RESULTS
Subject Profiles
Title |
Original Release Date/Form |
Genre |
Gold Standard |
Story Template Algorithm |
Charlotte’s Web |
1952/Children’s Book |
Children’s |
Classic |
Classic |
Chronicles of Narnia: Prince Caspian |
1951/Children’s Novel |
Fantasy/Children’s |
Classic |
Classic |
Heaven Can Wait |
1978/Film |
Comedy |
Non-Classic |
Non-Classic |
High Noon |
1952/Film |
Western |
Classic |
Classic |
Indiana Jones: Raiders of the Lost Ark |
1981/Film |
Action/Adventure |
Classic |
Non-Classic |
Iron Man |
1963/Comic Book à 2008/Film |
Action/Adventure |
N/A |
Predicted Classic |
Jaws |
1974/Novel |
Horror |
Classic |
Classic |
Lady Hawke |
1985/Film |
Romance |
Non-Classic |
Non-Classic |
Rocky |
1976/Film |
Drama |
Classic |
Classic |
Tuck Everlasting |
1975/Young Adult Novel |
Children’s |
Classic |
Classic |
War Games |
1983/Film |
Science Fiction |
Non-Classic |
Non-Classic |
This chart shows the stories that were studied in this project. Seven stories were identified as classics by the Gold Standard as per the criteria mentioned in the Methods section: Charlotte’s Web, Prince Caspian, High Noon, Raiders of the Lost Ark, Jaws, Rocky, and Tuck Everlasting. The Story Template algorithm correctly identified six of the seven classics as classics. Raiders of the Lost Ark was identified as a non-classic by The Story Template algorithm.
Three stories were identified as non-classics by the Gold Standard as per the criteria mentioned in the methods section: Heaven Can Wait, Lady Hawke, and War Games. The Story Template algorithm correctly identified all three of these non-classics as non-classics.
For fun, one modern story was included in this study to forecast whether it had the potential to become a classic. Iron Man was predicted to be a classic by The Story Template algorithm.
Sensitivity, Specificity, and Accuracy
Statistical Analysis
2x2 Contingency Table of Binary Data
|
Template Classic |
Template Non-Classic |
Total |
Gold Standard Classic |
6 |
1 |
7 |
Gold Standard Non-Classic |
0 |
3 |
3 |
Total |
6 |
4 |
10 |
This chart: 2x2 Contingency Table of Binary Data classifies the binary data (yes/no) into a table. The Story Template correctly identified six of the seven gold standard classics, and misidentified one classic as a non-classic. The Story Template also correctly identified three of three non-classics.
The above results were analyzed using a Fisher’s exact test, and were shown to be statistically significant (p < 0.03 for a two-tailed analysis). This significant result suggests that The Story Template accuracy (positive and negative detections compared to a “true” standard) is a real phenomenon, not due to chance. The study design of 16 factors, though, clouds strong interpretation of this p value as more than a suggestive trend.