES and perceived self-reliance is investigated, which is in turn expected to be associated with expectations from the government regarding health. The higher the expectations from the government regarding health, the more mistrust is expected.
By studying the role of SES and migration background in relation to self-reliance and the expectations from the government regarding health, this thesis produces insights that potentially contribute to the effectiveness of preventive interventions aimed at groups with a migration background
or low SES, as the government wishes to do (Hoogervorst, 2006). Understanding the way SES influences citizens’ expectations from the government regarding health allows designing interventions that respond to the specific needs of the population groups in question, which potentially neutralizes the negative reactions and subsequently contributes to the success of the interventions. Eventually, this may contribute to the improvement of health systems in terms of efficiency as is an aim of the G20 (G20 Argentina, 2018). At the same time, improving the health of citizens with a low SES or migration background can yield large cost savings for the Dutch society (Post et al., 2010).
In the theory section, the relationships between the concepts SES, migration background, perceived self-reliance, expectations from the government regarding health, and trust in institutions are described, and a theoretical model based on the formulated hypotheses is introduced.
THEORY
Previous studies showed that trust is needed for participation in health initiatives and that mistrust is hindering health improvements (Wang et al., 2009; Giordano, & Lindström, 2015).
Therefore, in this thesis, I aim to explain why citizens with a low SES are more likely to mistrust the government as this antecedes health outcomes. I propose that an important and under-theorized explanation for the relationship between SES and trust is based in perceived self-reliance and subsequently in expectations towards the government. Specifically, I propose that perceived self- reliance is even lower for citizens with a low SES that are not originally from the Netherlands. In the following sections, I elaborate on these predictions.
Socio-economic status and trust in institutions
Previous research defines SES in various ways. When relating to health, SES refers to “an attempt to capture an individual’s or group’s access to the basic resources required to achieve and maintain good health” (Shavers, 2007: 1013). In the Netherlands, SES is often indicated by education level as it easy to measure and is a predictor for other indicators of SES, such as income and occupation (Ruijsbroek, Wijga, Kerkhof, Koppelman, Smit, & Droomers, 2011). According to Ross, Mirowsky and Pribesh (2001), there is a direct influence of SES on trust in institutions. They found that mistrust
mistrust among its citizens. In addition, these citizens lack the social and economic resources to achieve desired ends that encourage trust. The absence of government institutions in the neighborhoods where they live makes them potentially feel abandoned and subsequently makes them more mistrusting towards the government.
In the Netherlands, citizens with a low SES report less trust in the government, in comparison to citizens with a high SES. Of this low SES group, only 28.9% trust the government ‘a lot’ to
‘moderately’, in contrast to 58.3% of citizens with a high SES (CBS Statline, 2018b). This is supported by a focus group study of De Ruyter (2017), where low SES and migration groups were more mistrustful towards the government. Based on the results from the CBS Statline (2018b), the findings of De Ruyter (2017) and those of Ross et al. (2001), it appears plausible to assume that SES directly influences the extent to which citizens trust the government. Therefore, I hypothesize:
Hypothesis 1: Higher SES is associated with greater trust in institutions.
The next section will explain why SES can be assumed to influence trust and considers the potential role of self-reliance and expectations from the government. Thus, SES is expected to influence trust in institutions indirectly, as well as directly. I elaborate on the indirect relation in the following sections.
Socio-economic status and perceived self-reliance
A possible explanation for the low levels of trust among citizens with a low SES is through perceived self-reliance and subsequently through expectations from the government. In this section, I illustrate the expected relationship between SES and perceived self-reliance. In the following section, I elaborate on the expected relationship between perceived self-reliance and expectations from the government.
Perceived self-reliance refers to the extent to which citizens experience a sense of personal mastery and is an aspect of psychosocial maturity (Pearlin, & Schooler, 1978). It is composed of an absence of excessive dependency on others, a sense of control over one’s life, and initiative (Greenberger, & Sørensen, 1974). According to Geis and Ross (1998), perceptions of powerlessness and low self-reliance are reinforced by negative and uncontrollable events that are more common in low SES areas, such as crime, vandalism and danger. Ross et al. (2001) found that citizens who live in
disadvantaged neighborhoods (low SES) and feel as if their life is shaped by forces outside one’s control (low perceived self-reliance) are more likely to develop mistrust than are others who feel a sense of personal control to avoid harm.
A sense of control over one’s life is an aspect of self-reliance (Greenberger, & Sørensen, 1974).
Young and Shorr (1986) found that citizens with a high SES are more likely to report an internal locus of control, meaning that they perceive more sense of control, in comparison to low SES citizens. In this research, I expect that, along the research of Ross et al. (2001), and Young and Shorr (1986), citizens with a low SES lack the social and economic resources they need to achieve their goals, which may lead them to feel less self-reliant.
In both subjective and objective self-reliance, citizens with a low SES report lower scores, due to lower educational and income levels. Therefore, they are more dependent on others (Galenkamp, Plaisier, Huisman, Braam, & Deeg, 2011). Health literacy, ‘the degree to which individuals can obtain, process, understand and communicate about health-related information needed to make informed health decisions’ (Berkman, Davis, & McCormak, 2010: 16), contributes to the extent to which one is self- reliant regarding health and can be referred to as objective health literacy. Citizens with a low SES have lower objective health literacy since health literacy is correlated with education and reading abilities.
Low health literacy is found among almost 29% of the Dutch population. Among them are mostly sick people, people older than 75, low educated people and people with financial problems (Health Literacy Survey Europe [HLS-EU] Consortium, 2012).
Based on the above findings, it appears plausible to assume that SES influences perceived self- reliance. Therefore, I hypothesize:
Hypothesis 2: Higher SES is associated with greater perceived self-reliance.
The next section will explain whether the positive relationship between SES and perceived self- reliance is expected for all citizens.
Migration background and perceived self-reliance
A focus group study showed that citizens with a migration background often have lower
have a low SES, also have a migration background. These specific groups are less healthy than indigenous people with a low SES. Therefore, I expect that the positive relationship between SES and perceived self-reliance is strengthened when one has a migration background. Thus, I hypothesize:
Hypothesis 3: The relationship between SES and perceived self-reliance is moderated by migration background, such that this relationship is stronger for citizens with a migration, as opposed to an indigenous, background.
The next section will explain why perceived self-reliance is expected to influence citizens’
expectations from the government.
Perceived self-reliance and expectations from the government
According to De Ruyter (2017), citizens have expectations from the government in the health domain about whether the government does enough, and who should be responsible for the health of citizens. The first, relate to citizens’ attitude towards the interference of the government and more specific, whether citizens think that the government does enough for its citizens regarding health. The latter refers to citizens’ attitude towards the distribution of responsibility regarding health and it refers to who they think should be responsible for the health of citizens.
De Ruyter (2017) showed in a focus-group study that low SES groups have more and higher expectations from the government. She speculates that because these groups report less perceived health-literacy, they expect the government to help and support them more than others who feel more self-reliant. In short, low SES groups experience a knowledge deficit when they compare themselves to others which influences their perceived self-reliance and subsequently their expectations from the government.
The difference in expectations from the government between SES groups is also found by a study conducted by Olsthoorn and Van der Torre (2018). They indicated that higher educated citizens are more willing to cut down on public healthcare costs, which suggests that they perceive themselves as more self-reliant and thus make less use of public healthcare than lower educated people. As Young and Shorr (1986) showed, low SES groups often have an external locus of control which means that they hold other parties responsible for important aspects of their life. Therefore, it is expected that these groups, because they feel less self-reliant and have an external locus of control, expect relatively more
from the government. Accordingly, I expect that high SES citizens have lower expectations from the government as they perceive themselves as more self-reliant. Therefore, I hypothesize:
Hypothesis 4: Perceived self-reliance is negatively associated with expectations from the government in such that greater perceived self-reliance is associated with less expectations from the government.
In the next section, I will explain why expectations from the government are expected to influence trust in institutions.
Expectations from the government and trust in institutions
In the current participation society, high expectations of citizens are often not met since the responsibility has shifted from government towards citizens. This is a possible explanation for the mistrust among low SES citizens, that was also found by De Ruyter (2017). Additionally, Ross et al.
(2001) showed that citizens in lower SES areas sometimes feel abandoned by the government since their neighborhoods often have fewer resources such as good schools or medical facilities. Therefore, they potentially feel let down by the government, which can evoke mistrust. Therefore, I hypothesize:
Hypothesis 5: In the context of the participation society, health-related expectations from the government are negatively associated with trust in institutions. The more citizens expect from the government, the less trust they will report in the government.
Taken together, the hypotheses derived above suggest the following overall associations between the constructs:
Hypothesis 6: The interactive effect of SES and migration background on trust in institutions is mediated through self-reliance and expectations from the government.
Summing up, Figure 1, on the following page, depicts the hypotheses formulated in the section above.
METHODOLOGY Procedure and participants
The hypotheses were tested by means of a questionnaire and respondents were recruited in front of Dutch supermarkets in different postal areas in the three northern provinces of the Netherlands. Postal areas varied from low to high SES in order to achieve a mix of respondents. Participation was voluntary and anonymous. After respondents confirmed participation, they were asked to fill in a hardcopy survey.
The survey (Appendix I) was written in Dutch, in order to make sure only Dutch citizens participated, as trust in the Dutch government was measured which is only relevant to Dutch citizens. It contained measures for SES, migration background, perceived self-reliance, expectations from the government, trust in institutions and several other variables that deemed sufficiently relevant for the research question at hand. Respondents were also asked to report demographic information such as age, gender, educational level and country of origin. In order to control for earthquake-related mistrust, respondents were asked whether they experienced personal earthquake damage. Completing the survey took respondents between five and ten minutes and happened in front of supermarkets.
One-hundred-fifty-one persons completed the survey (Mage = 44.81, SDage = 17.24; range 18 to 90; 83 females). This corresponds to data from CBS Statline (2018c), showing that in the northern provinces, 50.1% is female. Respondents’ level of education varied from ‘no education’ (1.3%), to
‘primary school / lower secondary school’ (12.6%), ‘upper secondary school / MBO’ (37.7%) ‘HBO / university’ (47.7%) and ‘other’ (0.7%). Most of the respondents lived in one of the three northern provinces of the Netherlands: Groningen (62.3%), Drenthe (23.2%), Friesland (13.2%), but a few of them lived elsewhere in the Netherlands (1.3%). Of the respondents, 81.5% was indigenous and others
Perceived Self-Reliance
_ Expectations from Government
_ +
Trust in Institutions Socio-Economic
Status Migration
Background
Figure 1 Theoretical model.
+ +
had a migration background (18.5%). This corresponds to data from CBS Statline (2018c), showing that in the northern provinces of the Netherlands, 88.2% is indigenous. Most of the respondents with a migration background were first-generation migrants (60.7%), and some were second-generation migrants (39.3%), I elaborate on this in the following section. Migrants came from Aruba, China, Colombia, Curacao, Indonesia, Iran, Korea, Philippines, Ukraine, Syria, Surinam and Thailand. The average standardized SES of the respondents, based upon their postal code, was -.41 (SD = 1.54), which is slightly lower than the Dutch average which is 0. For the purpose of this study, I have transformed the standardized SES scores that were provided by Volksgezondheid en Zorg [Public Health and Health Care] (2018). I elaborate on this in the following paragraph. The transformation resulted in 29.8% of the respondents with a low SES. Three respondents had to be excluded from the main analyses due to incomplete information.
Measures
Socio-economic status. In Dutch research, income, level of education and postal code are often
used as indicators for SES (Van Oyen, Deboosere, Lorant, & Charafeddine, 2011). To measure SES in this study, respondents were asked to report their postal codes in order to avert defensiveness and discomfort that would be caused by asking about income. Postal codes were matched to the corresponding SES scores. These scores are provided by Public Health and Health Care as continuous, standardized, scores (2018). I converted the scores into a dichotomous variable in which negative values, and values equal to 0 (£ 0) referred to a low SES and positive values (> 0) referred to a high SES.
Further, I corrected for education level and students. This meant that I gave education a higher weight than postal code, in order to correct for the Dutch phenomenon of “scheefwoners”, which refers to highly educated people who remain living in neighborhoods where rent is lower, to save money. A score of 0 (low SES) based on postal code of a highly educated respondent (education level HBO / University) was therefore corrected to a 1 (high SES). Also, scores of respondents that were currently studying, while living in a low SES neighborhood (0), were corrected to a 1. This resulted in a dichotomous variable in which a score of 0 referred to a low SES and a score of 1 referred to a high SES.
Migration background. The country to which a person has the closest ties, based on his or her
parents’ country of birth and his or her own country of birth, can be referred to as migration background.
One could differentiate between first- and second-generation migration background, of which the first is defined as a person’s country of birth. A person who is not born in the Netherlands, of whom at least one parent is also not born in the Netherlands has a first-generation migration background. A person who is born in the Netherlands of whom at least one parent is not born in the Netherlands has a second- generation migration background. A person who is born in the Netherlands, of whom both parents are also born in the Netherlands has no migration background and can be called indigenous or native (CBS Statline, 2018c). Migration background can therefore vary from indigenous, second-, to first generation migration background. In the main analyses, I did not differentiate between first- and second-generation background because the number of respondents in these groups was too low to consider them separately.
Therefore, I created a dichotomous in which a score of 0 referred to respondents with a first- or second- generation migration background and a score of 1 referred to indigenous respondents. I treated respondents with a western migration background (for instance Germans) as indigenous since these cultures are more similar to the Dutch culture, in comparison to non-western cultures (Hofstede, 2019).
Perceived self-reliance. In this research, self-reliance was measured through a self-reported
‘sense of mastery’ scale, as perceptions are more likely to influence attitude, in comparison to absolute self-reliance (Ajzen, 1991). The original ‘sense of mastery scale’ was created by Pearlin and Schooler (1978) and contained seven items. However, according to Gadalla (2009), two items are better to leave out. Therefore, only the first five items of the scale were used of which ‘There is little I can do to change many of the important things in my life’ is an example item. Respondents reported on a 5-point Likert scale, ranging from ‘completely disagree’ (1), to ‘completely agree’ (5). The items were reverse-coded because they were negatively formulated. A high value therefore reflected greater perceived self- reliance. After reliability analysis, the items were combined into a scale named ‘perceived self-reliance’
(Cronbach’s α = .80).
Expectations from the government. According to De Ruyter (2017), citizens have expectations from the government in the health domain about whether the government does enough, and who should be responsible for the health of citizens. The first, expectations from the government about whether the
government does enough is used in the main analyses because, according to De Ruyter (2017), these expectations differed by SES. The second, responsibility regarding health, is described in the following paragraph and was used in the additional analyses. The expectations from the government were measured by asking to what extent the respondents thought that ‘the government does enough for its citizens regarding health’. Respondents reported on a 5-point Likert scale, ranging from ‘completely disagree’ (1) to ‘completely agree’ (5). The variable was reverse-coded, such that a high value reflected higher expectations.
Responsibility regarding health. Respondents’ attitudes towards who should be responsible
for health, were measured by asking ‘who do you think should be responsible for the health of citizens?’.
This variable was used in the additional analyses. Respondents reported on a 7-point scale varying from
‘citizens are responsible for their own health’ (1), to ‘the government is responsible for the health of the citizens’ (7). I did not differentiate between the seven levels but treated the lowest values (< 4) as
‘citizens should be responsible’; the median as ‘neutral’; and the highest values (> 4) as ‘government should be responsible’, in order to gain more insights. A higher score reflected the attitude that the government should be responsible, and a lower score reflected the attitude that citizens should be responsible.
Trust in institutions. Trust was measured by asking respondents the World Values Survey Trust
Question, which is a valid indicator for trust (Johnson, & Mislin, 2012). Respondents reported their answer to ‘Generally speaking, would you say that most people can be trusted, or that you can’t be too careful in dealing with people?’ on a 10-point scale, varying from ‘most people can be trusted’ (10), to
‘you can’t be too careful in dealing with people’ (1). The responses were transformed into a 5-point scale and combined with four questions about trust in the government. Respondents reported on a 5- point Likert scale varying from ‘not at all’ (1), to ‘to a very high extent’ (5), to items such as ‘To what extent do you have trust in the local government?’. The five items were combined into the variable ‘trust in institutions’ (α = .73). The higher the score, the greater the trust in institutions.
Objective health literacy. The Dutch version of the Newest Vital Sign (NVS; Shah, West,
were related to a Ben & Jerry’s ice cream nutrition label (Appendix II). The total number of correct answers was listed, and the more correct answers, the greater the health literacy.
Subjective health literacy. In order to measure subjective health literacy, respondents were
asked how much they think they know about health and reported this on a 10-point scale varying from (1) ‘nothing’, to (10) ‘everything’. The responses were transformed into a 5-point scale. Second, respondents were asked to what extent they think they were able to ‘obtain, process, understand and communicate about health-related information needed to make informed health decisions. Respondents reported on a 5-point Likert scale varying from (1) ‘not at all’, to (5) ‘to a large extent’. Both items were combined into one variable as there was significant correlation (r = .26, p = .001) in such that a high value referred to higher subjective health literacy.
Factors facilitating and hindering a healthy lifestyle. In addition to the main analyses that
considered trust as an antecedent for healthy behavior, it deemed also interesting to look at other factors influencing a healthy lifestyle. Therefore, two questions investigating other factors influencing a healthy lifestyle, were added to the survey: ‘what factors around you help you to live healthy?’ and ‘what factors around you hinder you to live healthy?’. Respondents could tick multiple boxes containing factors introduced by a study of the WRR (2018). Facilitating factors were: knowledge and abilities, having enough to spend, living in a neighborhood with enough opportunities to live healthy and a social environment where a healthy lifestyle is the norm. Factors hindering a healthy lifestyle were: not having enough money, few healthy alternatives the neighborhood, lack of knowledge and abilities about health.
Both questions ended with a box: ‘other factors, such as:’, where respondents could fill in a factor themselves.
Control variables. Respondents were asked to report gender, birth year, highest level of
education, whether they were currently studying and whether they experienced personal earthquake damage (18.1% of the respondents reported that they had experienced personal earthquake damage).
Data analysis
First, a descriptive analysis was performed. Second, means, standard deviations and correlations were analyzed. In order to test the hypotheses, a simple moderation model was carried out, as well as a serial multiple mediation model. These correspond with model no. 1 and 6 by Hayes (pre-defined,
PROCESS macro, Hayes, 2012). The additional analyses were performed with the use of a Chi-square test.
RESULTS Descriptive statistics and correlations
In Table 1, descriptive statistics and correlation coefficients of all the variables under study are displayed. The results indicate a positive association between SES and trust in institutions (r = .33, p <
.01), providing support for hypothesis 1. Also, the relationship between SES and perceived self-reliance was positive, thereby providing support for hypothesis 2 (r = .26, p < .01). The relationship between expectations from the government and trust in institutions was significantly negative (r = -.36, p < .01) providing support for hypothesis 5. Even though a relationship between perceived self-reliance and expectations from the government is unlikely because of the absence of a significant correlation, a sequential mediation model was tested, after the first analysis that focused on the moderating role of migration background. Since gender and migration background significantly correlate with SES and perceived self-reliance, they were controlled for in the main analyses.
Table 1 Means, standard deviations, correlations and Cronbach’s alpha’s
M SD 1 2 3 4 5 6
1. Gender (1 = male; 2 = female) 1.55 .50 - 2. Migration Background
(0 = indigenous; 1 = migration) .19 .39 .12 -
3. Socio-Economic Status .70 .46 .17* -.32** -
4. Perceived Self-Reliance 3.77 .71 .02 -.25** .26** (.80) 5. Expectations from Government 3.12 .94 .06 .01 -.17* -.13 -
6. Trust in Institutions 3.54 .87 .13 -.14 .33** .47** -.36** (.73)
* correlation significant at .05 level (two-tailed) **correlation significant at .01 level (two-tailed) Cronbach’s a on the diagonal
Main analyses
The hypothesis whether the relationship between SES and perceived self-reliance is moderated by migration background was tested first, by a simple moderation analysis (Model 1 as described in PROCESS) with z-standardized variables (Hayes, 2012). SES, being the independent variable, was coded as a dichotomous variable, being either high (1) or low (0). The dependent variable, perceived
self-reliance, was a continuous variable and migration background, the moderator, was coded as indigenous (0) or migration background (1). In this model I controlled for gender.
As the results in Table 2 demonstrate, the expected relationship was not significant for indigenous citizens as the confidence interval included 0 (CI95% -.052, .327), but it was significant for migrants (CI95% .058, .740). Figure 2 is a graphical representation of the conditional effect of migration background and shows that for respondents with a migration background, perceived self-reliance is lower when they have a lower SES, while perceived self-reliance is higher when they have a high SES.
Therefore, the positive relation between SES and perceived self-reliance is reinforced by migration background. The effect of the moderator was not significant for natives, which means that the relationship between SES and perceived self-reliance did not significantly differ for indigenous respondents.
Table 2. Conditional effect of migration background on the relationship between socio-economic status and perceived self-reliance.
Perceived self-reliance
Effect Boot SE LLCI* ULCI*
Indigenous (0) .14 .096 -.052 .327
Migration background (1) .40 .172 .058 .740
*Based on 5.000 bootstrap interval
Indigenous
Migration Background
Figure 2 Graphical representation of the conditional effect of migration background on the relationship between socio-economic status and perceived self-reliance
In order to test the remaining hypotheses proposing that perceived self-reliance and expectations from the government sequentially mediate the impact of SES on trust in institutions, I performed a sequential mediation analyses (Model 6 as described in PROCESS) with z-standardized variables (Hayes, 2012). Figure 3 is a graphical representation of all the paths for the full process model, while the coefficients are displayed in Table 3 on the next page.
The results demonstrate that the expected positive direct effect of SES on trust was significant (b = .16, t = 2.16, p = .033) supporting hypothesis 1 and indicating that the higher the SES, the more trust in institutions can be expected. Also, the expected positive relationship between SES and perceived self-reliance was significant, providing support for hypothesis 2 (b = .19, t = 2.22, p = .028).
Furthermore, the results indicate that there is a significant negative relationship between expectations from the government and trust in institutions (b = -.28, t = -4.02, p < .001), providing support for hypothesis 5. The effect of perceived self-reliance on expectations from the government was not significant (b = -.11, t = -1.31, p > .100), rejecting hypothesis 4. Obviously, because perceived self- reliance was not significantly related to expectations from the government, the specific indirect effect of SES on trust in institutions through both perceived self-reliance and expectations from the government was not significant either, rejecting hypothesis 6 (a1a3b2 = .01; CI95%-.001, .026). However, the specific indirect effect through self-reliance only was significant (a1b1 = .07; CI95% .010, .156); as was the indirect effect through expectations from the government only (a2b2 = .05; CI .006, .122),
a2 = -.18*
Perceived Self-Reliance
Trust in Institutions Socio-Economic
Status
Expectations from Government a3 = -.11
a1 = .19*
b2 = -.28***
*** p < .001, ** p < .01, * p < .05
Figure 3 Observed associations of the sequential multiple mediation model.
Direct effect =.16*
b1 = .38***
Table 3 Results of sequential mediation analysis (PROCESS, Hayes 2012)
Dependent variable: Perceived Self-Reliance
R R2 F df1 df2 p
.28 .08 4.13 3.00 144.00 .008
Coeff SE t p
Constant .05 .26 .17 .862
Socio-Economic Status .19 .09 2.22 .028
Dependent variable: Expectations from Government
R R2 F df1 df2 p
.22 .05 1.85 4.00 143.00 .122
Coeff SE t p
Constant -.24 .27 -.88 .378
Perceived Self-Reliance -.11 .09 -1.31 .191
Socio-Economic Status -.18 .09 -2.01 .046
Dependent variable: Trust in Institutions
R R2 F df1 df2 p
.59 .35 15.09 5.00 142.00 .000
Coeff SE t p
Constant -.35 .23 -1.53 .127
Perceived Self-Reliance .38 .07 5.35 .000
Expectations from Government -.28 .07 -4.02 < .001
Socio-Economic Status .16 .08 2.16 .033
Total effect model
Dependent variable: Trust in Institutions
R R2 F df1 df2 p
.34 .12 6.34 3.00 144.00 .000
Coeff SE t p
Constant -.26 .26 -1.00 .318
Socio-Economic Status .30 .09 3.45 .001
Total, direct and indirect effects Total effects of Socio-Economic Status on Trust in Institutions
Effect SE t p
.30 .09 3.45 .001
Direct effects of Socio-Economic Status on Trust in Institutions
Effect SE t p
.16 .08 2.16 .033
Indirect effects of Socio-Economic Status on Trust in Institutions
Effect Boot SE LLCI ULCI
Total .13 .049 .040 .239
Ind1: .07 .035 .010 .156
Ind2: .01 .006 -.001 .026
Ind3: .05 .028 .006 .122
Indirect effect key
Ind1: Socio-Economic Status à Perceived Self-Reliance à Trust in Institutions
Ind2: Socio-Economic Status à Perceived Self-Reliance à Expectations from Government à Trust in Institutions
Ind3: Socio-Economic Status à Expectations from Government à Trust in Institutions
Coeff coefficient, LLCI lower limit confidence interval, ULCI upper limit confidence interval
Summing up, the proposition that a higher SES is associated with greater perceived self-reliance, which in turn is associated with lower expectations from the government and subsequently leads to more trust in the government, was not supported by the analysis. Hence, this study does not yield support that the relationship between SES and trust in institutions is sequentially mediated by perceived self-reliance and expectations from the government.
Combining the two models, the results indicate that the positive relation between SES and perceived self-reliance is reinforced for migrants. Also, the results indicate that the relationship between SES and trust in institutions can be explained through perceived self-reliance, and, separately, through expectations from the government.
Additional analyses
I measured several other variables that were not part of the process model hypotheses, but which I deemed sufficiently relevant for the research question at hand. These included health literacy satisfaction, objective and subjective health literacy, attitudes towards the responsibility regarding health, and factors respondents experience as facilitating and hindering healthy lifestyles. In the next section I will report the additional analyses that I was able to perform with this data. A Chi-square test was executed to determine whether scores for the aforementioned variables differed by SES and migration background. I report the results in the following paragraphs. The last two paragraphs contain the factors facilitating and hindering a healthy lifestyle.
Health literacy satisfaction. A Fisher’s Exact test was run to determine whether satisfaction
with health literacy levels differed by SES or migration background, since the requirements for a Chi- square test were not met. Both tests were not significant, which indicates that respondents with high (M
= .93, SD = .25) and low (M = .91, SD = .29) SES were equally likely to be (dis)satisfied with their health literacy level (p = .733). The satisfaction levels of natives (M = .93, SD = .25) and migrants (M
= .89, SD = .32) also did not differ significantly (p = .429). Altogether, SES nor migration background do not seem to influence the extent to which one is satisfied with health literacy.
Objective health literacy. In total, 43% of the respondents answered between 0 and 2 questions
22.69, p < .001. Respondents with a low SES answered less questions about the nutrition label correctly (M = .38, SD = .58), in comparison to respondents with a high SES who were more likely to answer more answers correctly (M = 1.08, SD = .86). Objective health literacy was also significantly different between natives and migrants, c2 (2, 151) = 11.80, p = .003. Migrants reported less correct answers (M
= .39, SD = .69) versus indigenous respondents (M = .98, SD = .85). In sum, objective health literacy scores differed significantly by SES and migration background.
Over- and underestimating health literacy. The difference between subjective and objective
health literacy was significantly different between respondents with a high and low SES, c2 (2, 151) = 10.68, p = .005. A positive score referred to overestimation, whereas a negative score referred to underestimation. Respondents with a low SES were more likely to overestimate their health literacy (M
= 1.76, SD = .53) compared to respondents with a high SES (M = 1.36, SD = .74). The difference between natives and migrants in estimated health literacy was marginally significant c2 (2, 151) = 3.72, p = .092, such that migrants were more likely to overestimate their health literacy (M = 1.71, SD = .54) and natives were more likely to underestimate (M = 1.03, SD = .73). Altogether, respondents with a low SES and migrants were more likely to overestimate themselves, while respondents with a high SES and natives were more likely to underestimate themselves.
Responsibility regarding health. The attitudes towards the responsibility of the government
regarding health did not significantly differ by SES c2 (2, 151) = 3.39, p = .183, but it differed between natives and migrants, c2 (2, 151) = 7.28, p = .026. Specifically, migrants were more likely to have the opinion that the government should be responsible for the health of citizens (M = .46, SD = .84) compared to native citizens who were more likely to have a neutral opinion (M = .02, SD = .88), meaning that both citizens and the government should be responsible. In summary, attitudes towards the responsibility of the government regarding health did not differ by SES but did differ by migration background.
Factors facilitating healthy behavior. The majority of the respondents, irrespective of SES or migration background, reported that knowledge and abilities (64.9%), and a social environment where living a healthy life is the norm (63.6%) were important factors for them in facilitating a healthy lifestyle.
Having enough money to spend was important to natives and respondents with a high SES (58.5%), but not to migrants or respondents with a low SES (75.6%). Living in a neighborhood with enough opportunities to live healthy was only important to the minority of the respondents (35.8%), independently of SES or migration background. Factors facilitating healthy behavior that were mentioned irrespective of SES or migration background, in the ‘other factors, such as:’- section, were:
luck, family members, personal trainer, sport, be an example to the kids, study, self-reflection and life experience.
Factors hindering healthy behavior. Independent of SES or migration background, the
majority of the respondents reported that none of the suggested factors were hindering them from living healthy. Of the respondents, only 15.9% reported that having not enough money hindered them from living healthy. Having few healthy alternatives in the neighborhood was only for 3.3% a factor hindering them and a lack of knowledge and abilities about health hindered only 17.2% of the respondents from living healthy. Factors that were mentioned in the ‘other factors, such as:’- section, irrespective of SES or migration background, were: lack of discipline, lack of time, people around, unhealthy alternatives are cheaper.
DISCUSSION
The aim of this study was to offer an explanation for the observation that citizens with a low SES are less healthy in comparison to citizens with a higher SES. Since trust in institutions is a strong predictor for health outcomes (Wang et al., 2009; Giordano, & Lindström, 2015), the relationship with SES was investigated. I expected that SES influences trust in institutions through perceived self-reliance and subsequently through expectations from the government and that the relationship between SES and perceived self-reliance is moderated by migration background. In the next section, I will exemplify the results of this research.
Discussing the results
Socio-economic status and trust in institutions. I found evidence supporting the expected direct relation between SES and trust in institutions, indicating that the higher the SES, the higher the trust in
Socio-economic status and perceived self-reliance. In line with my predictions and previous
research (Ross et al., 2001), I found evidence supporting a positive relationship between SES and perceived self-reliance in such that higher SES was related to higher perceived self-reliance. This means that different socio-economic groups have different perspectives with respect to self-reliance. This does not say anything about whether they are actually less self-reliant.
Moderating role of migration background. In this research, the relationship between socio-
economic status and perceived self-reliance is moderated by migration background, in such that it is reinforced for migrants. Therefore, for respondents with a low SES, perceived self-reliance was lower when they had a migration background, opposite to those with an indigenous background. This corresponds with the findings of De Ruyter (2017), who found that migrants experience a lack of knowledge about the health domain which might contribute to their feeling of being less self-reliant.
Parallel mediation. Against predictions that perceived self-reliance was negatively associated
with expectations from the government, there was no significant relationship found, implying that perceived self-reliance and expectations from the government are not related. The lack of evidence for this relationship indicates that supporting people to become more self-reliant, would probably not change their attitude towards the responsibilities of the government regarding health. However, I did find that both perceived self-reliance and expectations from the government, separately, significantly influence trust in institutions. Because of these distinct indirect effects, parallel mediation is indicated.
This means that there are several processes through which SES influences trust in institutions. Future research could investigate whether one or the other has more influence on trust in institutions, by testing a parallel mediation model. This would provide more thorough insights in the antecedents of trust in institutions and potentially lead to more efficient ways to influence it and subsequently health outcomes.
Also, considering the indirect effect of SES on trust in institutions through perceived self-reliance, this research demonstrates that making people feel more self-reliant would influence the levels of trust in institutions. Therefore, it also potentially influences the health outcomes of these groups, since the relationship between perceived self-reliance and trust in institutions was significant. This, ironically, implies that the people who need the government the least (because they feel very self-reliant), trust the
government the most. Indicating that in order to increase health outcomes by increasing trust in institutions, people should feel more self-reliant.
Expectations from the government and trust in institutions. The prediction that expectations
from the government are negatively related to trust in institutions was supported, indicating that expectations from the government antecede trust in institutions. Future research could investigate why this relationship is found so that it could be altered, in order to improve health outcomes. A potential solution is proposed by the research of Bouckaert and Van de Walle (2003) who state that trust in institutions depends on the extent to which citizens’ preferences match the perceived actual functioning of the government. This means that in order to improve trust in institutions, either preferences or perceived actual functioning of the government has to be altered. Future research could investigate what antecedes citizens’ preferences and perceived actual functioning of the government in order to look for opportunities to improve the match between both factors influencing trust in institutions.
Health literacy. As the additional analyses demonstrate, both respondents with a low and high
SES and migrants and natives were equally likely to be satisfied with health literacy levels. This is interesting because De Ruyter (2017) found that migrants with a low SES experience a knowledge deficit. This indicates, among other things, that either they do not mind the knowledge deficit, or they did not experience the knowledge deficit. The second explanation is more likely, as the additional analyses demonstrated that low SES respondents and migrants were more likely to overestimate health literacy scores. This phenomenon is called the ‘Dunning-Kruger effect’. In general, this means that people who are less competent, tend to overestimate their own abilities. More competent citizens on the other hand, tend to underestimate their own abilities. The explanation for this effect is that less competent citizens lack the metacognitive abilities to realize that they are less competent (Kruger, &
Dunning, 1999). The fact that citizens with a low SES are more likely to overestimate themselves might cause them to be less aware of the impact of their health choices and potentially less eager to learn how to improve their health. By improving the skills of these citizens, and thus increasing their metacognitive abilities, they become more able to recognize the limitations of their abilities (Kruger, & Dunning,
into account and make it easier to make a healthy choice. For instance, by creating nutrition labels that are clear and easy to understand or making unhealthy products more expensive. Lastly, objective health literacy differed by SES and migration background in such a way that with low SES respondents or migrants, a lower health literacy was more likely. This is in line with the previously mentioned study by De Ruyter (2017) and complements my speculations about the Dunning-Kruger effect.
Who should be responsible? The finding that migrants were more likely to think that the
government should be responsible for health, versus a neutral opinion by natives, can possibly be explained by cultural differences. Hofstede (2019) demonstrated that non-western cultures are more likely to score low on the ‘individualism dimension’ which addresses ‘the degree of interdependence a society maintains among its members’. The Netherlands scores very high on this dimension, meaning that it is an individualistic culture. This potentially explains the difference in attitude towards the responsibilities of the government between natives and migrants. As most migrants in this study came from non-western countries, that are often collectivist societies where people belong to groups that take care of each other, it is no surprise that this group was more likely to think that the government should be responsible for health. In order to make sure that these groups do not feel disappointed because certain expectations are not met, the government could involve these groups when developing and evaluating government policies in order to be able to respond to the needs and wishes of these specific groups.
Demotivated respondents. A large number of respondents, varying in educational background,
answered no more than 2 of the NVS-questions correctly. There are multiple potential explanations for this finding. First, respondents filled in the survey while they were on their way to do the groceries or to bring the groceries home. It might not have been a good moment for them to fill in a survey and answer five questions about a Ben & Jerry nutrition label. Second, the nutrition label could have made them nervous and feel judged. As health is nowadays turned into a moral subject, this can provoke strong emotions (Rozin, Markwith, & Stoess, 1997) and cause them to withdraw from the survey. Third, the way the nutrition information was presented is too annoying or difficult to read and understand in a short period of time for all sorts of people with different backgrounds and SES, and therefore they become demotivated. In order to stimulate healthy choices, the way nutrition information is presented can be
