add bechdel data and script

README.md

@@ -2,3 +2,5 @@ Article Date | Headline | File or folder
 ---|---------|-------------
 March 17, 2014 | [FiveThirtyEight’s NCAA Tournament Predictions](http://fivethirtyeight.com/interactives/march-madness-predictions) | `march-madness-predictions`
 March 27, 2014 | [The NCAA Bracket: Checking Our Work](http://fivethirtyeight.com/datalab/the-ncaa-bracket-checking-our-work) | `historical-538-ncaa-tournament-model-results.csv`
+April 1, 2014 | [The Dollar-And-Cents Case Against Hollywood’s Exclusion of Women](http://fivethirtyeight.com/features/the-dollar-and-cents-case-against-hollywoods-exclusion-of-women) | `bechdel`
+

bechdel/analyze-bechdel.R

@@ -0,0 +1,79 @@
+# Analyze movie data from the following webistes: www.bechdeltest.com and www.the-numbers.com,
+# calculate summary statistics and conduct basic regression analysis to test whether movies
+# which pass the Bechdel test have better or worse in boxoffice profits.
+
+# By Andrew Flowers ([email protected])
+# See also http://fivethirtyeight.com/features/the-dollar-and-cents-case-against-hollywoods-exclusion-of-women/
+# Dependent files: "movies.csv"
+
+# Install and load required packages
+# install.packages(c("gdata", "cwhmisc")) 
+library(gdata)
+library(cwhmisc)
+
+# Load data
+rawData<-read.csv("movies.csv", na.strings="#N/A")
+
+# Select movies pre-1990, and format $-denominated data fields
+rawData<-rawData[rawData$year>1989,]
+
+# International-only gross profits (which equal total profits minus domestic profits)
+rawData$intOnly<-rawData$intgross_2013.-rawData$domgross_2013.
+
+# Return on Investment (ROI) measures
+rawData$ROI<-rawData$intgross_2013./rawData$budget_2013. #  Total ROI
+rawData$ROI1<-rawData$domgross_2013./rawData$budget_2013. #  Domestic ROI
+rawData$ROI2<-rawData$intOnly/rawData$budget_2013.  #  International ROI
+
+# Divide movies into FAIL and PASS divisions
+failMovies<-rawData[rawData$binary=="FAIL",]
+passMovies<-rawData[rawData$binary=="PASS",]
+
+# Include a "generous" category (which includes both "ok" and "dubious" movies)
+generous<-rbind(rawData[rawData$clean_test=="ok",], rawData[rawData$clean_test=="dubious",])
+
+# Print medians: ROI and budget
+median(failMovies$ROI, na.rm=T)
+median(passMovies$ROI, na.rm=T)
+median(rawData$ROI, na.rm=T)
+
+median(failMovies$budget_2013.)
+median(passMovies$budget_2013.)
+median(rawData$budget_2013.)
+
+# Distributions and logs
+hist(rawData$budget_2013.)
+hist(log(rawData$budget_2013.))
+
+hist(rawData$intgross_2013.)
+hist(log(rawData$intgross_2013.))
+
+hist(rawData$ROI)
+hist(log(rawData$ROI))
+
+# Linear regression models
+
+# Movies with higher budgets make more gross revenues
+summary(lm(log(intgross_2013.)~log(budget_2013.), data=rawData))
+
+# Bechdel dummy is not significant
+summary(lm(log(intgross_2013.)~log(budget_2013.)+factor(binary), data=rawData))
+
+# Movies with higher budgets have lower ROI
+summary(lm(log(ROI)~log(budget_2013.), data=rawData))
+
+# Bechdel dummy is not significant
+summary(lm(log(ROI)~log(budget_2013.)+factor(binary), data=rawData))
+
+# ROI #1 (domestic) used in chart
+median(generous$ROI1, na.rm=T)
+median(rawData$ROI1[rawData$clean_test=="men"], na.rm=T)
+median(rawData$ROI1[rawData$clean_test=="notalk"], na.rm=T)
+median(rawData$ROI1[rawData$clean_test=="nowomen"], na.rm=T)
+
+# ROI #2 (international) used in chart
+median(generous$ROI2, na.rm=T)
+median(rawData$ROI2[rawData$clean_test=="men"], na.rm=T)
+median(rawData$ROI2[rawData$clean_test=="notalk"], na.rm=T)
+median(rawData$ROI2[rawData$clean_test=="nowomen"], na.rm=T)
+