WK11Assgn

1 Categorical Variables with Chi-Square Tests: Afrobarometer DrPh, Walden University RSCH-8210R: Quantitative Reasoning and Analysis Dr. Nancy Rea August, 2020

2 Categorical Variables with Chi-Square Tests: Afrobarometer Categorical analysis provides information on relationships between other types of variables that are not continuous (Frankfort-Nachmias et al., 2020). The Afrobarometer dataset will be used to verify the utilization of the dataset by conducting the mean age of the respondents (IBM SPSS vs25, 2015). Three scenarios will be presented with data analysis. The analysis will include a research question, hypothesis and explanation of the variables. The interpretation of the data will encompass the case summary, crosstabulation, chi-square tests, systemic measures, and assumptions. As in previous papers, social change implications will be addressed. Verification of Dataset Table 1, 37.39 is the mean of the age of respondents. Table 1 Age of Respondents N Valid 10232 Missing 85 Mean 37.39 Median 34.00 Std. Deviation 14.863 Minimum 18 Maximum 105 From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Scenario 1: Categorical Analysis Research Question and Hypotheses Research question for scenario one is to what extent is there an association between trust in the police and democracy (dichotomous). Null hypothesis there is no association between trust in the police and democracy. Alternate hypothesis there is an association between trust in the police and democracy.

4 Data Interpretation Tables 4-6 Table 4 depicts the crosstabulation table of democracy and trust of the police. When comparing countries with a democracy and not a democracy and their trust in the police, in countries where there is no democracy 19.8% have no trust in the police and in countries with democracy 80.2% have no trust in the police based on 2,126 of the respondents that answered the question Table 4. No democracy 12.6% has a little trust in the police and with a democracy 87.4% have little trust in the police, which is how 2,363 of the respondents that answered the question Table 4. No democracy 9.3% have somewhat trust in the police and with a democracy 90.7% have somewhat trust in the police of the 2,515 respondents that answered the question Table 4. In Table 4, no democracy 5.6% have a lot of trust in the police and with a democracy 94.4% have a lot of trust in the police of the 2,446 respondents that answered the question. Table 5 provides the chi-square test results, X 2 (3) = 243.109; p < .001 and based on p < .05 significance level, there is significance. The null hypothesis is rejected. The assumption of cell expected frequencies should be less than 5 for at least 20% of the cells (Laureate Education, 2019). Table 5 shows the cell count at 0 cells (0.0%) have expected count less than 5, which means the assumption is met. Table 6 will be evaluated because Table 4 is larger than a 2x2 model and phi test will not be used but instead the Cramer’s V will be used to determine the strength of the effect. Table 6 has the Cramer’s V as .160, which reveals the has a low association (AcaStat Statistical Software, n.d.).

5 Table 4

7 Table 5 Chi-Square Tests Value df Asymptotic Significance (2- sided) Pearson Chi-Square 243.109 a 3 .000 Likelihood Ratio 239.354 3 .000 Linear-by-Linear Association 233.990 1 .000 N of Valid Cases 9450 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 244.77. From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Table 6 Symmetric Measures Value Approximate Significance Nominal by Nominal Phi .160 .000 Cramer's V .160 .000 N of Valid Cases 9450 From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Scenario 2: Categorical Analysis Research Question and Hypotheses Research question for scenario two is to what extent is there an association between respondents that reside in rural, urban or semi-urban settings and trust in the police. The null hypothesis is there is no association between respondents that reside in rural, urban, or semi- urban settings and trust in the police. The alternative hypothesis there is an association between respondents that reside in rural, urban, or semi-urban settings and trust in the police. Variables and Assumptions

8 The independent variable is respondents that reside in rural, urban, or semi-urban settings and the dependent variable is trust in the police. The variables levels are mutually exclusive; therefore, the assumption is fulfilled. Descriptive Statistics Table 7 Table 7 reveals, respondents that reside in rural, urban, or semi-urban settings ( M = 1.63 SD = .515, N = 10317), and trust in the police ( M = 1.57, SD = 1.10, N = 10098). The DV is trust in the police and IV is reside in rural, urban, or semi-urban settings. Table 7 Descriptive Statistics N Minimum Maximum Mean Std. Deviation Urban or Rural Primary Sampling Unit 10317 1 3 1.63 .515 Q59h. Trust police 10098 0 3 1.57 1.103 Valid N (listwise) 10098 From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Data Interpretation Tables 8 Table 8 shows, 10,098 of people answered the questions, which reflects 97.9% and 219 respondents did not answer the questions, which reflects 2.1% of the missing data. Ta ble 8 Case Processing Summary Cases Valid Missing Total N Percent N Percent N Percent Urban or Rural Primary Sampling Unit * Q59h. Trust police 10098 97.9% 219 2.1% 10317 100.0% From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015.

10 Table 9 Urban or Rural Primary Sampling Unit * Q59h. Trust police Crosstabulation Q59h. Trust police Total Not at all Just a little Somewh at A lot Urban or Rural Primary Sampling Unit Urban Count 1044 1053 1039 730 3866 Expected Count 866.4 957.9 1029.5 1012.3 3866.0 % within Urban or Rural Primary Sampling Unit 27.0% 27.2% 26.9% 18.9% 100.0% % within Q59h. Trust police 46.1% 42.1% 38.6% 27.6% 38.3% % of Total 10.3% 10.4% 10.3% 7.2% 38.3% Rural Count 1193 1409 1609 1855 6066 Expected Count 1359.4 1503.0 1615.3 1588.3 6066.0 % within Urban or Rural Primary Sampling Unit 19.7% 23.2% 26.5% 30.6% 100.0% % within Q59h. Trust police 52.7% 56.3% 59.8% 70.2% 60.1% % of Total 11.8% 14.0% 15.9% 18.4% 60.1% Semi- Urban Count 26 40 41 59 166 Expected Count 37.2 41.1 44.2 43.5 166.0 % within Urban or Rural Primary Sampling Unit 15.7% 24.1% 24.7% 35.5% 100.0% % within Q59h. Trust police 1.1% 1.6% 1.5% 2.2% 1.6% % of Total 0.3% 0.4% 0.4% 0.6% 1.6% Total Count 2263 2502 2689 2644 10098 Expected Count 2263.0 2502.0 2689.0 2644.0 10098. 0 % within Urban or Rural Primary Sampling Unit 22.4% 24.8% 26.6% 26.2% 100.0% % within Q59h. Trust police 100.0% 100.0% 100.0% 100.0 % 100.0% % of Total 22.4% 24.8% 26.6% 26.2% 100.0% From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015.

11 Table 10 Chi-Square Tests Value df Asymptotic Significance (2- sided) Pearson Chi-Square 204.896 a 6 .000 Likelihood Ratio 209.423 6 .000 Linear-by-Linear Association 182.724 1 .000 N of Valid Cases 10098 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 37.20. From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Table 11 Symmetric Measures Value Approximate Significance Nominal by Nominal Phi .142 .000 Cramer's V .101 .000 N of Valid Cases 10098 From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Scenario 3: Categorical Analysis Research Question and Hypotheses Research question for scenario three is to what extent is there an association between extent of democracy and current economic conditions. Null hypothesis there is no association between the extent of democracy and current economic conditions. Alternate hypothesis there is an association between extent of democracy and current economic conditions. Variables and Assumptions

13 Table 14-16 Interpretation Table 14 Table 14 shows the crosstabulation of the extent of democracy and the current economic conditions. In reviewing the extent of democracy, the present economic condition as very bad ( N = 2505) was as follows, 17.4% ( N = 436) not a democracy, 39.4% ( N = 987) democracy with major problems, 29.5% ( N = 739) democracy with minor problems, 13.7% ( N = 343) full democracy Table 14. Table 14 those economic conditions fairly bad ( N = 2686) responses is as follow, 12.7% ( N = 341) not a democracy, 36% ( N = 968) democracy with major problems, 35.1% ( N = 943) democracy with minor problems, 16.2% ( N = 434) full democracy. Those economic conditions neither good or bad ( N = 1564) responses is as follow, 8.8% ( N = 137) not a democracy, 28.7% ( N = 449) democracy with major problems, 43.6% ( N = 682) democracy with minor problems, 18.9% ( N = 296) full democracy in Table 14. Table 14 also depicts the present economic condition as fairly good ( N = 2297) was as follows, 6.8% ( N = 157) not a democracy, 25% ( N = 575) democracy with major problems, 42.6% ( N = 979) democracy with minor problems, 25.5% ( N = 586) full democracy. Lastly, Table 14 shows, the present economic condition as very good ( N = 432) was as follows, 5.3% ( N = 23) not a democracy, 22% ( N = 95) democracy with major problems, 36.3% ( N = 157 democracy with minor problems, 36.3% ( N = 157) full democracy Table 14. Table 15 provides the chi-square test results, X 2 (12) = 500.943; p < .001 and conventional significance level has been set at p < .05, significance has been demonstrated. Reject the null hypothesis. Table 16 shows the cell count at 0 cells (0.0%) have expected count less than 5, which means the assumption is fulfilled. Cramer’s V will be used to determine the strength of the effect and based on Table 16 the value is .133, which has a low association (AcaStat Statistical Software, n.d.).

14 Table 14

16 Fairly bad Count 341 968 943 434 2686 Expected Count 309.8 870.6 991.2 514.3 2686.0 % within Q3a. Country's present economic condition 12.7% 36.0% 35.1% 16.2% 100.0% % within Q42. Extent of democracy 31.2% 31.5% 26.9% 23.9% 28.3% % of Total 3.6% 10.2% 9.9% 4.6% 28.3%

17 Neither good nor bad Count 137 449 682 296 1564 Expected Count 180.4 506.9 577.2 299.5 1564.0 % within Q3a. Country's present economic condition 8.8% 28.7% 43.6% 18.9% 100.0% % within Q42. Extent of democracy 12.5% 14.6% 19.5% 16.3% 16.5% % of Total 1.4% 4.7% 7.2% 3.1% 16.5%

19 Very good Count 23 95 157 157 432 Expected Count 49.8 140.0 159.4 82.7 432.0 % within Q3a. Country's present economic condition 5.3% 22.0% 36.3% 36.3% 100.0% % within Q42. Extent of democracy 2.1% 3.1% 4.5% 8.6% 4.6% % of Total 0.2% 1.0% 1.7% 1.7% 4.6% Total Count 1094 3074 3500 1816 9484 Expected Count 1094.0 3074.0 3500.0 1816.0 9484. 0 % within Q3a. Country's present economic condition 11.5% 32.4% 36.9% 19.1% 100.0 % % within Q42. Extent of democracy 100.0% 100.0% 100.0% 100.0% 100.0 % % of Total 11.5% 32.4% 36.9% 19.1% 100.0 % From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015.

20 Table 15 Chi-Square Tests Value df Asymptotic Significance (2- sided) Pearson Chi-Square 500.943 a 12 .000 Likelihood Ratio 492.594 12 .000 Linear-by-Linear Association 440.844 1 .000 N of Valid Cases 9484 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 49.83. From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Table 16 Symmetric Measures Value Approximate Significance Nominal by Nominal Phi .230 .000 Cramer's V .133 .000 N of Valid Cases 9484 From “Afrobarometer: Dataset B,” IBM SPSS v25, 2015. Social Implications Scenario One The social implications of trust in the police and democracy based on the variable chosen to be used in scenario one. The chi-square test shows an association between trust in the police and democracy, but the strength of the effect size is weak. In the study, countries with democracy have more confidence in the police and judicial system than those in an authoritarian government (Hsieh & Boateng, 2015). Regression analysis was done that showed significance between democracy and trust in the police and judicial system with the ( X 2 = 1068.53; p < .001)