Normalizing Star Wars

It's been about three weeks since I went to see Star Wars: Episode III on opening night and the hype about it being the most popular movie ever has sort of died down a bit. The media based it's claim of Episode III being the most popular movie ever based on it's record setting take of $158.5 million over it's opening weekend (or technically opening three days starting early Thursday morning at 12:01). I've always wondered about the validity using the opening weekend box office take to measure the popularity of a movie compared to movies in the past. To me using opening weekend box office as a metric seems to suffer from a lack of normalization. In other words, we are measuring something and comparing it across time when that something has a time dependence (ie the cost of a movie ticket is increasing over time, and dollars are not constant). This is a problem that I encountered a lot while analysing stats for my thesis in grad school: the raw number of train derailments has increased but since train traffic had doubled over a certain time period, the rate of accidents per train-mile had decreased. It's really the rate that is important, not the raw number. In a similar manner, I figured that the best measure of a movie's popularity would be the percent of the US population that saw the movie during it's opening weekend. So I decided to set about "normalizing" some of the Star wars numbers to see if Episode III truly was more popular than some older releases.

My first step was to convert $158.5 million into an estimate of the number of actual people who attended the movie. A quick google search revealed the average cost of movie tickets over time and for 2005 it is at $8.50, translating into 18.6 million people attending Episode III during it's first three days. The estimated population f the United States for 2005 is 296 million. Thus, 6.3% of the population of the United States saw Episode III during it's first three days in theatres. I could normalize further by breaking down the population pyramid to remove those under the age of 13 and those over 60 who are unlikely to attend this movie to get a more accurate percentage of the target market but I didn't have quite enough data to go to that level of detail. So we'll run with 6.3& for Ep III.

This figure is easily going to trump any recent releases, but could the inflation in movie ticket prices actually make on older release more popular?

Of the top 80 opening weekends ever (measured by raw unadjusted dollars) only three qualify as being not-so-recent: #54 Jurrassic Park with $47million in 1993, #60 Batman Returns with 46 million in 1992, and #78 Batman with $40 million in 1989. With early 90s average ticket prices around $5, each of these movies drew about 8 million people opening weekend or approximately 3 percent of the US population. So Episode III easily trumps the early 90s blockbusters.

What about the other movie that ended a Star Wars trilogy, Return of the Jedi which smashed records when it was released in 1983? ROTJ took in $23 million on its opening weekend, but at that time the average movie ticket was $2.50, so that translates into about 9.2 million people. The population of the US in 1983 was 234 million, so 3.9% of the US population saw ROTJ opening weekend. So Episode III is about twice as popular as Return of the Jedi when measured in this manner. Interestingly, these figures show that ROTJ was more popular in its time than Jurrasic Park, Batman or Batman Returns.

What of the other original trilogy? Star Wars took in $1.5 million its first weekend, but that was back in 1977, before the idea of seeing a summer blockbuster opening weekend had even entered the conscience of the public. So only 0.4% percent of the US population at the time saw it opening weekend. The Empire Strikes back fared a bit better, with a take of $6.4 million translating into 1.4% of the US population.

Based on these figures, one could say that Episode III is the most popular movie of the past 30 years measured by percent of US population seeing it during its first three days of release. Interestingly, if the top movies were ranked this was, ROTJ (ranked number 230 in terms of all time opening weekend gross) would move all the way up to number 5 behind Episode III, Spider-Man, Shrek 2, and Harry Potter: Prisoner of Azkaban, and into a virtual tie with Indiana Jones and the Temple of Doom which took in $25 million in 1984.

Of course, one can not overlook the fact that despite earning $1.5 million its first week, Star Wars went on to have a total box office gross of over $780 million. So these numbers are far from perfect and there-in lies the trouble with using opening weekend numbers to gauge popularity. Movies vary wildly in the percentage of total gross earned on the opening weekend. Of the Top 80 ranked by constant dollar gross, Batman from 1989 has the lowest opening weekend percentage with 16.1% ($40.5 million out of $251 million). Hulk meanwhile, is the biggest box office staller with an opening weekend percentage of 47% ($62 million out of $132 million). So opening weekends can be poor indicators of long term popularity but after one week in theatres and a media hungry for an answer on success, it's the only data available. So we're stuck with it. However misleading it might be.