library(readxl)
library(rcompanion)

# Chi-Square Test of Independence
# Research Question: Is there an association between a customer's banking type and whether they have experienced a security concern?
# Variables: Banking type (Online banking, In-person banking), Security concern (Yes, No)
# Test: Chi-Square Test of Independence
# H0: There is no association between banking type and security concern.
# Ha: There is an association between banking type and security concern.


banking_data <- read_excel("C:/Users/DELL/OneDrive - Saint Louis University/AA 5221/Final Project/banking_data.xlsx")
View(banking_data)

observed <- table(banking_data$Banking_Type, banking_data$Security_Concern)
observed

##                    
##                      No Yes
##   In-person banking  95  36
##   Online banking     69 100

barplot(
  observed,
  beside = TRUE,
  col = rainbow(nrow(observed)),
  legend = rownames(observed)
)

chi_result <- chisq.test(observed)
chi_result

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  observed
## X-squared = 28.641, df = 1, p-value = 8.712e-08

cramer_v <- rcompanion::cramerV(observed)
cramer_v

## Cramer V 
##   0.3157

# A Chi-Square Test of Independence was conducted to determine if there was an association between banking type (online banking or in-person banking) and security concern (yes or no).
# The results showed that there was an association between the two variables, χ²(1) = 28.64, p < .001.
# The association was moderate (Cramer's V = .32).




library(ggpubr)

## Loading required package: ggplot2

# Pearson or Spearman Correlation
# Research Question: Is there a relationship between a customer's age and their perceived security?
# Variables: Age (years), Perceived security (1 to 10)
# Test: Spearman Correlation
# H0: There is no relationship between age and perceived security.
# Ha: There is a relationship between age and perceived security.


ggscatter(
  banking_data,
  x = "Age",
  y = "Perceived_Security",
  add = "reg.line",
)

# The relationship is linear.
# The relationship is negative.
# There are no outliers.


mean(banking_data$Age)

## [1] 45.43333

sd(banking_data$Age)

## [1] 14.33751

median(banking_data$Age)

## [1] 45

mean(banking_data$Perceived_Security)

## [1] 6.659

sd(banking_data$Perceived_Security)

## [1] 1.450851

median(banking_data$Perceived_Security)

## [1] 6.8

hist(banking_data$Age,
     breaks = 15,
     col = "skyblue",
     border = "white")

hist(banking_data$Perceived_Security,
     breaks = 15,
     col = "firebrick",
     border = "white")

#Data for Age appears abnormally distributed.
#Data for Perceived_Security appears normally distributed.


shapiro.test(banking_data$Age)

## 
##  Shapiro-Wilk normality test
## 
## data:  banking_data$Age
## W = 0.98716, p-value = 0.009103

shapiro.test(banking_data$Perceived_Security)

## 
##  Shapiro-Wilk normality test
## 
## data:  banking_data$Perceived_Security
## W = 0.99428, p-value = 0.3216

#Age was abnormally distributed, W = 0.987, p = .009.
#Perceived_Security was normally distributed, W = 0.994, p = .322.


cor.test(
  banking_data$Age,
  banking_data$Perceived_Security,
  method = "spearman"
)

## Warning in cor.test.default(banking_data$Age, banking_data$Perceived_Security,
## : cannot compute exact p-value with ties

## 
##  Spearman's rank correlation rho
## 
## data:  banking_data$Age and banking_data$Perceived_Security
## S = 5856707, p-value = 1.01e-07
## alternative hypothesis: true rho is not equal to 0
## sample estimates:
##        rho 
## -0.3015049

# A Spearman correlation was conducted to test the relationship between age (Mdn = 45.00) and perceived security (Mdn = 6.80).

# There was a statistically significant relationship between the two variables, ρ = -.30, p < .001.

# The relationship was negative and weak.

# As age increased, perceived security decreased.



# Independent T-Test or Mann-Whitney U
# Research Question: Is there a difference in perceived security between online banking customers and in-person banking customers?
# Variables: Banking type (Online banking, In-person banking), Perceived security (1 to 10)
# Test: Independent T-Test
# H0: There is no difference in perceived security between online banking and in-person banking customers.
# Ha: There is a difference in perceived security between online banking and in-person banking customers.


library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(effectsize)

## 
## Attaching package: 'effectsize'

## The following object is masked from 'package:rcompanion':
## 
##     phi

library(effsize)


banking_data %>%
  group_by(Banking_Type) %>%
  summarise(
    Mean = mean(Perceived_Security, na.rm = TRUE),
    Median = median(Perceived_Security, na.rm = TRUE),
    SD = sd(Perceived_Security, na.rm = TRUE),
    N = n()
  )

## # A tibble: 2 × 5
##   Banking_Type       Mean Median    SD     N
##   <chr>             <dbl>  <dbl> <dbl> <int>
## 1 In-person banking  7.47    7.6  1.18   131
## 2 Online banking     6.03    6.1  1.32   169

hist(banking_data$Perceived_Security[banking_data$Banking_Type == "In-person banking"],
     breaks = 15,
     col = "skyblue",
     border = "white",
     main = "Perceived Security: In-person Banking Customers",
     xlab = "Perceived Security (1 to 10)")

hist(banking_data$Perceived_Security[banking_data$Banking_Type == "Online banking"],
     breaks = 15,
     col = "firebrick",
     border = "white",
     main = "Perceived Security: Online Banking Customers",
     xlab = "Perceived Security (1 to 10)")

#Data for In-person_banking appears normally distributed.
#Data for Online_banking appears normally distributed.

ggboxplot(banking_data, x = "Banking_Type", y = "Perceived_Security",
          color = "Banking_Type",
          palette = "jco",
          add = "jitter")

# The Online_banking boxplot does not have outliers.
# The In-person_banking boxplot does not have outliers.


shapiro.test(banking_data$Perceived_Security[banking_data$Banking_Type == "In-person banking"])

## 
##  Shapiro-Wilk normality test
## 
## data:  banking_data$Perceived_Security[banking_data$Banking_Type == "In-person banking"]
## W = 0.98985, p-value = 0.4541

shapiro.test(banking_data$Perceived_Security[banking_data$Banking_Type == "Online banking"])

## 
##  Shapiro-Wilk normality test
## 
## data:  banking_data$Perceived_Security[banking_data$Banking_Type == "Online banking"]
## W = 0.9935, p-value = 0.6572

#The In-person_banking group is normally distributed, (p > .05).
#The Online_banking group is normally distributed, (p > .05).


t.test(Perceived_Security ~ Banking_Type, data = banking_data, var.equal = TRUE)

## 
##  Two Sample t-test
## 
## data:  Perceived_Security by Banking_Type
## t = 9.8089, df = 298, p-value < 2.2e-16
## alternative hypothesis: true difference in means between group In-person banking and group Online banking is not equal to 0
## 95 percent confidence interval:
##  1.153301 1.732222
## sample estimates:
## mean in group In-person banking    mean in group Online banking 
##                        7.471756                        6.028994

cohens_d_result <- effectsize::cohens_d(Perceived_Security ~ Banking_Type, data = banking_data, pooled_sd = TRUE)
print(cohens_d_result)

## Cohen's d |       95% CI
## ------------------------
## 1.14      | [0.90, 1.39]
## 
## - Estimated using pooled SD.

# An Independent T-Test was conducted to determine if there was a difference in perceived security between in-person banking and online banking customers.
# In-person banking scores (M = 7.47, SD = 1.18) were significantly different from online banking scores (M = 6.03, SD = 1.32), t(298) = 9.81, p < .001.
# The effect size was large, Cohen's d = 1.14.



# Dependent T-Test or Wilcoxon Signed-Rank
# Research Question: Is there a difference in perceived security before versus after exposure to a cybersecurity awareness message?
# Variables: Security_Before (1 to 10), Security_After (1 to 10)
# Test: Wilcoxon Signed-Rank
# H0: There is no difference in perceived security before versus after the cybersecurity awareness message.
# Ha: There is a difference in perceived security before versus after the cybersecurity awareness message.


library(effsize)
library(rstatix)

## 
## Attaching package: 'rstatix'

## The following objects are masked from 'package:effectsize':
## 
##     cohens_d, eta_squared, omega_squared

## The following object is masked from 'package:stats':
## 
##     filter

banking_data2 <- read_excel("C:/Users/DELL/OneDrive - Saint Louis University/AA 5221/Final Project/banking_data2.xlsx")
View(banking_data2)


Before <- banking_data2$Security_Before
After <- banking_data2$Security_After

Differences <- After - Before


mean(Before, na.rm = TRUE)

## [1] 5.941667

median(Before, na.rm = TRUE)

## [1] 5.9

sd(Before, na.rm = TRUE)

## [1] 0.9881144

mean(After, na.rm = TRUE)

## [1] 6.823333

median(After, na.rm = TRUE)

## [1] 6.8

sd(After, na.rm = TRUE)

## [1] 0.922665

hist(Differences,
     breaks = 15,
     col = "blue",
     border = "white")

#Data for the difference scores appears abnormally distributed.


boxplot(Differences,
        main = "Distribution of Score Differences (After - Before)",
        ylab = "Difference in Scores",
        col = "blue",
        border = "darkblue")

# The difference scores boxplot has two outliers.


shapiro.test(Differences)

## 
##  Shapiro-Wilk normality test
## 
## data:  Differences
## W = 0.90366, p-value = 0.0001793

#Shapiro-Wilk Difference Scores
#The data is abnormally distributed, (p < .001).

wilcox.test(Before, After, paired = TRUE, na.action = na.omit)

## 
##  Wilcoxon signed rank test with continuity correction
## 
## data:  Before and After
## V = 13, p-value = 2.91e-11
## alternative hypothesis: true location shift is not equal to 0

df_long <- data.frame(id = rep(1:length(Before), 2), time = rep(c("Before", "After"), each = length(Before)), score = c(Before, After))

wilcox_effsize(df_long, score ~ time, paired = TRUE)

## # A tibble: 1 × 7
##   .y.   group1 group2 effsize    n1    n2 magnitude
## * <chr> <chr>  <chr>    <dbl> <int> <int> <ord>    
## 1 score After  Before   0.859    60    60 large

#A Wilcoxon Signed-Rank Test was conducted to determine if there was a difference in perceived security before exposure to a cybersecurity awareness message versus after exposure.
#Before scores (Mdn = 5.90) were significantly different from after scores (Mdn = 6.80), V = 13, p < .001.
#The effect size was large, r = .86.