The attribution theory of poverty has been a significant focus of academic inquiry, particularly from the 1970's to 2010. However, the research is further scant in the later years on this topic. Nevertheless, this study attempts to explore the most relevant literature on the theory, focusing on how demographic groups can be targeted for charitable purposes by examining their beliefs about poverty and charity.

Initially, Simmel (1908), cited in Oorschot and Halman (2000), argued that beliefs about poverty shape welfare policies, with public attitudes determining whether poverty is an individual or societal issue. In the 1970's, studies (Alston and Dean, 1972) found that white Americans predominantly blamed the poor for their circumstances, with younger and more educated respondents being less sympathetic. The attribution theory, developed by Heider (1958), categorizes explanations of behavior into internal (personal) and external (situational) factors, later linked to poverty by Feagin (1972). Feagin identified individualistic, structural, and fatalistic explanations for poverty.

Subsequent research has expanded on these ideas. Feather (1974) found that low-income Australians attributed poverty to societal factors. Furnham (1982) revealed that Labor Party supporters blamed societal factors, while conservatives pointed to individualism. On the other hand, race influences poverty attributions as the difference of opinion could be observed between different races in the United States of America (Hunt, 1996, 2002, 2004, 2016; Bastias et al., 2024). Similar patterns were observed in Russia and Estonia during economic downturns (Stephenson, 2000). Therefore, the first hypothesis of this study is:

Sociodemographic factors have a significant impact on the beliefs about poverty (Hunt, 1996; Bastias et al., 2019). Different demographic factors are tested against the individualistic and structural explanations of poverty. For instance, a study conducted in Malaysia finds that younger adolescents tended to blame societal factors for poverty, while older ones held both individualistic and structural beliefs (Halik and Webley, 2011), and a similar pattern could be found in the other countries (Bergmann and Todd, 2019; Alcañiz-Colomer et al., 2023). Similarly, women were more likely to endorse individualistic explanations (Halik et al., 2012) contrasting with earlier findings (Kluegel and Smith, 1986; Hunt, 1996, 2004). However, in some recent studies, it is found that women give structural explanations of poverty as compared to their male counterparts (Terol Cantero et al., 2023). Similarly, Khan et al. (2022), in the context of Malays, found that women with structuralist views are more philanthropic, highlighting the role of beliefs in shaping pro-poor attitudes. Therefore, this study hypothesizes that:

Additionally, to answer the second objective of the study, this research sheds light on how these beliefs influence people's willingness to help those experiencing poverty. For example, research shows that individuals who attribute poverty to individualistic factors tend to oppose pro-poor policies, whereas those who endorse structural explanations of poverty are more sympathetic toward the poor and favor such policies (Feagin, 1972; Feather, 1974). Similarly, Tuennerman-Kaplan (2001) suggests that people with structural beliefs about poverty are likelier to engage in philanthropy. However, previous studies have not provided empirical evidence to support these claims. This gap is addressed by Khan et al. (2022), who found that women are more likely to hold structuralist views, with their philanthropic nature being a key factor behind this perspective. They also identified the sociodemographic groups who have positive opinions about poverty. However, the study is limited to the Malays ethnic group only. Therefore, this study focuses on inter-ethnic comparison where beliefs about charity mediate the relationship between sociodemographic factors and structural beliefs about poverty.

To achieve the objectives of this study, the next section presents the methodology, detailing the sampling size and the development of the questionnaire. Following this, the results are reported in the finding section and contextualized within the discussion section. Finally, the conclusion provides a summary of the findings, highlights the study's limitations, and discusses its implications.

This study uses a non-probability sampling technique due to the non-availability of proper databases and limited resources. Selangor is selected as a sample because of its diverse population and the non-probability sampling technique is used to select the respondents for the study. Next, a sample size of 384 is identified based on Krejicie and Morgan's sample size table (Krejicie and Morgan, 1970); however, around 700 questionnaires were distributed.

The quota sampling technique is then applied, dividing the sample into nine districts. To reduce sample bias, first, the sample is categorized based on the population of Malays, Chinese, and Indians in Selangor and then further classified into males and females of each ethnic group. The details regarding sample distribution can be seen in Table 1.

The data was collected by approaching the respondents personally. The respondents were informed about the purpose of data collection and their consent was taken before handing over the questionnaire. Later, the incomplete and unfilled questionnaires were discarded, and 520 responses were retained for final analysis.

The questionnaire for this study is adapted from the Khan et al. (2022) study. In their research, four of six independent variables are adopted from previous studies such as Hunt (1996, 2002, 2004), while the authors introduce two. This study considers 11 sociodemographic factors (independent variables) to cover the broader perspective. The factors include ethnicity, gender, marital status, occupation, residence, monthly expenditure, age, dependents, income, wealth, and education.

The dependent variables in their study, i.e., structural, individualistic, and fatalist factors, are adopted from Feather (1974), Furnham (1982), Smith and Stone (1989), and Hunt (2004). The responses were measured on the Likert scale of 1 (Totally disagree) to 6 (Totally agree). Next, the study constructed beliefs about charity variables. This variable is used as a mediating variable in their research. In the current study, the same variable is used for mediation analysis. The reliability analysis for these variables based on Cronbach alpha is reported in Table 2.

Statistical software packages such as SPSS version 20 and Stata version 17 are used to perform the analysis. Due to the extended sample of this study, factor analysis is performed, and the results are extracted using a principal component method with Eigenvalues of >1. Items carrying a factor loading of >0.5 are retained for each dependent variable. The results are reported in Table 3. Moreover, the factor analysis results of the mediating variable are represented in Table 4.

In the initial model, monthly expenses and marital status were statistically significant at 10%. This suggests that, when comparing Malay, Indian, and Chinese females, there is a modest inclination to believe that poverty is mostly attributed to a lack of education and abilities. However, when demographic factors are considered in the second model of individualistic beliefs, the significance of the coefficients for marital status and monthly expenses has improved. Additionally, inherited wealth substantially impacts the development of individualistic beliefs. Education plays a crucial role in shaping Individuals' perspectives, especially regarding socioeconomic status. Specifically, females with limited educational opportunities who possess inherited wealth tend to adopt more individualistic ideologies that perpetuate the cycle of poverty than males.

In the second model of structuralist beliefs, the coefficient for Chinese ethnicity is positively and statistically significant at the 1% level. The coefficient of inherited wealth is also statistically significant at the 1% level. Education has a notable negative impact on structural beliefs. It may be inferred that, when compared to individuals of other ethnic backgrounds, Chinese females with higher levels of education are less likely to support the socioeconomic element as a cause of poverty. On the other hand, women with higher degrees of inherited wealth seem to have a stronger inclination toward embracing the structuralist perspective on poverty compared to men. The predominant tendency is to attribute poverty mostly to societal factors or the absence of income-generating alternatives.

In the initial model, when comparing Malay ethnicity to Chinese and Indian ethnicities, the latter two groups exhibit a statistically significant positive influence on fatalistic attitudes at the 1% level. The coefficient of gender shows a positive and statistically significant relationship at the 1% level. Males of Chinese and Indian descent exhibit a higher tendency to endorse fatalistic ideas regarding poverty compared to females. In contrast, the occupation variable shows a significant negative association with fatalistic attitudes at a 5% level. Chinese and Indian males engaged in private-sector employment are less likely to endorse fatalistic beliefs regarding poverty.

After accounting for demographic factors, the second model enhances the relevance of private sector occupation. The analysis reveals a statistically significant association between marital status and fatalistic belief at a 5% level. Married males tend to hold the opinion that certain individuals are predetermined to experience poverty. The data also indicates a statistically significant negative relationship between income and fatalistic thoughts about poverty at the 1% level. This suggests that individuals with lower income levels place greater emphasis on fatalistic beliefs about poverty. The coefficient of education has exhibited a positive relationship in this context, with a significance level of 5%. The presence of inherited wealth significantly positively influences fatalistic beliefs, as determined at a 5% level.

Table 8 explains the role of Charity (the mediator variable M) in the relationship between demographic factors (independent variable X) and structural beliefs about poverty (the dependent variable Y) for the Malay ethnic group. The results show that the total and direct effects of gender on the structuralist belief about poverty are insignificant, respectively. However, when we included charity as a mediating variable in our model, the relationship between gender and structuralist belief became significant. Thus, the indirect effect of gender on structuralist beliefs through charity is statistically significant. These findings indicate that gender does not directly affect the structural belief about poverty rather, but through charity, it does.

Similarly, the relationship between education and income and structuralist factors and wealth and structuralist factors is also mediated by charity. With increased income levels, Malays become more sympathetic toward the poor because of their philanthropic nature. Conversely, with the decreased education level, Malays have become more sympathetic toward the poor because of their philanthropic nature. The results are aligned with the study of Khan et al. (2022), which suggests that beliefs about charity mediate the relationship between these variables.

Table 9 shows the mediation analysis results for the Chinese ethnic group. The results show that the beliefs about charity mediate the relationship between gender and structuralist beliefs and education and structuralist beliefs. It means that Chinese males are more concerned about people experiencing poverty because of their strong philanthropic beliefs. However, this result contrasts with the Malay group, where the same result holds for the female group. Similarly, the low-level education holders are more concerned about the poor and hold societal factors responsible for poverty because they are more philanthropic.

Table 10 reports the mediation analysis results of the Indian ethnic group. The results show that the relationship between gender, education, income, and wealth with structuralist beliefs is mediated by beliefs about charity. Like the Chinese group, Indian males hold more structuralist views about poverty because of their strong philanthropic beliefs. Also, with a decrease in education level, the structuralist beliefs strengthen due to the strong charitable beliefs of the Indian ethnic group. Like Malays, with the increase in income and wealth, the beliefs of the Indians become more structuralist due to their strong beliefs about doing charity.