A Study with the Story Template

My daughter is doing a school project using my story template algorithm, and her goal is to examine classic and non-classic stories to see if there is any difference in the story structure. She decided to use just movies since the analysis is quicker, although still not trivial: for each story she broke it into a list of scenes, timed each scene, then calculated percentages of whole for duration and placement. Her definition of a classic was a film that had been adapted from a novel, had one or more sequels, and/or was recognized as the epitome of its genre. Furthermore, the original novel or film must have been made at least 25 years ago (1985 or earlier), since it takes at least about a generation to be recognized as a classic. Non-classics were films in that same time period that did not fit the "classic" criteria. She tried to choose films from a variety of genres.

She chose well-known movies:

Classics:
Charlotte's Web
Prince Caspian
Tuck Everlasting
High Noon
Raiders of the Lost Ark
Jaws
Rocky

Non-Classics:
Heaven Can Wait
War Games
LadyHawke

To analyze these movies, she studied the Story Template and then broke it down into 16 specific testable points. Note that I did NOT directly involve myself in the construction of her study; I gave her the template, then she worked with an advisor to create the research protocol. I was available to her and her advisor for an informed opinion if there were any questions.

She did an amazing amount of work, and I'm so proud of her! Most importantly, and contrary to all of my expectations, she found a real difference in structure between the classics and nonclassics.

15 of the 16 points were present in all the stories. However, 1 of the points was present in 6 out of 7 classics, but in NONE of the non-classics. This blew me away. When she did a Fisher's exact t-test for binary data on the presence or absence of this one variable in identifying a classic, in a two-tailed test (which is harder to reach significance), she had a p value of 0.03, considered significant. (The p value means that if you did this test 100 times, in 3 out of 100 trials you would expect to obtain these results by chance. Scientific standards typically accept a p < 0.05 to be considered significant, meaning that the scientist is probably measuring a real phenomenon). What this result indicates is that the presence of this one story point is highly correlated to having a "classic" whereas its absence means it is linked to being a non-classic.

Remember, though, that these aren't clean statistics since in the original project design she was looking at 16 variables. The likelihood with this many variables is that one might reach a level of significance with one of the variables just by chance. HOWVER, 1) none of the other points changed -- they were all present in both classics and non-classics; and 2) this point makes a lot of sense to me that it might distinguish the lasting stories from the throwaways. At the minimum, it seems to be important to remember to include this point. It sure can't hurt!

I'm sure you're wondering what this variable is? Well, this blog entry is already long, so I'm going to save that till Wednesday. In the meantime, I'd love to hear what you think it might be. Happy writing.