You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257642636

Perceptions of poverty attributions in Europe: A multilevel mixture model


approach

Article  in  Quality & Quantity · May 2014


DOI: 10.1007/s11135-013-9843-3

CITATIONS READS

20 454

2 authors:

Leonor Pereira da Costa José G. Dias


Universidade Lusófona de Humanidades e Tecnologias ISCTE-Instituto Universitário de Lisboa
12 PUBLICATIONS   270 CITATIONS    90 PUBLICATIONS   1,199 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FEINPT View project

CLAVE - The social development of human values in childhood and early adolescence View project

All content following this page was uploaded by Leonor Pereira da Costa on 21 May 2014.

The user has requested enhancement of the downloaded file.


Qual Quant
DOI 10.1007/s11135-013-9843-3

Perceptions of poverty attributions in Europe:


a multilevel mixture model approach

Leonor Pereira da Costa · José G. Dias

© Springer Science+Business Media Dordrecht 2013

Abstract The perception of the causes of poverty has long been recognized as a very impor-
tant factor in the broader study of this social phenomenon. This study covers 15 European
Union countries and considers three types of poverty attributions: individualistic, societal,
and fatalistic. The individualistic view perception believes the poor are responsible for their
situation, the societal perspective blames society for poverty and the fatalistic view considers
poverty to be the result of bad luck or fate. A multilevel mixture model with three clusters
of countries and six clusters of individuals was identified. Despite the generalization of the
social explanations of poverty at the individual or micro level, there are also groups that
emphasize more individualistic explanations, blaming the poor for their condition. At the
country or macro level, the most developed cluster believes in the individualistic and fatal-
istic causes of poverty, whereas the least developed clusters explain poverty based on the
injustices of society. There is diversity in the way these countries perceive poverty.

Keywords Multilevel mixture models · Poverty · Social attributions · Latent class analysis

1 Introduction

Poverty is a complex phenomenon as it results from a combination of many factors. Several


studies on the causes of poverty have concluded that there are economic and non-economic
factors that can lead to poverty (e.g., Dewilde 2004). Other authors concluded that poverty is
a multidimensional concept and factors such as the employment situation, low income level,
family dimension, marital status, alcoholism, education are significantly related to poverty
(e.g., Nándori 2011; Kim et al. 2010). In addition to these individual characteristics, Kim
et al. (2010) also found that poverty can be explained by the context of the country where

L. P. da Costa (B) · J. G. Dias


Instituto Universitário de Lisboa (ISCTE-IUL), Business Research Unit (UNIDE-IUL),
Lisbon, Portugal
e-mail: mleonorpc@gmail.com

123
L. P. da Costa, J. G. Dias

the poor live, namely state involvement and contributions to public social policies for the
reduction of poverty.
Our study focuses on the perceptions of the causes of poverty that have been of great
importance to a broader understanding of this social phenomenon. These perceptions have
direct implications on social interaction with the poor and, indirectly, on the legitimacy of
social support measures and the fight against poverty. Combating poverty and social exclusion
is only possible with the commitment of all social actors, hence the importance of studying
the perceptions of poverty and its causes.
Most studies on attitudes towards poverty focus on the causal attributions that indi-
viduals make to explain why the poor are poor on the assumption that these attributions
may be closely related to the attitudes of the poor (Cozzarelli et al. 2001). The theory of
causal attributions (Heider 1958) assumes that individuals try to understand why people do
what they do and this process may involve attributing one or more causes to the observed
conduct.
Studies on poverty consider three types of attribution for poverty: individualistic, societal
or structural, and fatalistic (Cozzarelli et al. 2001; Bullock 1999). In the first group, people
blame the poor for their own poverty; e.g., they consider the behavior of the poor to be
undesirable due to addictions like drugs and alcohol, laziness, weakness and lack of com-
mitment to overcome obstacles or low intelligence (Morçöl 1997; Cozzarelli et al. 2001).
From the societal perspective, social causes predominate and individuals blame society itself
or contextual factors (political, economic, cultural) for poverty, e.g., society did not make
jobs available for the poor or paid them poorly, the lack of government support or discrim-
inatory behavior from other citizens (Reutter et al. 2006). Finally, people who see poverty
as a result of the poor’s bad luck or fate e.g. they suffer from an illness/disability or have
been victims of an accident that has determined their future, attribute poverty to fatalistic
causes.
The study of the different perceptions of the causes of poverty is relevant in that
these perspectives have an impact on how each member of the society interacts with
the poor and, more generally, increases the legitimacy of policies to fight poverty. For
instance, it is believed that people who explain poverty in individualistic terms tend to
be less willing to accept spending on social security and social protection policies than
those who attribute poverty to social causes (van Oorschot and Halman 2000; Zucker
and Weiner 1993). On the other hand, the fatalistic view of poverty can lead to society’s
resignation towards poverty as it has no social responsibility for the phenomenon. This
resignation hinders society’s mobilization and active participation towards resolving the
problem.
When studying human behavior, cultural orientation is important as it explains how indi-
viduals process information. Consequently, it influences the formation and change of attitude
and the development of specific behaviors. National studies have therefore been developed in
different countries: Canada (Reutter et al. 2006), United States (Cozzarelli et al. 2001; Hunt
1996), Australia (Feather 1983), the Netherlands (van Oorschot and Halman 2000), Britain
(Furnham 1982), Finland (Niemelä 2008), Lebanon (Abouchedid and Nasser 2001), Iran
(Hayati and Karami 2005), Croatia (Ljubotina and Ljubotina 2007), among others. Although
comparisons cannot be made between these studies as different methodologies of question-
naire design, sampling, and data analysis were used (Nasser et al. 2002), there are a number
of studies that compare countries (Nasser et al. 2002; van Oorschot and Halman 2000; Kreidl
2000).
The studies undertaken by the European Commission (Eurobarometer surveys) conclude
that poverty is increasingly being explained by social causes in European countries (European

123
Perceptions of poverty attributions in Europe

Commission 1977, 1990, 2007, 2010). Between 1976 and 1990, the comparison between the
countries showed that individuals agreed less with the idea that poverty was caused by laziness
or lack of will, especially in countries where this idea had previously prevailed (United
Kingdom, Ireland, Luxembourg). Moreover, whereas 37 % of the respondents perceived
social injustice as a cause of poverty in 2007, this went up to 47 % in 2009 with increases
in all countries except Poland (European Commission 2010). Denmark was the only country
that did not indicate this as the major cause; indeed, 32 % of the Danes feel that the poor are
poor due to bad luck.
van Oorschot and Halman (2000) also analyzed data from the 1990 European Values Study
and concluded that the dimension of social blame is the most mentioned cause for people
living in poverty, and fatalism (bad luck) the least indicated. In this study, the countries were
organized into three groups: the Netherlands, West Germany, United Kingdom and Ireland
are in the pluralistic group that assigns the causes of poverty to all perspectives considered
(van Oorschot and Halman 2000); Sweden, Norway and France showed a more structuralist
position and do not consider the fatalism dimension; the United States of America, Italy,
Austria, Spain, Canada and Portugal assigned more blame as the cause of poverty.1 Out-
side of this group, Belgium revealed an individualistic perspective and Denmark a sense of
fatalism.
It appears that the Nordic countries tend to demonstrate a more structural attitude towards
the causes of poverty. However, comparative studies indicate that the Finnish people deviate
from the Scandinavian pattern as they are more likely to attribute poverty to individualistic and
fatalistic explanations (Niemelä 2008). On the other hand, social injustice is seen as the main
cause of poverty in the Netherlands (Halman and van Oorschot 1999). For van Oorschot and
Halman (2000), the Czech Republic stood out compared to other Eastern European countries:
a large proportion of Czechs blame the poor for their own state of poverty, while a small
proportion attribute it to social injustice. For these authors, although most Eastern European
countries have a structural approach, the blame for poverty is assigned to the poor more
frequently than in the Western countries.
This study aims to investigate Europeans’ perceptions of the causes of poverty and analyze
the differences and similarities between different countries. This type of cross-cultural study
has to take into account factor invariance by estimating a single model (see, e.g., Savelkoul
et al. 2012). We propose a multilevel taxonomy that models respondent and country levels,
simultaneously. Multilevel analysis studies the interactions between the different levels of a
hierarchical system (Kreft and de Leeuw 1998; Hox 2002). The differences and similarities in
the perceptions of the causes of poverty among individuals from different European countries
are analyzed. Multilevel mixture analysis allows the heterogeneity between the countries and
between individuals to be structured at the same time, identifying relatively homogenous
clusters at country and individual levels in light of poverty attributions. By retrieving these
differences and similarities, we can establish groups and explain what types of groups exist
at both levels: in a first level of analysis, we can group individuals within each country
according to their attitudes towards poverty; in addition, at a second level, and on the basis
of the distribution of groups formed at the first level, we can group countries with a similar
structure of individual clusters, i.e., two countries will be part of the same group if they have
similar profiles at the individual level. The remainder of the paper is organized as follows.
Section 2 describes the methodology. Section 3 introduces the sample and the indicators used

1 The second wave of the European Social Survey took place in 1990 in some non-European countries such
as Canada and United States of America.

123
L. P. da Costa, J. G. Dias

in this study. Section 4 provides the results and Section 5 ends the paper with concluding
remarks.

2 Methodology

Our aim is to provide a typology of the European countries in terms of perceptions of poverty
by estimating a multilevel mixture model (Vermunt 2003). This model extends the tradi-
tional mixture model and is important in that individuals living in the same country who
share specific characteristics are taken into account. The construction of the individuals’
identity, attitudes, and behaviors is inevitably influenced by their country. Using a multilevel
mixture model, a particular phenomenon can be analyzed at two or more distinct levels with
the simultaneous clustering at each level. Thus, the multilevel mixture model provides a
simultaneous analysis at two levels, for example: individual level, i.e., their profile within
each country in terms of their perceptions of poverty; country level, i.e., the similarities and
differences between European countries in this context.
A multilevel mixture model with two levels is defined by:

 nj  T 

M   
K
f (yi ) = P(W j = m) P(Z i j = t|W j = m) P(Yi jk |Z i j = t)
m=1 i=1 t=1 k=1

where j = 1, . . . , J is the set of countries, i = 1, . . . , n j corresponds to individuals, and


n j is the total number of individuals in country j. The response of individual i in country
j on indicator k = 1, . . . , K is Yi jk . The aim is to identify two latent discrete variables:
Z i j ∈ {1, . . . , T } and W j ∈ {1, . . . , M} represent clusters of individuals and countries,
respectively. T and M are the number of clusters at individual and country level, respectively.
The components of the model in the above equation are:

1. P(W j = m) is the probability that country j belongs to cluster m of countries. It gives


the size of the clusters at country level;
2. P(Z i j = t|W j = m) gives the probability that individual i in country j belongs to cluster t
of individuals conditional on the membership in cluster m of countries. These parameters
give the differences between the clustering of countries, more specifically, the relative
size of each cluster of individuals;
3. P(Yi jk |Z i j = t) gives the probability of individual i expressing a certain opinion con-
ditional on being in cluster of individuals t. It is precisely these latter probabilities of
conditioned responses to the individual belonging to a particular cluster that allow us to
understand the differences between individuals regarding the dependent variables. Here,
Yi jk are considered to be independent of each other within each cluster (local indepen-
dence). Moreover, the clusters in the lower level Z i j are mutually independent given
Wj.

Mixture model parameters can be estimated by maximum likelihood using either the EM
algorithm or extensions (Vermunt 2003; Dias and Vermunt 2007; Pirani 2012). The decision
on the number of clusters is based on the analysis of the Bayesian information criterion (BIC),
which takes into account model fit (the log-likelihood) and model complexity (number of
parameters).

123
Perceptions of poverty attributions in Europe

3 Sample and measures

The data set used in this research was collected by Eurobarometer 2007 (Papacostas 2007) in
15 European Union member states: Austria, Belgium, Denmark, Finland, France, Germany,
Great Britain, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, and
Sweden. Two of the countries surveyed—Great Britain and Germany, were divided into two
regions, West and East Germany and Great Britain and Northern Ireland, respectively. Thus,
the upper level of the analysis contains 17 groups (13 countries and 4 regions) and the lower
level is defined by 15504 individuals aged over 15 years.
At the individual level we use six dichotomous variables plus a nominal variable with
five response categories. Each of the three types of poverty attributions is measured by
two dichotomous variables (Table 2). In the individualistic or internal attribution, the
explanations given were the poor’s addictive behavior and inertia; in the structural attri-
bution, people explained poverty by referring to either the poor’s low level of benefits
or low income; the fatalists explained poverty by illness/disability or (the fact that indi-
viduals suffered from) a family breakdown. The seventh nominal variable is on why
there are people in need. The possible answers correspond to one of the abovementioned
attributions.

4 Results

We estimated 100 models for M = 1, . . . , 10 and T = 1, . . . , 10. We did not obtain a


minimum value for the BIC measure. After analyzing the cluster sizes and cluster profiles,
we chose the model with M = 3 and T = 6, i.e., three clusters at country level and six at
individual level. Although social causes are the most quoted by individuals (van Oorschot
and Halman 2000), these results show heterogeneity within clusters.
At the individual level, the size of six clusters ranges from 9.4 to 33.0 % (Table 1).
The perception of the causes of poverty in each cluster is described in Table 2. It gives
the estimated probability for each cause of poverty within each individual level cluster,
P(Yi jk |Z i j = t).
Firstly, the results show a separation between individuals who have a more individualistic-
oriented perspective to the causes of poverty and those who argue that society is the main
cause of poverty: two clusters have a more societal tendency explaining poverty, whereas three
clusters explain poverty based on intrinsic characteristics of the poor. The fatalistic causes
only stand out as the main causes for one of the clusters, combined with the explanation of
drug and alcohol consumption. In short, it is individuals in the third and sixth clusters that are
more society oriented, whereas those in the first, second, and fifth clusters consider that the
poor themselves are responsible for their poverty. The fatalistic causes are only mentioned
by the second cluster. The fourth cluster is the largest, and in general can be characterized
precisely by mild opinions on the dimensions assessed.

Table 1 Cluster sizes (individual level)

t =1 t =2 t =3 t =4 t =5 t =6

P(Z i j = t) 0.075 0.178 0.167 0.330 0.094 0.157

123
L. P. da Costa, J. G. Dias

Table 2 Cluster probabilities at individual level

Attributions Individual-level clusters Aggregate

1 2 3 4 5 6

Social benefits/Pensions 0.152 0.034 0.546 0.000 0.128 0.997 0.302


are not sufficiently high
in our country
Their current jobs don’t 0.249 0.060 0.998 0.256 0.090 0.000 0.325
pay enough
They suffer from 0.164 0.587 0.047 0.255 0.989 0.187 0.291
alcoholism, drug abuse
or other addictions
They don’t do enough to 0.992 0.112 0.026 0.018 0.247 0.030 0.129
overcome their
problems
They suffer from 0.106 0.498 0.110 0.241 0.152 0.228 0.221
long-term illness or
disability
They suffered some 0.135 0.311 0.094 0.299 0.183 0.201 0.220
family breakdown, or
have lost a family
member
Bad luck 0.148 0.279 0.166 0.230 0.155 0.216 0.207
Laziness and lack of 0.428 0.086 0.088 0.166 0.530 0.113 0.185
determination
Unfair society 0.187 0.224 0.533 0.350 0.130 0.468 0.361
Part of progress 0.145 0.226 0.137 0.108 0.102 0.124 0.131
Don’t Know 0.092 0.185 0.076 0.145 0.083 0.079 0.115

Focusing first on the structural clusters, it is individuals in the sixth cluster who attribute
poverty mainly to the fact that pensions or social benefits are not sufficiently high in their
country, with a very high probability (0.997). In this cluster, there is a null probability that
poverty is due to low wages (another social cause). The results also support the structural
orientation of this cluster as a result of the high probability of individuals stating that social
injustice (0.468) explains the existence of people in need.
The second structural cluster is the third one. This is distinguished from the sixth cluster
mainly because individuals in this cluster consider that both low social benefits (0.546) and
also low wages (0.998) are the main causes for being poor. Within this group, there is a null
probability of individuals advocating any of the internal causes. When asked about why there
are people living in need, it is likely individuals will report that it is due to social injustice
(0.533). Thus, this cluster is the most pro societal in terms of causes of poverty.
In the more individualistic group of clusters, the reasons stated for people being poor or
excluded from the society are heterogeneous.
Cluster 5 (9 % of the sample) contains the individuals who attribute poverty largely
to individual causes, and particularly that people are poor or excluded in our society as a
result of addictions such as drug or alcohol abuse (0.989). The second most important item
characterizing this cluster is the perception that the poor do not do enough to overcome their
problems (0.247). Respondents in this cluster also give laziness and lack of determination
as a cause for poverty (0.530). In the first cluster, almost all individuals considered laziness

123
Perceptions of poverty attributions in Europe

to be a cause for poverty (0.992). Laziness and lack of determination is also considered a
relevant reason for poverty for the individuals in this cluster (0.428).
Combining individualistic and fatalistic reasons to explain poverty, the second cluster
refers to the importance of addictions to this state (0.587). Moreover, it refers that poverty is
due to long-term illness and diseases (0.498) of the poor and family problems (0.311). This
is the only cluster of individuals where the fatalistic explanations gain relevance.
The fourth cluster is the largest with one third of the individuals. It is the most heteroge-
neous cluster and individuals do not present a strong position on any of these dimensions. They
consider structural, individualistic and fatalistic causes of poverty, simultaneously. However,
this group does not blame society for poverty (because the probability of low social benefits
is null); they also do not blame individuals for not doing enough to overcome their status
(0.018). This is in agreement with Kreidl (2000): there are different views to explain the
causes of poverty and they are not mutually exclusive. For instance, a strong belief in an
individualistic cause does not invalidate that the individual can also believe that poverty is
caused by society (Kreidl 2000). Table 3 summarizes the clusters at individual level based
on the discussion of the most important dimensions highlighted above.
Table 4 provides the clustering at country level, i.e., how countries join each other. At
first, we see that most countries under analysis have a maximum probability of belonging to
a given cluster. Luxembourg is an exception as it overlaps the three clusters (mostly the third
cluster).
The smaller cluster comprises three of the most developed countries: Denmark, the Nether-
lands, and Sweden. The countries of the central and Western Europe are divided into two
clusters. Most countries belong to cluster 3. Figure 1 provides the distribution of the countries
across the three clusters.

Table 3 Labels of clusters at


individual level 1 Lazy and with lack of determination
2 Addicted and with bad luck
3 Society victims
4 Pluralistic vision
5 Addicted and lazy
6 Low social benefits victims

Table 4 Country-level clusters

Country clusters Relative weights Country Probability

1 0.188 Denmark, the Netherlands and Sweden 1.000


Luxembourg 0.050
2 0.241 Germany (East), Greece, France and Italy 1.000
Luxembourg 0.003
3 0.571 Belgium, Germany (West), Spain, Finland 1.000
Ireland, Austria, Portugal 1.000
Great Britain and Northern Ireland 1.000
Luxembourg 0.946

123
L. P. da Costa, J. G. Dias

Fig. 1 Country level classification (maximum a posteriori probability)

For each cluster of countries we computed the average Human Development Index (HDI)
values to characterize their level of development (Table 5). Although these countries have
an overall high HDI, cluster one has the highest and is composed of the countries with a
more advanced Welfare State. The third cluster has the lowest HDI and comprises almost all
countries of Western and Central Europe.
The distribution of the individual-level cluster within country-level clusters, i.e., the con-
ditional probability of individuals belonging to the individual cluster t, given the respondent
i belongs to the country-cluster m, P(Z i j = t|W j = m), is given in Table 6.
In cluster one, individuals attribute the causes of poverty to the poor’s addiction and bad
luck. In these countries, a well established Welfare State takes full responsibility for ensuring
public services and the population’s protection. Thus, that may explain why individuals do
not blame the society for the existence of poverty in the country.
The second cluster has a more pluralistic view of the causes of poverty: although they
think that poor people are poor because they are lazy and unwilling, they also believe that they
are victims of society and low social benefits. But in this country-level cluster, the pluralistic
group that assigns all causes to poverty is well represented. This is the group of countries

123
Perceptions of poverty attributions in Europe

Table 5 Country-level clusters

Country clusters Country HDI 2007 Min–Max Poverty risk Min–Max (%)
(average) (average, %)

1 Denmark, the 0.961 0.955–0.964 12 11–12


Netherlands,
Sweden
2 Germany (East), 0.950 0.942–0.961 17 13–20
Greece, France,
Italy
3 Belgium, Germany 0.950 0.909–0.965 16 12–20
(West), Spain,
Finland, Ireland,
Austria, Portugal,
Great Britain,
Northern Ireland,
Luxembourg
Source UNDP (2009) and Eurostat (2010)

Table 6 Distribution of individual-level clusters within country-level clusters

Individual-level clusters Country-level clusters Aggregate

m=1 m=2 m=3

t=1 Lazy and with lack of determination 0.051 0.103 0.070 0.075
t=2 Addicted and with bad luck 0.685 0.002 0.085 0.178
t=3 Society victims 0.048 0.306 0.147 0.167
t=4 Pluralistic vision 0.064 0.406 0.385 0.330
t=5 Addicted and lazy 0.064 0.000 0.143 0.094
t=6 Low social benefits victims 0.088 0.184 0.169 0.157

that blames society most for poverty. If we analyze the poverty risk index, we see that this
cluster has more risk than the other clusters.
The third cluster is very similar to the second one with regard their composition of indi-
vidual clusters. The main difference with the second cluster is that more individuals give
addiction and laziness as the reason for being poor. In these countries, the group of individ-
uals that blame society for poverty is not as representative as in the second group.
Based on the fact that these two clusters have a greater risk of poverty than the first, it
can be concluded that one of the factors distinguishing the clusters is precisely individuals’
involvement with poverty (e.g., their potential economic situation and their own financial
situation). Given this personal experience, the respondents may be more sympathetic to this
social phenomenon. Thus, those either in need or interacting with those in need tend to
analyze the situation and assign it to external factors, blaming society for poverty. On the
other hand, groups of individuals without difficulties consider poverty to be the outcome of
individual behavior.
At the country level, the human development of the country and the risk of poverty
help explain country clustering: more developed countries are grouped in cluster one
(Denmark, Sweden, and the Netherlands); Western and Central Europe is divided into two
clusters. Overall, we conclude that the cluster containing the most developed countries, with

123
L. P. da Costa, J. G. Dias

higher socioeconomic status and attaining higher educational levels, gives individualistic and
fatalistic causes for poverty. The clusters in Central and Western Europe show greater het-
erogeneity in how they perceived poverty, explaining it mainly by social injustice.

5 Conclusion

This research performed a joint analysis of perceptions on poverty across 15 European coun-
tries before the enlargement to the East. The statistical analysis of the attributions of poverty
from a multilevel and multidimensional point of view proves an excellent tool for identifying
taxonomies at different levels of analysis. Apart from profiling the individuals within coun-
tries based on the individual responses, it provides a country level typology. This study vali-
dates the importance of the hierarchical structure in these studies: individuals are inevitably
influenced by their own country’s cultural issues, history and policies, and these influences
play a major role in their opinion and perception of the causes of poverty (Pirani 2012). This
cross-country study provides a deeper and richer data analysis than traditional clustering
methods as it takes into account different levels of analysis simultaneously.
Further research can explore different social phenomena in cross-cultural settings, comb-
ing different manifest variables, countries being surveyed, and individual units (e.g., compa-
nies, households). Some research focused on the study of socio-demographic variables that
may influence the perception of the causes of poverty (Hunt 1996; Bullock 1999; Cozzarelli
et al. 2001). However, there is no clear evidence of the role of these variables as predictors
of causal attributions of poverty as there are diverse results in several different studies (Coz-
zarelli et al. 2001). In future research we would like to extend these results into a predictive
framework by taking into account the characteristics of the individuals (e.g., sex, age, acad-
emic qualifications, economic situation, or work occupation). We would also like to further
the analysis of the influence of personal experience, namely regarding individuals’ economic
situation.

Acknowledgments This research was supported by the Fundação para a Ciência e Tecnologia (Portugal),
Grant PTDC/CS-DEM/108033/2008.

References

Abouchedid, K., Nasser, R.: Poverty attitudes and their determinants in Lebanon’s plural society. J. Econ.
Psychol. 22, 271–282 (2001)
Bullock, H.: Attributions for poverty: a comparison of middle-class and welfare recipient attitudes. J. Appl.
Soc. Psychol. 29(10), 2059–2082 (1999)
Cozzarelli, C., Wilkinson, A., Tagler, M.: Attitudes toward the poor and attributions for poverty. J. Soc. Issues
57(2), 207–227 (2001)
Dewilde, C.: The multidimensional measurement of poverty in Belgium and Britain: a categorical approach.
Soc. Indic. Res. 68, 331–369 (2004)
Dias, J., Vermunt, J.: Latent class modeling of website users’ search patterns: implications for online market
segmentation. J. Retailing Consum. Serv. 14(6), 359–368 (2007)
European Commission (1977) The perception of poverty in Europe (Eurobarometer 5), Brussels. http://ec.
europa.eu
European Commission (1990) The perception of poverty in Europe in 1989 (Eurobarometer 32), Brussels.
http://ec.europa.eu
European Commission (2007) Poverty and exclusion (Special Eurobarometer 279/Wave 67.1), Brussels. http://
ec.europa.eu

123
Perceptions of poverty attributions in Europe

European Commission (2010) Poverty and social exclusion (Special Eurobarometer 321/Wave 72.1), Brussels.
http://ec.europa.eu
Eurostat (2010) Database with at-risk-of-poverty rate after social transfers. http://epp.eurostat.ec.europa.eu
Feather, N.: Causal attributions and beliefs about work and unemployment among adolescents in state and
independent secondary schools. Aust. J. Psychol. 35, 211–232 (1983)
Furnham, A.: Explanations for unemployment in Britain. Eur. J. Soc. Psychol. 12, 335–352 (1982)
Halman, L., van Oorschot, W.: Popular perceptions of poverty in Dutch Society. WORC Paper 99.11.01,
Tilburg University, Tilburg (1999)
Hayati, D., Karami, E.: Typology of causes of poverty: the perception of Iran farmers. J. Econ. Psychol. 26,
884–901 (2005)
Heider, F.: The Psychology of Interpersonal Relations. Wiley, New York (1958)
Hox, J.: Multilevel Analysis: Techniques and Applications. Lawrence Erlbaum Associates, Mahwah (2002)
Hunt, M.: The individual, society, or both? A comparison of black, latino, and white beliefs about the causes
of poverty. Soc. Forces 75(1), 293–322 (1996)
Kim, K., Lee, Y., Lee, Y.: A multilevel analysis of factors related to poverty in welfare states. Soc. Indic. Res.
99, 391–404 (2010)
Kreft, I., de Leeuw, J.: Introducing Multilevel Modeling. Sage, London (1998)
Kreidl, M.: Perceptions of poverty and wealth in Western and post-communist countries. Soc. Justice Res.
13(2), 151–176 (2000)
Ljubotina, O., Ljubotina, D.: Attributions of poverty among social work and non-social work students in
Croatia. Croatian Med. J. 48, 741–749 (2007)
Morçöl, G.: Lay explanations for poverty in Turkey and their determinants. J. Soc. Psychol. 137(6), 728–738
(1997)
Nándori, E.: Subjective poverty and its relation to objective poverty concepts in Hungary. Soc. Indic. Res.
102, 537–556 (2011)
Nasser, R., Abouchedid, K., Khashan, H.: Perceptions of the causes of poverty comparing three national
groups: Lebanon, Portugal, and South Africa. Curr. Res. Soc. Psychol. 8(7), 101–118 (2002)
Niemelä, M.: Perceptions of the causes of poverty in Finland. Acta Sociologica 51(1), 23–40 (2008)
Papacostas A (2007) Eurobarometer 67.1: Cultural values, poverty and social exclusion, developmental aid
and residential mobility, Brussels. http://www.icpsr.umich.edu
Pirani E (2012) Evaluating contemporary social exclusion in Europe: a hierarchical latent class approach.
Qual. Quant. 47(2), 923–941
Reutter, L.I., Veenstra, G., Stewart, M.J., Raphael, D., Love, R., Makwarimba, E., McMurray, S.: Public
attributions for poverty in Canada. Can. Rev. Sociol. Anthrop. 43(1), 1–22 (2006)
Savelkoul, M., Scheepers, P., van der Veld, W., Hagendoorn, L.: Comparing levels of anti-Muslim attitudes
across western countries. Qual. Quant. 46(5), 1617–1624 (2012)
UNDP (2009) Human Development Report. Overcoming barriers: human mobility and development. UNDP,
Washington, DC. http://hdr.undp.org
van Oorschot, W., Halman, L.: Blame or fate, individual or social: an international comparison of popular
explanations of poverty. Eur. Soc. 2(1), 1–28 (2000)
Vermunt, J.K.: Multilevel latent class models. Sociol. Methodol. 33, 213–239 (2003)
Zucker, G., Weiner, B.: Conservatism and perceptions of poverty: an attributional analysis. J. Appl. Soc.
Psychol. 23, 925–943 (1993)

123
View publication stats

You might also like