We took the dataset on UA in 2012 to do research about how people evaluate their goverment and actions of politicians. The future work will be a comparison of these attitudes between different years to show changes. Maybe later we even can say about how Maydan evolve - because of people (if their attitudes changed to worse) or it was created artificially by someone (if attitudes of people didn't change). Now we want to describe one particular dataset to show the variables that we need in our work:

We assume that there is significant difference between how males and females evaluate how democratic Ukraine:

H0 - there is no significant difference between genders.

H1 - there is significant difference between genders.

## 
##  Shapiro-Wilk normality test
## 
## data:  as.numeric(DA2$dmcntov)
## W = 0.96228, p-value < 0.00000000000000022

QQ-plot and Shapiro-Wilk normality test show that our variable haven't normal distribution.

## 
##  Bartlett test of homogeneity of variances
## 
## data:  DA2$dmcntov by DA2$gndr
## Bartlett's K-squared = 0.063856, df = 1, p-value = 0.8005

## Levene's Test for Homogeneity of Variance (center = median)
##         Df F value Pr(>F)
## group    1  0.7668 0.3813
##       1958

The p-values of Levene Test=0.3813 and Bartlett Test=0.8005 both above the significance level of 0.05 so we can assume that variances are equal.

## 
##  Two Sample t-test
## 
## data:  DA2$dmcntov by DA2$gndr
## t = -4.2878, df = 1958, p-value = 0.00001893
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.7281001 -0.2710835
## sample estimates:
##   mean in group Male mean in group Female 
##             3.722449             4.222041

With p-value=0.00001995 we reject H0 hypothesis, therefore, different groups define how Ukraine democratic overall.

We assume that females and males voted differently - gender had influence on voting behaviour on the last national elections.

H0 - there is no significant association between gender and voting behaviour.

H1 - there is significant association between gender and voting behaviour.

##       Voted
## Gender Male Female
##    Yes  543    976
##    No   192    249

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  DA2$vote and DA2$gndr
## X-squared = 8.5204, df = 1, p-value = 0.003512

With p-value=0.003512 we can decline H0-hypothesis and say that there is significant association between categories. From that we can assume that gender may have influence on voting behaviour on the last national elections.

We assume that there is significant difference between adherents of different political parties according to their satisfaction with the national government:

H0 - there is no significant difference between adherents of different political parties.

H1 - there is significant difference between adherents of different political parties.

## 
##  Shapiro-Wilk normality test
## 
## data:  as.numeric(politics$stfgov)
## W = 0.91427, p-value < 0.00000000000000022

QQ-plot and Shapiro-Wilk normality test show that our variable haven’t normal distribution.

## Levene's Test for Homogeneity of Variance (center = median)
##        Df F value     Pr(>F)    
## group   5  5.7439 0.00003163 ***
##       887                       
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
##  Bartlett test of homogeneity of variances
## 
## data:  politics$stfgov by politics$prtcldua
## Bartlett's K-squared = 26.356, df = 5, p-value = 0.00007611

The p-values of Levene Test=0.00003163 and Bartlett Test=0.00007611 above the significance level of 0.05 so we can assume that variances are unequal. From that we set "var.equal = FALSE"

oneway.test(politics$stfgov~politics$prtcldua, var.equal = FALSE)

## 
##  One-way analysis of means (not assuming equal variances)
## 
## data:  politics$stfgov and politics$prtcldua
## F = 61.338, num df = 5.00, denom df = 165.21, p-value <
## 0.00000000000000022

The p-value of OneWay.ANOVA=0.00000000000000022 above the significance level of 0.05 so we can assept H1-hypothesis and can assume that there is difference between adherents of different political parties according to their satisfaction with the national government. But because of not normal distribution will be better to use non-parametric Kruskal Test

kruskal.test(politics$stfgov~politics$prtcldua)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  politics$stfgov by politics$prtcldua
## Kruskal-Wallis chi-squared = 240.63, df = 5, p-value <
## 0.00000000000000022

The p-value of Kruskal-Wallis=0.00000000000000022 so we have the same p-value with ANOVA and the same conclusion.

Our data was someway obvious cause Party of Regions was governing party at 2012 so this party have significant dofference with other parties in satisfaction with national goverment, but we proved it statistically.