So Christenfeld decided to put spoilers to the test in the most straightforward way possible: by spoiling stories for people. In the initial experiment , his team had subjects read short stories from various genres. One group simply read a story and rated how much they liked it at the end.
The other group did the same, but the researchers spoiled the narrative, as if by accident, by giving them a short introduction. That one ends well. So there isn't any thought that with these great works of fiction, knowing the ending is going to ruin them. No one watches a romantic comedy truly wondering if the couple will be happy in the end.
With a detective story, you can safely assume the detective will eventually solve the case. And it turns out we didn't make them better.
Extra knowledge about a work of art makes it more enjoyable; when a spoiler is worked into the story itself, it simply makes for a flawed tale. If spoiling key plot points improves a story, perhaps the plot itself is simply a distraction that keeps us from enjoying the rest of it — the sensory descriptions, the character development, the satire, the artistry.
Spoilers clear away the need to think about the plot and allow you to enjoy the rest of the story more. Spoiler alert!
They found none. So they created their own by collecting more than 1. The reviews were collected from Goodreads, a social networking site that allows people to track what they read, and share thoughts and reviews with other readers.
Researchers found that spoiler sentences tend to clump together in the latter part of reviews. But they also found that different users had different standards to tag spoilers, and neural networks needed to be carefully calibrated to take this into account.
In addition, the same word may have different semantic meanings in different contexts. Identifying and understanding these differences is challenging, Wan said. Researchers trained SpoilerNet on 80 percent of the reviews on Goodreads, running the text through several layers of neural networks.
Christenfeld and Leavitt ran three experiments with a total of 12 short stories. Three types of stories were studied: ironic-twist, mystery and literary. Each story -- classics by the likes of John Updike, Roald Dahl, Anton Chekhov, Agatha Christie and Raymond Carver -- was presented as-is without a spoiler , with a prefatory spoiler paragraph or with that same paragraph incorporated into the story as though it were a part of it.
Each version of each story was read by at least 30 subjects. Data from subjects who had read the stories previously were excluded. Subjects significantly preferred the spoiled versions of ironic-twist stories, where, for example, it was revealed before reading that a condemned man's daring escape is all a fantasy before the noose snaps tight around his neck.
The same held true for mysteries. Knowing ahead of time that Poirot will discover that the apparent target of attempted murder is, in fact, the perpetrator not only didn't hurt enjoyment of the story but actually improved it. Subjects liked the literary, evocative stories least overall, but still preferred the spoiled versions over the unspoiled ones. The answers go beyond the scope of the study, but one possibility is perhaps the simplest one: that plot is overrated.
What the plot is is almost irrelevant. It's also possible that it's "easier" to read a spoiled story. Other psychological studies have shown that people have an aesthetic preference for objects that are perceptually easy to process. But the researchers are careful to note that they do not have a new recipe for writers to follow. After all, spoilers helped only when presented in advance, outside of the piece.
When the researchers inserted a spoiler directly into a story, it didn't go over quite as well.
0コメント