An analysis of movies released between 1970-2013 that pass the Bechdel test, and looking at their association with movie budget, profit, and other factors.
The Bechdel test is an indicator of the active presence of women in the entire field of film and other fiction, and calls attention to gender inequality in fiction. Some studies indicate that films that pass the test perform better financially while receiving a lower budget, compared to those that do not. Thus, through this project, we intend:
Female representation is growing by day in every sector, with particular attention given to it in film. We, a group of movie nerdgirls, wanted to look for progressive movies that signify higher representation of women or where women are completely centered in the story. The Bechdel Test has been a simple and popular measure of feminism in works of fiction. The rules of the test were first defined back in 1985 in a comic strip featuring two queer women who could not find a movie that
We came across articles talking about the importance of these tests in the industry, and some previous studies that used this test to determine if the movie had significant representation for female character(s). Using this information, we further wanted to explore the prioritization of these movies in the industry and their success through an examination of their budget, profits, and ratings.
Initially, our motivation for this project was prompted by the following questions: * How many movies from 1970-2013 pass the Bechdel test and has the number changed over time? * What is the association between the movies that pass the Bechdel test and the movie’s budget? * What is the association between the movies that pass the Bechdel test and the movie’s profit?
As we delved into the data, we further wanted to explore: * What is the association between movies that pass the Bechdel test and their review ratings?
FiveThirtyEight: The FiveThirtyEight Bechdel
dataset contains data about 1794 films released from 1970 to 2013.
We used this data to examine the relationship between the prominence of
women in a film and that film’s budget and profits. The dataset consists
of variables including title, year,
budget, domestic and international gross revenue, budget
and revenues adjusted for 2013 inflation, and the Bechdel test
(evaluated both as binary pass/fail and categorically evaluating the
three criteria).
Bechdel Test: This database consists of a list
of 9802 movies released between 1874 and 2022 (as of Dec 8th, 2022), and
results for whether they pass or fail the Bechdel test. Variables
include title of the film, date when the film
was added, reviews and comments, and
IMDB ratings. The data can be accessed through
API.
The-Numbers: This is a film industry data
website that tracks box office revenue in a systematic and algorithmic
way. We accounted for the budget and revenue
variables for the movies using this website. The data can be accessed
through API.
IMDb: The Internet Movie Database
(IMDb) is an online database containing information and statistics about
movies, TV shows, and video games as well as actors, directors, and
other film industry professionals. We used the information about the
year of release, runtime, genre,
language and rating of the films.
Tidy Tuesday Project: This is the main source for our dataset which provides weekly releases of raw datasets for users to wrangle and analyse. The data itself originates from FiveThirtyEight that contains movies ranging from 1970 to 2013, and merges data from the other sources: IMDb, The-Numbers.com, and BechdelTest.com.
We imported the following datasets from the Tidy Tuesday project
using read_csv from ‘readr’ library. This dataset was
divided into 2 files: movies.csv and
raw_bechdel.csv, initially.
raw_bechdel = readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-03-09/raw_bechdel.csv')
movies = readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-03-09/movies.csv')We mainly used movies.csv for our analysis. movies.csv included movies from 1970 to 2013; and contains the following variables year, imdb, title, test, clean_test, binary, budget, domgross, intgross, code, budget_2013, domgross_2013, intgross_2013, period_code, decade_code, imdb_id, plot, rated, response, language, country, writer, metascore, imdb_rating, director, released, actors, genre, awards, runtime, type, poster, imdb_votes, error
For data cleaning,
We checked for the variables and converted the following
variables from character to numeric form as
required:domgross, intgross,
domgross_2013, intgross_2013 using as.numeric
function in the tidyverse library.
The ‘binary’ variable in character form was converted to logical form to indicate whether the movie had passed or failed the Bechdel test.
Numeric ‘Decade_code’ variable was re-coded and re-leveled into a
factor decade variable according to the year of release. 5
levels were created for this ‘decade’ variable ranging from 1970-79,
1980-89, 1990-99, 2000-09, 2010 to 2013.
Re-coded and re-leveled the clean_test variable into
4 levels corresponding to the Bechdel Test criteria:
Created an award_winner variable that is
TRUE if the movie has won at least 1 Oscar, Golden Globe,
or BAFTA award
Split the genre character variable into dummy-coded
variables indicating whether the movie falls under up to three of 20
different genres
Created a logical variable pass_bechdel that
indicates whether the movie passes or fails the Bechdel Test based on
the bechdel_score variable
Created profit and ROI variables for
movies. We used worldwide box office revenue instead of domestic
revenue.
profit = intgross_2013 – budget_2013ROI = profit / budget_2013Renamed key variables to relevant names, and removed extra variables
movies_df = movies %>%
mutate(test = as.factor(test),
clean_test = fct_recode(clean_test,
"Less than 2 women" = "nowomen",
"Don't talk to each other" = "notalk",
"Only talk about men" = "men",
"Dubious" = "dubious",
"Passes Bechdel" = "ok"),
clean_test = fct_relevel(clean_test, c("Less than 2 women", "Don't talk to each other",
"Only talk about men", "Dubious", "Passes Bechdel")),
binary = ifelse(clean_test == "Dubious" | clean_test == "Passes Bechdel", TRUE, FALSE),
domgross = as.numeric(domgross),
intgross = as.numeric(intgross),
domgross_2013 = as.numeric(domgross_2013),
intgross_2013 = as.numeric(intgross_2013),
decade_code = case_when(year >= 1970 & year < 1980 ~ "1970-1979",
year >= 1980 & year < 1990 ~ "1980-1989",
year >= 1990 & year < 2000 ~ "1990-1999",
year >= 2000 & year < 2010 ~ "2000-2009",
year >= 2010 & year < 2020 ~ "2010-2013"),
decade_code = as.factor(decade_code),
title = str_replace(title, "'", "'"),
title = str_replace(title, "&", "&"),
title = str_replace(title, "à", "à"),
title = str_replace(title, "å", "å"),
title = str_replace(title, "ä", "ä"),
runtime = as.numeric(str_replace(runtime, " min", "")),
award_winner = ifelse(str_detect(awards, "Won") & (str_detect(awards, "Golden Globe") |
str_detect(awards, "Oscar") | str_detect(awards, "BAFTA")), T, F),
profit = intgross_2013 - budget_2013,
ROI = profit/budget_2013) %>%
separate(genre, into = c("g1", "g2", "g3"), sep = ", ") %>%
rename("pass_bechdel" = binary,
"bechdel_score" = clean_test,
"decade" = decade_code) %>%
mutate(
action = ifelse(g1 == "Action" | g2 == "Action" | g3 == "Action", TRUE, FALSE),
adventure = ifelse(g1 == "Adventure" | g2 == "Adventure" | g3 == "Adventure", TRUE, FALSE),
animation = ifelse(g1 == "Animation" | g2 == "Animation" | g3 == "Animation", TRUE, FALSE),
biography = ifelse(g1 == "Biography" | g2 == "Biography" | g3 == "Biography", TRUE, FALSE),
comedy = ifelse(g1 == "Comedy" | g2 == "Comedy" | g3 == "Comedy", TRUE, FALSE),
crime = ifelse(g1 == "Crime" | g2 == "Crime" | g3 == "Crime", TRUE, FALSE),
documentary = ifelse(g1 == "Documentary" | g2 == "Documentary" | g3 == "Documentary", TRUE, FALSE),
drama = ifelse(g1 == "Drama" | g2 == "Drama" | g3 == "Drama", TRUE, FALSE),
family = ifelse(g1 == "Family" | g2 == "Family" | g3 == "Family", TRUE, FALSE),
fantasy = ifelse(g1 == "Fantasy" | g2 == "Fantasy" | g3 == "Fantasy", TRUE, FALSE),
history = ifelse(g1 == "History" | g2 == "History" | g3 == "History", TRUE, FALSE),
horror = ifelse(g1 == "Horror" | g2 == "Horror" | g3 == "Horror", TRUE, FALSE),
music = ifelse(g1 == "Music" | g2 == "Music" | g3 == "Music", TRUE, FALSE),
musical = ifelse(g1 == "Musical" | g2 == "Musical" | g3 == "Musical", TRUE, FALSE),
mystery = ifelse(g1 == "Mystery" | g2 == "Mystery" | g3 == "Mystery", TRUE, FALSE),
romance = ifelse(g1 == "Romance" | g2 == "Romance" | g3 == "Romance", TRUE, FALSE),
sci_fi = ifelse(g1 == "Sci-Fi" | g2 == "Sci-Fi" | g3 == "Sci-Fi", TRUE, FALSE),
sport = ifelse(g1 == "Sport" | g2 == "Sport" | g3 == "Sport", TRUE, FALSE),
thriller = ifelse(g1 == "Thriller" | g2 == "Thriller" | g3 == "Thriller", TRUE, FALSE),
war = ifelse(g1 == "War" | g2 == "War" | g3 == "War", TRUE, FALSE),
western = ifelse(g1 == "Western" | g2 == "Western" | g3 == "Western", TRUE, FALSE)
) %>%
mutate(across(action:western, ~replace_na(., FALSE))) %>%
select(year, title, bechdel_score, pass_bechdel, budget_2013:intgross_2013, decade, imdb_id, metascore, imdb_rating, award_winner, runtime, profit, ROI, action:western)
skimr::skim(movies_df) %>%
knitr::kable(digits = 3) %>%
kableExtra::kable_styling(bootstrap_options = c("striped", "hover"), font_size = 12) %>%
kableExtra::scroll_box(width = "100%", height = "300px")| skim_type | skim_variable | n_missing | complete_rate | character.min | character.max | character.empty | character.n_unique | character.whitespace | factor.ordered | factor.n_unique | factor.top_counts | logical.mean | logical.count | numeric.mean | numeric.sd | numeric.p0 | numeric.p25 | numeric.p50 | numeric.p75 | numeric.p100 | numeric.hist |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| character | title | 0 | 1.000 | 1 | 83 | 0 | 1768 | 0 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA |
| character | imdb_id | 0 | 1.000 | 6 | 8 | 0 | 1794 | 0 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA |
| factor | bechdel_score | 0 | 1.000 | NA | NA | NA | NA | NA | FALSE | 5 | Pas: 803, Don: 514, Onl: 194, Dub: 142 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA |
| factor | decade | 0 | 1.000 | NA | NA | NA | NA | NA | FALSE | 5 | 200: 840, 201: 438, 199: 337, 198: 125 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | pass_bechdel | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.527 | TRU: 945, FAL: 849 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | award_winner | 202 | 0.887 | NA | NA | NA | NA | NA | NA | NA | NA | 0.119 | FAL: 1402, TRU: 190 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | action | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.247 | FAL: 1351, TRU: 443 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | adventure | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.200 | FAL: 1435, TRU: 359 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | animation | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.060 | FAL: 1686, TRU: 108 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | biography | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.043 | FAL: 1717, TRU: 77 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | comedy | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.312 | FAL: 1234, TRU: 560 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | crime | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.144 | FAL: 1535, TRU: 259 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | documentary | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.002 | FAL: 1790, TRU: 4 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | drama | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.412 | FAL: 1055, TRU: 739 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | family | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.060 | FAL: 1687, TRU: 107 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | fantasy | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.100 | FAL: 1614, TRU: 180 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | history | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.027 | FAL: 1745, TRU: 49 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | horror | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.095 | FAL: 1623, TRU: 171 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | music | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.021 | FAL: 1757, TRU: 37 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | musical | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.012 | FAL: 1773, TRU: 21 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | mystery | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.080 | FAL: 1651, TRU: 143 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | romance | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.133 | FAL: 1556, TRU: 238 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | sci_fi | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.114 | FAL: 1590, TRU: 204 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | sport | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.016 | FAL: 1766, TRU: 28 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | thriller | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.169 | FAL: 1490, TRU: 304 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | war | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.016 | FAL: 1765, TRU: 29 | NA | NA | NA | NA | NA | NA | NA | NA |
| logical | western | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | 0.007 | FAL: 1782, TRU: 12 | NA | NA | NA | NA | NA | NA | NA | NA |
| numeric | year | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 2.002552e+03 | 8.980000e+00 | 1970.0 | 1998.000 | 2005.000 | 2.009000e+03 | 2.013000e+03 | ▁▁▂▅▇ |
| numeric | budget_2013 | 0 | 1.000 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 5.546461e+07 | 5.491864e+07 | 8632.0 | 16068917.500 | 36995786.000 | 7.833790e+07 | 4.614359e+08 | ▇▂▁▁▁ |
| numeric | domgross_2013 | 18 | 0.990 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 9.517478e+07 | 1.259653e+08 | 899.0 | 20546593.750 | 55993640.500 | 1.216784e+08 | 1.771683e+09 | ▇▁▁▁▁ |
| numeric | intgross_2013 | 11 | 0.994 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 1.978380e+08 | 2.835079e+08 | 899.0 | 33232603.500 | 96239640.000 | 2.414790e+08 | 3.171931e+09 | ▇▁▁▁▁ |
| numeric | metascore | 378 | 0.789 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 5.900100e+01 | 1.723800e+01 | 11.0 | 47.000 | 60.000 | 7.200000e+01 | 1.000000e+02 | ▁▅▇▆▂ |
| numeric | imdb_rating | 202 | 0.887 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 6.760000e+00 | 9.630000e-01 | 2.1 | 6.200 | 6.800 | 7.400000e+00 | 9.300000e+00 | ▁▁▅▇▂ |
| numeric | runtime | 203 | 0.887 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 1.109890e+02 | 2.091300e+01 | 55.0 | 97.000 | 107.000 | 1.220000e+02 | 2.710000e+02 | ▅▇▁▁▁ |
| numeric | profit | 11 | 0.994 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 1.421113e+08 | 2.547185e+08 | -120328632.0 | 5802238.500 | 54966804.000 | 1.645158e+08 | 3.024172e+09 | ▇▁▁▁▁ |
| numeric | ROI | 11 | 0.994 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 5.197000e+00 | 2.081800e+01 | -1.0 | 0.346 | 1.657 | 4.042000e+00 | 4.305170e+02 | ▇▁▁▁▁ |
This resulting dataset movies_df contains 1794
observations and 36 variables, indicating information about each movie’s
Bechdel test score, budget, revenue, genre, and ratings.
We looked at the overall distribution of movies from 1970 to 2013 that passed or failed the Bechdel test, and distribution of movies as per Bechdel test scores.
movies_df %>%
group_by(pass_bechdel) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N)) %>%
rename("Passes Bechdel Test" = pass_bechdel) %>%
knitr::kable(digits = 3) %>%
kableExtra::kable_styling(bootstrap_options = c("striped", "hover"))| Passes Bechdel Test | N | Proportion |
|---|---|---|
| FALSE | 849 | 0.473 |
| TRUE | 945 | 0.527 |
movies_df %>%
group_by(bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N)) %>%
rename("Bechdel Test Criterion" = bechdel_score) %>%
knitr::kable(digits = 3) %>%
kableExtra::kable_styling(bootstrap_options = c("striped", "hover"))| Bechdel Test Criterion | N | Proportion |
|---|---|---|
| Less than 2 women | 141 | 0.079 |
| Don’t talk to each other | 514 | 0.287 |
| Only talk about men | 194 | 0.108 |
| Dubious | 142 | 0.079 |
| Passes Bechdel | 803 | 0.448 |
The findings suggested that over half of the movies, i.e. 52.7% movies in the data passed the Bechdel Test. Stratifying by levels of Bechdel scores, it was seen that among the movies that passed the test, a small proportion had a dubious passing status. Among those that failed the test, most movies had women who didn’t talk to each other.
We conducted a further exploratory data analysis to answer our main questions:
seventies_df = movies_df %>%
filter(decade == "1970-1979") %>%
group_by(decade, bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N))
eighties_df = movies_df %>%
filter(decade == "1980-1989") %>%
group_by(decade, bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N))
nineties_df = movies_df %>%
filter(decade == "1990-1999") %>%
group_by(decade, bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N))
thousands_df = movies_df %>%
filter(decade == "2000-2009") %>%
group_by(decade, bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N))
tens_df = movies_df %>%
filter(decade == "2010-2013") %>%
group_by(decade, bechdel_score) %>%
summarise(N = n()) %>%
mutate(Proportion = N/sum(N))
decades = bind_rows(seventies_df, eighties_df, nineties_df, thousands_df, tens_df)
decades %>%
plot_ly(x = ~decade, y = ~Proportion*100, type = "bar", color = ~bechdel_score, colors = "RdYlGn") %>%
layout(barmode = "stack",
title = "Distribution of movies passing Bechdel Test Criteria by decade",
xaxis = list(title = "Decade"),
yaxis = list(title = "Percentage (%)"))This concluded that the proportion of movies that passed the Bechdel test increased over the decades. Due to a smaller sample of movies in the last decade (2010-13), we that the results seemed to have had leveled-off since 2000s. Also, the proportion of movies that decisively passed the test is still below 50%.
movies_df %>%
mutate(pass_bechdel = ifelse(pass_bechdel == TRUE, "Pass", "Fail")) %>%
plot_ly(x = ~budget_2013, y = ~pass_bechdel, type = "box", text = ~title, color = ~pass_bechdel, colors = c("red4", "green4")) %>%
layout(xaxis = list(title = "Movie budget ($)"),
yaxis = list(title = "Bechdel Test"),
showlegend = FALSE)movies_df %>%
plot_ly(x = ~budget_2013, y = ~bechdel_score, type = "box", text = ~title, color = ~bechdel_score, colors = c("darkred","orange1", "green4")) %>%
layout(xaxis = list(title = "Movie budget ($)"),
yaxis = list(title = "Bechdel Criterion"),
showlegend = FALSE,
title = "Movie budgets (2013-adjusted) stratified by Bechdel Test criteria")It was seen that the movies that passed the Bechdel test appeared to have smaller budgets than movies that don’t pass. On stratifying by the categorical Bechdel score, the budgets were not significantly different from each other. This possibly suggests that Hollywood puts more money behind male-only films than films in which women talk to each other.
movies_df %>%
mutate(pass_bechdel = ifelse(pass_bechdel == TRUE, "Pass", "Fail")) %>%
plot_ly(x = ~profit, y = ~pass_bechdel, type = "box", text = ~title, color = ~bechdel_score, colors = c("darkred","orange1", "green4")) %>%
layout(boxmode = "group",
xaxis = list(title = "Profit ($)"),
yaxis = list(title = "Bechdel Criterion"),
showlegend = FALSE,
title = "Movie profits stratified by Bechdel Test criteria")Movies failing the Bechdel test seemed to make more profit as compared to the ones that passed the test. The distribution of these profits across movies that passed and failed the Bechdel test was not appreciably different. Upon stratifying, there were no notable differences in the profits between Bechdel criteria.
We also checked for ROI to check for the ratio of profit to the budget (2013 inflation). Because ROI was heavily right-skewed, we used log-transform for this variable to visualize differences in the ROI distribution between movies of differing Bechdel criteria.
movies_df %>%
mutate(pass_bechdel = ifelse(pass_bechdel == TRUE, "Pass", "Fail")) %>%
plot_ly(x = ~log(ROI), y = ~pass_bechdel, type = "box", text = ~title, color = ~bechdel_score, colors = c("darkred","orange1", "green4")) %>%
layout(boxmode = "group",
xaxis = list(title = "Log(Return on Investment)"),
yaxis = list(title = "Bechdel Criterion"),
showlegend = FALSE,
title = "Movie ROI stratified by Bechdel Test criteria")Despite having slightly lower budgets, we could see that the primacy of women’s roles in movies did not negatively impact movies’ financial performance. This contradicts the claims that films with stronger female characters perform any worse at the box office than films without them.
We wanted to see how the fans received the movies that passed the Bechdel test. We plotted IMDB ratings (fan ratings) against Metacritic Scores (aggregate scores based on critic reviews) and stratified by Bechdel Score.
movies_df %>%
plot_ly(x = ~imdb_rating, y = ~metascore/10, color = ~bechdel_score, type = "scatter",
mode = "markers", colors = c("darkred","orange1", "green4"), text = ~title, alpha = 0.6) %>%
layout(yaxis = list(title = list(text = "Metascore", standoff = 5), gridcolor = "lightgray"),
xaxis = list(title = "IMDB Rating"), gridcolor = "gray",
title = "Movie IMDB and Metacritic Scores stratified by Bechdel Test criteria")There was no appreciable difference between the movies that pass and fail the Bechdel test in terms of their performance. There were more films passing the Bechdel Test with higher critic ratings relative to fan ratings, and more films that failed the Bechdel Test with lower critic ratings relative to fan ratings. One potential reason can be inherent biases held by fans to be more critical of films with female-dominant roles, while critics may be more likely to isolate their own biases while judging films.
To test and determine the strength of association, we used linear
regression models for categorical Bechdel score with
outcomes budget, profit,
imdb_rating and Metascores individually.
For these linear models, we used the
"less than 2 women" as the reference category. The
intercept was used to compare movies that fail all dimensions of the
Bechdel test with the movies that pass all dimensions of the Bechdel
test.
Because the budget was heavily right skewed, we log-transformed it to enforce normal distribution.
log_budget = movies_df %>% ggplot(aes(x = log(budget_2013))) + geom_histogram(alpha = 0.8, color = "white") +
labs(
x = "Budget ($, 2013-adjusted)",
y = "Count",
title = "Distribution of movie budgets")
ggplotly(log_budget)We quantitatively tested the association of the log of budget with categorical Bechdel score using linear regression model:
log_budget= movies_df %>%
lm(log(budget_2013)~bechdel_score, data=.)%>%
broom::tidy()
knitr::kable(log_budget, digits=3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 17.341 | 0.116 | 150.109 | 0.000 |
| bechdel_scoreDon’t talk to each other | 0.154 | 0.130 | 1.183 | 0.237 |
| bechdel_scoreOnly talk about men | -0.226 | 0.152 | -1.488 | 0.137 |
| bechdel_scoreDubious | -0.078 | 0.163 | -0.478 | 0.633 |
| bechdel_scorePasses Bechdel | -0.310 | 0.125 | -2.475 | 0.013 |
Based on parameter estimate = -0.31 with a p value = 0.01, passing all 3 Bechdel dimensions had a negative relationship with the Budget of the movies. i.e., the movies that passed all 3 Bechdel dimensions had a lower budget than movies that failed every Bechdel dimension.
Movies that failed every Bechdel dimension had an average budget
of $33,963,942
Movies that passed all 3 Bechdel dimensions had an average budget
of $24,910,750
Because the profit was heavily right skewed, we
log-transformed the data to enforce normal distribution. Additionally,
before the logarithm was applied, 1 was added to the base value to
prevent applying a logarithm to a 0 value. This was later subtracted
while calculating the results.
profit = movies_df %>% ggplot(aes(x = log(profit))) + geom_histogram(alpha = 0.8, color = "white") +
labs(
x = "Profit",
y = "Count",
title = "Distribution of Profits")
ggplotly(profit)We quantitatively tested the association of profit with categorical Bechdel score using linear regression model. We regressed profit against the categories adjusting for budget.
profit = movies_df %>%
lm(log(profit+1)~bechdel_score, data =.) %>%
broom::tidy()
knitr::kable(profit, digits=3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 18.139 | 0.160 | 113.117 | 0.000 |
| bechdel_scoreDon’t talk to each other | 0.080 | 0.179 | 0.447 | 0.655 |
| bechdel_scoreOnly talk about men | -0.012 | 0.208 | -0.057 | 0.955 |
| bechdel_scoreDubious | 0.071 | 0.224 | 0.317 | 0.751 |
| bechdel_scorePasses Bechdel | -0.282 | 0.173 | -1.625 | 0.104 |
Based on the parameter estimate = -0.044, Profits of movies that passed all 3 dimensions of the Bechdel test had lower profits than movies that failed all dimensions of the Bechdel test. However since the p-value is 0.7 > 0.05, the difference was not found to be statistically significant.
Movies that failed all dimensions of the Bechdel test made an
average profit of $75,526,942
Movies that passed all dimensions of the Bechdel test made an
average profit of $57,082,034
imdb = movies_df %>% ggplot(aes(x = imdb_rating)) + geom_histogram(alpha = 0.8, color = "white") +
labs(
x = "IMDB rating",
y = "Count",
title = "Distribution of IMDB")
ggplotly(imdb)This graph showed an approximately
normal distribution.
We quantitatively tested the association of IMDB rating with categorical Bechdel scores using linear regression
imdb = movies_df %>%
lm(imdb_rating~bechdel_score, data=.) %>%
broom::tidy()
knitr::kable(imdb, digits=3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 6.944 | 0.085 | 81.848 | 0.000 |
| bechdel_scoreDon’t talk to each other | -0.032 | 0.095 | -0.335 | 0.738 |
| bechdel_scoreOnly talk about men | -0.074 | 0.113 | -0.653 | 0.514 |
| bechdel_scoreDubious | -0.208 | 0.121 | -1.723 | 0.085 |
| bechdel_scorePasses Bechdel | -0.344 | 0.092 | -3.729 | 0.000 |
Based on the Parameter estimate = -0.344 with a p value = 0.00019, movies that passed all 3 dimesions of Bechdel test had a negative relationship with IMDB ratings.
Movies that passed all dimensions of the Bechdel test received
0.34 points lower IMDB ratings than movies
that failed all dimensions of the Bechdel test.
meta = movies_df %>% ggplot(aes(x = metascore)) + geom_histogram(alpha = 0.8, color = "white") +
labs(
x = "Meta score",
y = "Count",
title = "Distribution of Metascore")
ggplotly(meta)The graph showed an approximately
normal distribution.
We quantitatively tested the association of Metascore with categorical Bechdel scores using linear regression.
meta = movies_df %>%
lm(metascore~bechdel_score, data=.) %>%
broom::tidy()
knitr::kable(meta, digits=3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 58.730 | 1.636 | 35.897 | 0.000 |
| bechdel_scoreDon’t talk to each other | 1.306 | 1.841 | 0.709 | 0.478 |
| bechdel_scoreOnly talk about men | -0.591 | 2.155 | -0.274 | 0.784 |
| bechdel_scoreDubious | 2.032 | 2.324 | 0.874 | 0.382 |
| bechdel_scorePasses Bechdel | -0.464 | 1.775 | -0.262 | 0.794 |
Based on the parameter estimate = -0.4642, movies that pass all dimensions of the Bechdel test had a negative relationship with Meta scores. However since the p-value 0.7 > 0.05, the difference was not statistically significant.
Movies that passed all dimensions of the Bechdel test received
0.46 points less in Metascore than movies
that failed all dimensions of the Bechdel test.
We used stepwise regression to check if the Bechdel score was an
influential predictor of movie performance as measured by
profit, budget_2013, and review ratings. The
setup of the regression models here follows the methodology utilized for
the linear regression models including log-transforming variables like
profit and budget_2013 to enforce a normal
distribution. For the review rating variable, we had a choice of using
the Metacritic scores or the IMDB scores. Because
imdb_rating proved to be statistically significantly
associated with the Bechdel test in the linear regression analysis, we
decided to use this variable over the metascore. We
utilized a stepwise algorithm that optimized model AIC from the
MASS package. First, we tested both the binary and
categorical variables for the Bechdel test in the algorithm. Upon
running the models, we found that the binary variable was more likely to
be selected as a part of the final model due to having fewer degrees of
freedom, so that became the variable used in the final models.
Initially, we included return on investment, ROI as an
outcome variable as well, but after running the models, we found that it
displayed the same results as the profit variable. Because
of this, it was removed from the final analyses.
We first assessed the Bechdel test as a potential predictor of
profit. Other potential predictors included in the model
were budget_2013, imdb_rating, and all genre
variables.
# Stepwise regression for profit
## logProfit ~ Bechdel (binary) + logBudget + IMDB + Genre
modelprofit = lm(log(profit +1) ~ pass_bechdel + log(budget_2013) + imdb_rating + action + adventure + animation + biography + comedy + crime + documentary + drama + family + fantasy + history + horror + music + musical + mystery + romance + sci_fi + sport + thriller + war + western, data = movies_df)
stepprofit <- MASS::stepAIC(modelprofit, direction = "both", trace = FALSE) %>% broom::tidy()
knitr::kable(stepprofit, digits = 3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 4.205 | 0.599 | 7.020 | 0.000 |
| log(budget_2013) | 0.646 | 0.030 | 21.721 | 0.000 |
| imdb_rating | 0.395 | 0.041 | 9.539 | 0.000 |
| adventureTRUE | 0.453 | 0.101 | 4.463 | 0.000 |
| biographyTRUE | -0.338 | 0.179 | -1.886 | 0.060 |
| dramaTRUE | -0.369 | 0.085 | -4.330 | 0.000 |
| horrorTRUE | 0.547 | 0.128 | 4.264 | 0.000 |
| musicTRUE | 0.481 | 0.262 | 1.834 | 0.067 |
| romanceTRUE | 0.197 | 0.107 | 1.848 | 0.065 |
| sci_fiTRUE | -0.210 | 0.113 | -1.851 | 0.064 |
| westernTRUE | -2.685 | 0.496 | -5.412 | 0.000 |
The second metric we assessed as a measure of movie success was movie
reviews as measured by imdb_rating. Other potential
predictors we included in the model were budget_2013,
runtime, profit, award_winner,
and all genre variables.
# Stepwise regression for imdb ratings
## IMDB ~ Bechdel (binary) + logBudget + runtime + award_winner + profit + Genre
modelIMDB = lm(imdb_rating ~ pass_bechdel + log(budget_2013) + profit + imdb_rating + action + adventure + animation + biography + comedy + crime + documentary + drama + family + fantasy + history + horror + music + musical + mystery + romance + sci_fi + sport + thriller + war + western, data = movies_df)
stepIMDB <- MASS::stepAIC(modelIMDB, direction = "both", trace = FALSE) %>% broom::tidy()
knitr::kable(stepIMDB, digits = 3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 8.427 | 0.312 | 26.991 | 0.000 |
| pass_bechdelTRUE | -0.255 | 0.043 | -5.943 | 0.000 |
| log(budget_2013) | -0.080 | 0.018 | -4.450 | 0.000 |
| profit | 0.000 | 0.000 | 14.631 | 0.000 |
| actionTRUE | -0.418 | 0.060 | -6.930 | 0.000 |
| adventureTRUE | -0.212 | 0.065 | -3.266 | 0.001 |
| animationTRUE | 0.320 | 0.092 | 3.477 | 0.001 |
| biographyTRUE | 0.336 | 0.102 | 3.299 | 0.001 |
| comedyTRUE | -0.412 | 0.055 | -7.490 | 0.000 |
| crimeTRUE | 0.140 | 0.062 | 2.268 | 0.023 |
| dramaTRUE | 0.199 | 0.056 | 3.526 | 0.000 |
| familyTRUE | -0.462 | 0.089 | -5.218 | 0.000 |
| fantasyTRUE | -0.116 | 0.070 | -1.655 | 0.098 |
| horrorTRUE | -0.684 | 0.081 | -8.438 | 0.000 |
| musicTRUE | -0.236 | 0.139 | -1.703 | 0.089 |
| romanceTRUE | -0.139 | 0.065 | -2.138 | 0.033 |
| thrillerTRUE | -0.242 | 0.063 | -3.834 | 0.000 |
| warTRUE | 0.233 | 0.156 | 1.499 | 0.134 |
As budget is a significant predictor of both profit and IMDB ratings,
we also assessed whether passing the Bechdel test is an influential
predictor of a movie’s budget. Other potential predictors we included in
the model were runtime and all genre variables.
# Stepwise regression for budget
## logBudget ~ Bechdel (binary) + runtime + Genre
modelbudget = lm(log(budget_2013) ~ pass_bechdel + runtime + action + adventure + animation + biography + comedy + crime + documentary + drama + family + fantasy + history + horror + music + musical + mystery + romance + sci_fi + sport + thriller + war + western, data = movies_df)
stepbudget <- MASS::stepAIC(modelbudget, direction = "both", trace = FALSE) %>% broom::tidy()
knitr::kable(stepbudget, digits = 3)| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 13.521 | 0.186 | 72.873 | 0.000 |
| runtime | 0.026 | 0.001 | 17.882 | 0.000 |
| actionTRUE | 0.817 | 0.076 | 10.814 | 0.000 |
| adventureTRUE | 0.622 | 0.082 | 7.613 | 0.000 |
| animationTRUE | 1.300 | 0.124 | 10.492 | 0.000 |
| biographyTRUE | 0.326 | 0.134 | 2.433 | 0.015 |
| comedyTRUE | 0.287 | 0.073 | 3.917 | 0.000 |
| crimeTRUE | 0.173 | 0.080 | 2.157 | 0.031 |
| documentaryTRUE | -1.750 | 0.539 | -3.246 | 0.001 |
| dramaTRUE | -0.247 | 0.071 | -3.474 | 0.001 |
| familyTRUE | 0.677 | 0.115 | 5.890 | 0.000 |
| fantasyTRUE | 0.549 | 0.093 | 5.904 | 0.000 |
| musicalTRUE | 0.531 | 0.237 | 2.239 | 0.025 |
| mysteryTRUE | 0.483 | 0.101 | 4.771 | 0.000 |
| romanceTRUE | 0.285 | 0.084 | 3.388 | 0.001 |
| sci_fiTRUE | 0.424 | 0.090 | 4.726 | 0.000 |
| sportTRUE | 0.571 | 0.208 | 2.751 | 0.006 |
| thrillerTRUE | 0.331 | 0.083 | 3.976 | 0.000 |
The Bechdel test was selected as an influential predictor only for
imdb_rating and not for profit or
budget_2013. The IMDB ratings for movies that passed the
Bechdel test were lower on average than for movies that failed. Based on
this, passing the Bechdel test can be considered a neutral to negative
factor to a movie’s potential success.
At first, we explored the proportion of the movies from our dataset that passed the Bechdel test. We then also wanted to see if there was a change in the trend over time for the proportion of movies that passed the Bechdel test. We saw that 52.7% movies in the data passed the Bechdel test. There was also an increase in the proportion of movies that passed the Bechdel test over each decade.
For further analysis, we tried to measure the success of these movies based on factors like the budget, profit, and their IMDb ratings. While looking at these associations, we saw that the movies that passed the test had lower budgets and made lesser profits as compared to those that failed the test. Additionally, the movies that passed and failed the test had no appreciable difference between the IMDb ratings. On stratifying, there were more films that passed the Bechdel test and had higher critic ratings (compared to general audience ratings), and there were more films that failed the Bechdel test with lower critic ratings relative to audience ratings. This was suggestive of inherent biases held by general audiences that make them more critical of films with female-dominant roles, while critics may have been more likely to isolate these biases while judging films.
We found that a movie passing the Bechdel test plays an influential role in predicting IMDb ratings, but not profits or budget. Movies that pass the Bechdel test, on average, have lower ratings than movies that fail.
In conclusion, while not all results were statistically significant, our analysis pointed to a similar pattern, where passing the Bechdel test had a negative influence on our measures of success.
The Superheroes dataset was obtained using the Selector
Gadget Chrome Extension.
We created a custom IMDB list containing
all released movies from Marvel Cinematic Universe, as well as one for
the DC Extended Universe. Several DC Superheroes movies not in the
Extended Universe were also added. Using the Selector Gadget, we pulled
variables title and year.
Data on budget,
domestic and international gross was pulled using the Selector Gadget
from Wikipedia.
Data on the Bechdel test results for each movie was merged from our
movies_df dataframe, as well as bechdeltest.com.
The
datasets obtained from each website were cleaned and merged to create
the Superheroes_finally.xlsx file used for the Superheroes dashboard.
The Dashboard contains 2 graphs. One is depicting the
proportion of movies that pass the Bechdel test in the DC Universe and
the Marvel universe, represented as a bar plot. We can observe that
marvel has a higher percentage of movies that pass the test.
The
second graph depicts movies that pass the test over the years. The
purpose of this scatter plot is to see the cluster of movies that pass
the test is bigger post-2016 compared to before 2016, while no movie
passed the test before 2008. This trend shows that the amount of
superhero movies passing the Bechdel test is increasing.
finally,
we put the search table in the dashboard for the user to easily be able
to look at which movies passed the Bechdel test and which failed.