Annals of Improbable Research, Vol. 5, No. 3, pg. 10.

Script an Astronomer, Then Reach for the Stars

Eric Schulman
National Radio Astronomy Observatory
Charlottesville, Virginia

Why do so many Hollywood movies fail? The fault lies not in their stars, but
in the professions of the characters. In its search for success, the movie industry
has under-utilized a proven winner: a movie is most likely to succeed when it
has astronomer characters.

1. Introduction
    The holy grail (Nadis 1996) of Hollywood is a formula to predict the popularity of a movie before it's made. Here we present just such an algorithm, which predicts a linear relationship between the perceived quality of a movie and the number of characters who are astronomers or astrophysicists. That this correlation hasn't been discovered before is not surprising because of the small number of astronomers in "the industry."

2. Methods
    We used the Internet Movie Database (IMDb; Bernhardt et al. 1997) to search for movies with astronomer or astrophysicist characters that had been evaluated by at least ten IMDb users. Seventeen such movies were found by searching for "astronomer" and "astrophysicist" in the character name and plot summary fields and for "astronomy" in the genre field.
    The number of astronomer (or astrophysicist; the terms will be used interchangeably hereafter) characters was estimated from the IMDb character lists, IMDb plot summaries, and the memory of the senior author for films (or movies; the terms will be used interchangeably hereafter) that had been previously viewed by this researcher.
    The referee suggested that we view each of the movies in our sample to more precisely determine the number of astronomer characters, but we decided that this would strain our research budget.

3. Results
    The number of astronomer characters and the IMDb rating [on a scale from 1 to 10, with 1 being worst and 10 being best; the lowest-rated movie we could find was Manos, the Hands of Fate (1966), with a rating of 2.3, and the top film we found was Star Wars (1977), with a rating of 8.9 (the referee pointed out that one of the films listed in Table 1 below has a rating of 9.2, but this film only had ten votes and we restricted the search we just described to films with at least fifty votes, on the grounds that this information was easily found in the IMDb)] are plotted in Figure 1 for each of the seventeen movies.
    Linear regression analysis revealed a slight correlation between the number of astronomers and the IMDb rating. Since this slight correlation didn't agree with our preconceived ideas (uh...we mean with our preliminary hypothesis), we needed to decide whether to discard the hypothesis or to discard some of the data (up to half; Schulman 1996; Schulman and Cox 1997; Schulman, Cox, and Schulman 1999).
    Since the eight movies having one astronomer character and IMDb ratings greater than 7.0 were obvious outliers, we discarded these and then re-ran the linear regression analysis, which resulted in a fit that supported our preliminary hypothesis (Figure 1). The reasons for discarding each outlying data point are summarized in Table 1.

Figure 1

Table 1: Obvious Outliers

Title (Year)                     IMDb   Reason for Discarding

Colpo di luna (1995)              9.2   This film is set in Italy.

Twelve Monkeys (1995)             8.2   This is a time travel movie,
                                        and we don't believe in time

The Sea Hawk (1940)               8.1   We don't believe that the
                                        Spanish Armada has anything
                                        to do with astrophysics.

Straw Dogs (1971)                 7.9   This film is set in England.

Addicted to Love (1997)           7.3   We never saw this movie so
                                        it couldn't have been all
                                        that good.

Moonlight and Valentino (1995)    7.3   The astrophysicist in this
                                        movie dies off-screen during
                                        the first scene. 

Top Gun (1986)                    7.2   The lead actress described
                                        this role as "just another 
                                        dumb astrophysicist," which
                                        we take offense at.

Roxanne (1987)                    7.1   This is a modernization of
                                        a French play. 'Nuff said.

4. Discussion
    The fact that there is a correlation between the number of astronomers and the quality of a movie shouldn't be surprising. Contact (1997) was very highly regarded and it has a large number of astronomer characters. On the other hand, everyone can think of many very bad films that don't have any astronomer characters [for example, The Incredibly Strange Creatures Who Stopped Living and Became Mixed-Up Zombies!!? (1964), with an IMDb rating of 3.1].
    Despite the lack of astronomers in the movie industry (Cox and Schulman 1998), astronomers are still over-represented in films compared to their numbers in society. More than 0.1% of movies with IMDb ratings have astronomer characters even though only about 0.001% of the population are astronomers. This factor of 100 over-representation is due to the unconscious realization on the part of movie producers that astronomer characters increase film sales.
    Note also that the seventeen films with astronomer characters have an average IMDb rating of 7.1, significantly higher than the average of 5.5 that would have been expected by chance [the referee pointed out that we had discarded eight of these movies in a previous section, but we suspect that this referee is none other than Dr. Earle E. Spamer, who has been known to publish results that are incorrect by five orders of magnitude (Spamer 1997)].

5. Conclusions
    We have shown that there is a relationship between the number of astronomer characters and the perceived quality of a film. The average film with no astronomer characters is only in the 50th percentile for quality, while the average film with at least one astronomer character is in the top 10% of movies; the average film with more than five astronomer characters is in the top 0.3% of movies.
    It is obvious that there are currently too few astronomers in the film industry, and yet it's also true that only about 25% of Ph.D. astronomers are able to stay in astronomy due to a shortage of permanent jobs in the field. Clearly, it's time for the smart studios to start hiring as many astronomers as they possibly can before their competition realizes what a gold mine they have been overlooking.