library(tidyverse)
library(openintro)
library(tinytex)
library(dplyr)
library(gridExtra)    
dfMasculinity <- read.csv("https://raw.githubusercontent.com/fivethirtyeight/data/master/masculinity-survey/raw-responses.csv", header= TRUE)

Introduction

My dataset is from the article, “What Do Men Think It Means To Be A Man?” The article examines whether the #MeToo movement has changed men’s thinking about masculinity. The article concludes that little has changed in how men regard their masculine identity or in how they think about their behavior in work or in relationships.

https://fivethirtyeight.com/features/what-do-men-think-it-means-to-be-a-man/ https://github.com/ericonsi/607_Assignment1/ https://rpubs.com/ericonsi/719305

dfMasculinity_Subset <- subset(dfMasculinity, select = c("q0001", "age3", "q0026", "q0005", "race2", "q0029", "q0035", "q0030"))
dfMasculinity_Subset <- rename(dfMasculinity_Subset, c(Feel_Masculine =  "q0001", Age = "age3", Orientation = "q0026", BelieveSocietyPressuresMen="q0005", Race="race2", Education="q0029", Region="q0035", State="q0030"))

Analysis

The study did not disaggregate the data. I was curious to see if there were any associations between self perception of masculinity and age, region and/or levels of education.

I looked first at the overall breakdown of how respondents answered this question. The majority of respondents reported feeling “Somewhat Masculine” or “Very Masculine.”

ggplot(data = dfMasculinity_Subset, aes(x = Feel_Masculine)) +
  geom_bar() + ggtitle("Respondent answers to 'How Masculine Do You Feel'")

Then I looked at how the percentages broke down for age, region and education. Because I was going to have to reproduce this exercise many times, I wrote a function to make it easier:

EH_PlotPercents<- function(filterVariable, filterCriteria, groupBy)
{
dfPercent <- dfMasculinity_Subset %>%
  filter(get(filterVariable) == filterCriteria) %>%
  group_by_at(groupBy) %>%
  summarise(count = n() ) %>%
  mutate( perc = count / sum(count) )

g <- ggplot(data = dfPercent, aes_string(x = groupBy, y= "perc")) +
  geom_col()+ ggtitle(filterCriteria) + theme(axis.text.x = element_text(angle = 90))
return(g)
}
g1 <- EH_PlotPercents("Age", "18 - 34", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
g2 <- EH_PlotPercents("Age", "35 - 64", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
g3 <- EH_PlotPercents("Age", "65 and up", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
grid.arrange(g1, g2, g3, ncol = 3)

For age, I compared all the categories in the dataset (18-34, 35-64 and >65). Not surprisingly, a higher percentage of millenial men are less masculine-identified than older men.

g4 <- EH_PlotPercents("Region", "New England", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
g5 <- EH_PlotPercents("Region", "West South Central", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
grid.arrange(g4, g5, ncol = 2)

When I compared the New England region to the West South Central region (e.g. Texas, Oklahoma, etc.), I found that a higher percentage of men in the West South Central identify as “Very Masculine” compared to those in New England (where I’m from.)

g7 <- EH_PlotPercents("Education", "Did not complete high school", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
g8 <- EH_PlotPercents("Education", "High school or G.E.D.", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
g9 <- EH_PlotPercents("Education", "Post graduate degree", "Feel_Masculine")
## `summarise()` ungrouping output (override with `.groups` argument)
grid.arrange(g7, g8, g9, ncol = 3)

The biggest surprise was to see how much less masculine-identified were those men who did not finish high school compared to those with more education. In fact, the more education a man had, the more likely he was to identify as masculine. If education is a proxy for type of employment, this contradicts the notion portayed in Hollywood and beyond that norms of masculinity play a more important role for men in manual labor jobs than for those in white collar jobs.

Conclusion

There are probably a number of ways to disaggregate this data that would lend more insight into how masculinity varies across a number of independent variables. This would be worth exploring before making general statements about masculinity and the effect of #MeToo.