For the first part, the Afrobarometer survey was used to analyze the factors that affect trust in the ruling party. Thus, the research question was the following.
RQ1: What factors influence trust in the ruling party of Africa’s residents?
Two hypotheses were formulated.
H1: The age of the respondent has a significant influence on the respondents’ trust in the ruling party.
H2: Perceived level of the country’s present economic condition has a significant influence on the respondents’ trust in the ruling party.
A regression model was created, where trust to the ruling party (Q59F) was used as the dependent variable, and age (Q1) and perceived country’s economic condition (Q3a) were used as the independent variables. The results of the analysis revealed that both independent variables were significant predictors of trust to the ruling party with p < 0.001. However, the overall effect size of the model was low, as the R2 coefficient was 0.051 (Wagner, 2019). After testing for the assumptions, no problems with the model were identified. The results of the analysis are provided in Tables 1-3.
Table 1. Descriptive statistics
Table 2. Model Summary
Table 3. Coefficients
The results of the analysis were controversial. On the one hand, both independent variables had a significant impact on the dependent variable, which demonstrated that both age and perceived economic conditions of the country could be used to predict trust in the ruling party. On the other hand, the effect size of the model was small, which demonstrated that manipulation of the independent variables would have a small effect on the independent variables. In other words, if the ruling party wanted to increase trust in it, increasing the country’s economic performance would be an inefficient strategy to follow. Thus, the model suggests that when trying to develop strategies for increasing the trust in the ruling party, policymakers should consider the age distribution of the population and the economic state of the country along with other possible predictors. Additional factors that affect trust in the ruling party should be included in the analysis to increase the effect size of the model.
Following the results of the analysis in Part 1, additional factors that may affect trust in the ruling party were assessed. The research question remained the same:
RQ1: What factors influence trust in the ruling party of Africa’s residents?
Two hypotheses were formulated.
H1: The gender of the respondent has a significant influence on the respondents’ trust in the ruling party.
H2: Perceived level of personal living conditions has a significant influence on the respondents’ trust in the ruling party.
A regression model was created, where trust to the ruling party (Q59F) was used as the dependent variable, and gender (Q101) and present living conditions (Q3b) were used as the independent variables. Gender was the dummy variable coded “1” for “Male” and “2” for “Female.” The results of the analysis revealed that both independent variables were significant predictors of trust in the ruling party with p < 0.001. However, the overall effect size of the model was low, as the R2 coefficient was 0.011, which was lower than the effect size achieved in Part 1 (Wagner, 2019). The coefficient for gender was 0.015, which implied that women tended to have greater trust in the ruling party than men. However, the difference was rather low, as the difference was only 0.015 on a scale from one to four. At the same time, an increase in satisfaction with living conditions also had a small effect on the trust in the ruling party. The coefficient for the variable was 0.099, which implied that the difference in trust to the ruling party between those who assessed their living conditions as “very bad” and those that thought that their conditions were “very good” was toughly 0.4 on the grid from one to four. SPSS output for the analysis is provided in Tables 1-5 below.
Table 1. Descriptive statistics
Table 2. Correlation analysis
Table 3. Model Summary
Table 4. Coefficients
Table 5. Residual statistics
According to Wagner (2019), the assumptions were not violated, which meant that the model was no biased. All the measurements were taken independently, and there are no significant correlations among the variables, demonstrating no-to-minimum collinearity. The analysis of residuals, including Cook’s D, DFFITS, and DFBETAS revealed no problems with highly influential cases, which was expected from Likert scale questions. The scatterplot of the residuals revealed no obvious pattern, which implies that the assumption of homoscedasticity was not violated. The results of the analysis suggest that the living conditions of the respondents had a small effect on trust in the ruling party, which suggests that policymakers need to manipulate more than just individual living conditions to increase trust in the ruling party.
References
Wagner, W. (2019). Using IBM® SPSS® statistics for research methods and social science statistics (7th ed.). SAGE Publications.