Likert Survey Analysis

Loading the libraries

library(dplyr)
library(janitor)
library(tidyr)
library(ggplot2)
library(readxl)
library(stringr)

Part 1

Loading the data that is in the data folder.

students <- read_excel("data/likert_survey.xlsx", sheet = "students") %>%
  clean_names() 

professors <- read_excel("data/likert_survey.xlsx", sheet = "professors")%>%
  clean_names()

Part 2

We are now looking into the total number of answers obtained by students, professors and if each question obtained the same number of answers.

To answer the first two questions I am using the tidyverse way, putting together the dataset, to then create long format data. This will allow me to exploit the power of grouping and summarising.

# let's first put the two datasets together

students <- students %>% 
  mutate(respondent = "student") # create a respondent column

professors <- professors %>% 
  mutate(respondent = "professor") # create a respondent column

# We can now bind together the two tibbles

total <- bind_rows(students, professors)

# use head(total) to inspect the element

# And now we can use pivot_longer to get it in a tidy format

total <- total %>% 
  pivot_longer(names_to = "response_type",
               values_to = "value",
               cols= c(-question_id, -respondent))

#head(total)

After having wrangled the data to have it in tidy format, I can respond to the first two questions.

total %>% 
  group_by(respondent) %>% 
  summarise(total_answer = sum(value)) %>% 
  ungroup() %>% 
  rename('Total Answer' = total_answer,
         Respondent = respondent) %>% 
  knitr::kable()

Respondent	Total Answer
professor	3654
student	7748

We are now checking if each question have the same number of answers.

total %>% 
  group_by(respondent, question_id) %>%  # we group_by respondent and question_id
  summarise(total_answer = sum(value)) %>% # we calculate the total answer per question and respondent
  ungroup() %>% # we ungroup!
  distinct(respondent, total_answer) %>%  # we keep only the unique values
  knitr::kable()

## `summarise()` has grouped output by 'respondent'. You can override using the
## `.groups` argument.

respondent	total_answer
professor	115
professor	113
professor	112
professor	111
professor	103
student	244
student	233
student	234
student	239
student	241
student	238
student	230

The answer is No as we can see that not all questions have been answered the same amount of time, for both professors and students. It means that some respondents left some questions unanswered.

Part 3

I am now comparing the distribution of answers of the students versus the professors.

total %>%
  ggplot(mapping = aes(x =response_type,
                       y = value,
                       fill = respondent)) +
  geom_col(position = "dodge") + 
  facet_wrap(vars(question_id)) + 
  labs(title = "Comparaison of answers for students vs professors",
       subtitle = "Each facet shows a different question",
       x="Response type", 
       y="Number of Responses",
       fill = "respondent") +
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 90)) # tilting the text

From this chart, we can see that basically for each question, students gave more answers.

Part 4

Now I am computing the mode for the students and professors for each of the questions.

In order to make the data nice to present in a table I select the response_type, respondent and question_id and use pivot_wider() so to have one line per question_id.

total %>%
  group_by(question_id, respondent) %>%
  slice_max(order_by = value, n = 1) %>% # we keep the top 1 value for each group
  ungroup() %>% 
  select(-value) %>% 
  pivot_wider(names_from = respondent, values_from = response_type) %>% 
  knitr::kable()

question_id	professor	student
0	totally_agree	totally_agree
1	agree	agree
2	agree	agree
3	agree	agree
4	agree	agree
5	agree	agree
6	agree	agree
7	disagree	disagree
8	disagree	disagree
9	disagree	disagree
10	agree	agree
11	agree	agree
12	agree	agree
13	agree	agree
14	agree	agree
15	totally_agree	totally_disagree
16	totally_agree	totally_disagree
17	agree	totally_disagree
18	agree	totally_disagree
19	agree	totally_disagree
20	totally_agree	totally_agree
21	agree	totally_agree
22	totally_agree	totally_agree
23	disagree	totally_agree
24	totally_agree	agree
25	agree	agree
26	agree	agree
27	agree	agree
28	totally_agree	agree
29	agree	agree
30	agree	agree
31	agree	disagree