Visualizations from Norm Deviation Project
Sara Lieber
07/26/2021
Project Background
In this project, participants were asked to describe a time when they witnessed someone deviate from a widely accepted norm. After reflecting on the instance of norm deviation, participants were asked follow-up questions regarding how their own attitudes toward the norm and perception of the norm in society more generally changed as a result of witnessing the norm deviation.
Import data
data <- import(here("Data", "norm_deviation_GS_SP19.xlsx"))
glimpse(data)## Rows: 200
## Columns: 27
## $ Duration__in_seconds_ <dbl> 664, 843, 1244, 1534, 3382, 776, 1132, 1467, 986~
## $ RecordedDate <dttm> 2019-04-01 12:42:36, 2019-04-01 14:50:36, 2019-~
## $ Q1.1 <chr> "20", "19", "18", "30", "19", "19", "18", "18", ~
## $ Q328 <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ~
## $ Q328_4_TEXT <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, ~
## $ Q1.3 <dbl> 10, 16, 16, 16, 14, 16, 10, 8, 17, 16, 16, 16, 1~
## $ Q1.3_17_TEXT <chr> NA, NA, NA, NA, NA, NA, NA, NA, "mixed race", NA~
## $ Q346 <chr> NA, "one of my friends in high school made the d~
## $ Q347.0 <chr> "religion norms", "gender norms- female cleanlin~
## $ Q348 <chr> "traditional house hold, people around believe i~
## $ Q349.0 <dbl> 5, 5, 1, 4, 6, 7, 5, 4, 6, 5, 5, 6, 4, 7, 5, 1, ~
## $ Q350 <chr> "School", "school environment", "This was within~
## $ Q351.0 <chr> "25", "when I witnessed the act and explanation,~
## $ Q352 <chr> "95 percent", "women", "It applies to anyone", "~
## $ Q353.0 <dbl> 4, 4, 4, 4, 6, 6, 6, 4, 4, 4, 6, 4, 4, 4, 4, 6, ~
## $ Q355.0 <dbl> 6, 7, 7, 7, 6, 5, 5, 7, 7, 5, 6, 7, 6, 7, 7, 1, ~
## $ Q356 <dbl> 7, 5, 7, 6, 5, 5, 4, 7, 5, 5, 6, 6, 5, 6, 4, 1, ~
## $ Q357 <dbl> 95, 100, 95, 97, 95, 80, 80, 97, 90, 95, 90, 95,~
## $ Q358 <dbl> 80, 98, 65, 95, 60, 50, 30, 97, 70, 0, 90, 90, 8~
## $ Q359 <dbl> 4, 7, 4, 7, 7, 8, 7, 1, 8, 5, 4, 7, 6, 7, 8, 1, ~
## $ Q360 <dbl> 4, 7, 1, 7, 6, 9, 7, 1, 8, 5, 4, 7, 6, 9, 9, 1, ~
## $ Q361 <dbl> 2, 1, 2, 1, 1, 2, 2, 2, 2, 1, 1, 2, 1, 1, 2, 2, ~
## $ Q363 <chr> NA, "made me think about why women are held to t~
## $ Q364 <dbl> 6, 1, 7, 6, 6, 5, 5, 7, 7, 5, 6, 7, 6, 5, 6, 1, ~
## $ Q365 <dbl> 5, 5, 7, 5, 7, 5, 5, 5, 5, 5, 6, 6, 5, 1, 5, 1, ~
## $ Q366 <dbl> 95, 99, 95, 97, 100, 60, 90, 100, 90, 95, 90, 90~
## $ Q367 <dbl> 80, 99, 65, 95, 100, 100, 30, 100, 70, 0, 90, 90~Cleaning Data
Re-naming columns
data_clean <- data %>%
rename(Age = Q1.1,
Gender = Q328,
Gender_Text = Q328_4_TEXT,
Ethnicity = Q1.3,
Ethnicity_Text = Q1.3_17_TEXT,
Deviation_Text = Q346,
Norm_Deviated = Q347.0,
Identity_Norm_Deviant = Q348,
Trust_Deviant_Worldview = Q349.0,
Location_Deviation = Q350,
Num_Ppl_Present = Q351.0,
Group_Norm_Applies = Q352,
Whether_Observer_in_Group = Q353.0,
Prior_Rate_Followed_Norm = Q355.0,
Prior_Observer_Norm_Endorsement = Q356,
Prior_Percent_Follow = Q357,
Prior_Percent_Believed = Q358,
Shared_Views_1 = Q359,
Shared_Views_2 = Q360,
Whether_Opinion_Changed = Q361,
Whether_Opinion_Changed_Text = Q363,
Post_Rate_Follow_Norm = Q364,
Post_Observer_Norm_Endorsement = Q365,
Post_Percent_Follow = Q366,
Post_Percent_Believe = Q367)Descriptions of Variables
Deviation_Text: Describe a time when you witnessed someone deviate from a norm by doing something that challenged conventions or by expressing an opinion that went against widely accepted ideas.
Norm_Deviated: What norm did this person deviate from?
Identity_Norm_Deviant: Describe the identity of the person who you witnessed deviate from a norm
Trust_Deviant_Worldview: How much do you trust this person’s way of seeing the world?
Location_Deviation: Where did the instance you described above take place?
Num_Ppl_Present: How many people were present?
Group_Norm_Applies: What group of people does the norm you described above apply to?
Whether_Observer_in_Group: Are you a part of this group?
Prior_Rate_Followed_Norm: Prior to witnessing someone deviate from this norm, how often did you follow this norm when you were in a situation where the norm applied?
Prior_Observer_Norm_Endorsement: Prior to witnessing someone deviate from this norm, how strongly did you believe that people should follow this norm?
Prior_Percent_Follow: Prior to witnessing someone deviate from this norm, what percentage of people in the group that this norm applies to did you think followed this norm? Fill in a percentage between 0 to 100
Prior_Percent_Believed: Prior to witnessing someone deviate from this norm, what percentage of people in your society did you think believed that people should follow this norm? Fill in a percentage between 0 to 100
Shared_Views_1: Please rate your agreement with the following statements about you and the person you described previously at the time that you witnessed them deviating from a norm: “I felt like we saw the world in the same way.”
Shared_Views_2: Please rate your agreement with the following statements about you and the person you described previously at the time that you witnessed them deviating from a norm: “I felt like we shared the same thoughts and feelings about things.”
Whether_Opinion_Changed: Did witnessing someone deviate from the norm you described above change your opinion about the original norm?
Whether_Opinion_Changed_Text: If yes, please explain how your opinion about the original norm changed.
Post_Rate_Follow_Norm: How often do you follow this norm when you are in a situation where the norm applies?
Post_Observer_Norm_Endorsement: How strongly do you believe that people should follow this norm?
Post_Percent_Follow: What percentage of people in the group that this norm applies to do you think follow this norm? Fill in a percentage between 0 to 100
Post_Percent_Believe: What percentage of people in your society do you think believe that people should follow this norm? Fill in a percentage between 0 to 100
Cleaning variable types
# str(data_clean)
data_clean <- data_clean %>%
mutate(Age = as.numeric(Age),
Gender = as.factor(Gender),
Ethnicity = as.factor(Ethnicity),
Whether_Observer_in_Group = as.factor(Whether_Observer_in_Group),
Whether_Opinion_Changed = as.factor(Whether_Opinion_Changed))
nrow(data_clean) # 200## [1] 200Recoding numerical variables
data_clean$Trust_Deviant_Worldview <- recode(data_clean$Trust_Deviant_Worldview, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5)
data_clean$Prior_Rate_Followed_Norm <- recode(data_clean$Prior_Rate_Followed_Norm, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5)
data_clean$Prior_Observer_Norm_Endorsement <- recode(data_clean$Prior_Observer_Norm_Endorsement, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5)
data_clean$Shared_Views_1 <- recode(data_clean$Shared_Views_1, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5, `8` = 6, `9` = 7)
data_clean$Shared_Views_2 <- recode(data_clean$Shared_Views_2, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5, `8` = 6, `9` = 7)
data_clean$Post_Rate_Follow_Norm <- recode(data_clean$Post_Rate_Follow_Norm, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5)
data_clean$Post_Observer_Norm_Endorsement <- recode(data_clean$Post_Observer_Norm_Endorsement, `1` = 1, `4` = 2, `5` = 3, `6` = 4, `7` = 5)
levels(data_clean$Whether_Opinion_Changed) <- c(0, 1)
levels(data_clean$Whether_Opinion_Changed)## [1] "0" "1"Labeling levels of nominal variables
levels(data_clean$Gender) <- c("Female", "Male", "Non-binary")
levels(data_clean$Ethnicity) <- c("American Indian or Alaska Native", "Asian", "Black or African American", "Hispanic, Latinx or Spanish Origin", "Middle Eastern or North African", "Native Hawaiian or Other Pacific Islander", "White", "Some other ethnicity or origin", "I prefer not to answer")
levels(data_clean$Whether_Observer_in_Group) <- c("Yes", "No")
levels(data_clean$Whether_Opinion_Changed) <- c("Yes", "No")Choose specific columns
data_short <- data_clean %>%
select(Age:Post_Percent_Believe)
data_short <- data_short %>%
filter(Whether_Opinion_Changed != "NA") %>%
droplevels()Bar Plots
Whether opinion changed after witnessing norm deviation
text_settings <- theme(text = element_text(size = 16)) +
theme(plot.title = element_text(size = 16)) +
theme(axis.text.x = element_text(size = 16)) +
theme(axis.text.y = element_text(size = 16)) +
theme(axis.ticks = element_blank())
ggplot(data_short, aes(x = Whether_Opinion_Changed)) +
geom_bar(na.rm = TRUE, fill = "coral") +
scale_x_discrete(na.translate = FALSE) +
labs(title = "Whether opinion changed after witnessing the norm deviation", x = "Whether Opinion Changed", y = "Frequency") +
theme_light() +
theme(plot.title = element_text(hjust=0.5)) +
text_settings
Investigating whether rates of opinion change differ between men and women
# Outcome variable by gender
data_short <- data_short %>%
filter(Gender != "NA",
Gender != "Non-binary") %>%
droplevels()Stacked bar graph
library(RColorBrewer)
ggplot(data_short, aes(x = Whether_Opinion_Changed, fill = Gender)) +
geom_bar(na.rm = TRUE) +
scale_x_discrete(na.translate = FALSE) +
labs(title = "Whether opinion changed after witnessing the norm deviation", x = "Whether Opinion Changed", y = "Frequency") +
scale_fill_brewer(palette = "Accent") +
theme_light() +
theme(plot.title = element_text(hjust=0.5)) +
text_settings
Side-by-side bar graph
ggplot(data_short, aes(x = Whether_Opinion_Changed, fill = Gender)) +
geom_bar(position = "dodge", na.rm = TRUE) +
scale_x_discrete(na.translate = FALSE) +
labs(title = "Whether opinion changed after witnessing the norm deviation", x = "Whether Opinion Changed", y = "Frequency") +
scale_fill_brewer(palette = "Accent") +
theme_light() +
theme(plot.title = element_text(hjust=0.5)) +
text_settings
String Manipulation
Text analysis of the Deviation_Text variable
- Deviation_Text: Describe a time when you witnessed someone deviate from a norm by doing something that challenged conventions or by expressing an opinion that went against widely accepted ideas.
Clean the string variables (i.e., remove punctuation, numbers, common words, whitespide, & convert to lowercase)
docs <- VCorpus(VectorSource(data_short$Deviation_Text)) %>%
tm_map(removePunctuation) %>% # removes punctuation
tm_map(removeNumbers) %>% # removes numbers
tm_map(tolower) %>% # converts to lowercase
tm_map(removeWords, stopwords("english")) %>% # removes common words (i.e., 'a', 'the')
tm_map(stripWhitespace) %>% # removes whitespace
tm_map(PlainTextDocument)
tdm <- TermDocumentMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
colnames(tdm) <- 1:nrow(data_short)
dtm <- DocumentTermMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
# head(dtm)Calculate frequency of each word
freq <- colSums(dtm)Gather columns representing unique words into a single variable column
dtm_gather <- gather(as.data.frame(dtm), key = unique_words, frequency, able:zumba)
# Create a new variable containing only the unique words
unique_words <- unique(dtm_gather$unique_words)Combine unique words and frequencies into a single dataframe
word_set <- cbind.data.frame(unique_words, freq)Arrange word set in descending order from most to least frequently used words
word_set_ordered <- word_set[order(-freq),]Wordcloud
ggwordcloud(word_set$unique_words, word_set$freq, colors = c("#1B9E77", "#D95F02", "#7570B3", "#E7298A", "#66A61E")) + labs(title = "Describe a time when you witnessed someone deviate from a norm by doing something that challenged \nconventions or by expressing an opinion that went against widely accepted ideas.")
Text analysis of the Norm_Deviated variable
- Norm_Deviated: What norm did this person deviate from?
docs <- VCorpus(VectorSource(data_short$Norm_Deviated)) %>%
tm_map(removePunctuation) %>% # removes punctuation
tm_map(removeNumbers) %>% # removes numbers
tm_map(tolower) %>% # converts to lowercase
tm_map(removeWords, c(stopwords("english"), "norm", "norms", "deviated")) %>% # removes common words (i.e., 'a', 'the'); I also removed words that re-stated the prompt (i.e., norm, norms, deviated)
tm_map(stripWhitespace) %>% # removes whitespace
tm_map(PlainTextDocument)
tdm <- TermDocumentMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
colnames(tdm) <- 1:nrow(data_short)
dtm <- DocumentTermMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
# head(dtm)Calculate frequency of each word
freq <- colSums(dtm)Gather columns representing unique words into a single variable column
dtm_gather <- gather(as.data.frame(dtm), key = unique_words, frequency, able:yelling)
# Create a new variable containing only the unique words
unique_words <- unique(dtm_gather$unique_words)Combine unique words and frequencies into a single dataframe
word_set <- cbind.data.frame(unique_words, freq)Arrange word set in descending order from most to least frequently used words
word_set_ordered <- word_set[order(-freq),]Wordcloud
ggwordcloud(word_set$unique_words, word_set$freq, colors = c("#1B9E77", "#D95F02", "#7570B3", "#E7298A", "#66A61E")) + labs(title = "What norm did this person deviate from?")
Text analysis of the Identity_Norm_Deviant variable
- Identity_Norm_Deviant: Describe the identity of the person who you witnessed deviate from a norm
docs <- VCorpus(VectorSource(data_short$Identity_Norm_Deviant)) %>%
tm_map(removePunctuation) %>% # removes punctuation
tm_map(removeNumbers) %>% # removes numbers
tm_map(tolower) %>% # converts to lowercase
tm_map(removeWords, stopwords("english")) %>% # removes common words (i.e., 'a', 'the'); I also removed words that re-stated the prompt (i.e., norm, norms, deviated)
tm_map(stripWhitespace) %>% # removes whitespace
tm_map(PlainTextDocument)
tdm <- TermDocumentMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
colnames(tdm) <- 1:nrow(data_short)
dtm <- DocumentTermMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
# head(dtm)Calculate frequency of each word
freq <- colSums(dtm)Gather columns representing unique words into a single variable column
dtm_gather <- gather(as.data.frame(dtm), key = unique_words, frequency, acquaintance :younger)
# Create a new variable containing only the unique words
unique_words <- unique(dtm_gather$unique_words)Combine unique words and frequencies into a single dataframe
word_set <- cbind.data.frame(unique_words, freq)Arrange word set in descending order from most to least frequently used words
word_set_ordered <- word_set[order(-freq),]Wordcloud
ggwordcloud(word_set$unique_words, word_set$freq, colors = c("#1B9E77", "#D95F02", "#7570B3", "#E7298A", "#66A61E")) + labs(title = "Describe the identity of the person who you witnessed deviate from a norm.")
Text analysis of the Whether_Opinion_Changed_Text variable
- Whether_Opinion_Changed_Text: If yes, please explain how your opinion about the original norm changed.
docs <- VCorpus(VectorSource(data_short$Whether_Opinion_Changed_Text)) %>%
tm_map(removePunctuation) %>% # removes punctuation
tm_map(removeNumbers) %>% # removes numbers
tm_map(tolower) %>% # converts to lowercase
tm_map(removeWords, stopwords("english")) %>% # removes common words (i.e., 'a', 'the'); I also removed words that re-stated the prompt (i.e., norm, norms, deviated)
tm_map(stripWhitespace) %>% # removes whitespace
tm_map(PlainTextDocument)
tdm <- TermDocumentMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
colnames(tdm) <- 1:nrow(data_short)
dtm <- DocumentTermMatrix(docs) %>% # This creates a document in which each unique word is a row, and each document (which in this case is each participant's response) is a column
as.matrix()
# head(dtm)Calculate frequency of each word
freq <- colSums(dtm)Gather columns representing unique words into a single variable column
dtm_gather <- gather(as.data.frame(dtm), key = unique_words, frequency, everything())
# Create a new variable containing only the unique words
unique_words <- unique(dtm_gather$unique_words)Combine unique words and frequencies into a single dataframe
word_set <- cbind.data.frame(unique_words, freq)Arrange word set in descending order from most to least frequently used words
word_set_ordered <- word_set[order(-freq),]Wordcloud
ggwordcloud(word_set$unique_words, word_set$freq, colors = c("#1B9E77", "#D95F02", "#7570B3", "#E7298A", "#66A61E")) + labs(title = "Explain how your opinion about the original norm changed.")