gwern comments on Make evidence charts, not review papers? [Link] - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (4)
Looking at that sleep one... I thought it looked nice initially, but then I tried to look at the information in more detail, and I became less impressed. The text was doing all the semantic heavy-lifting.
Much of the heavy lifting is also done by the assignment of numbers and colors to indicate the impact of the experiment on a hypothesis. That's much easier to grok as a whole than plain text. I can also easily make quick judgments from the chart that are much more difficult to do from a review paper, such as "later experiments generally oppose this hypothesis, and only early experiments strongly support it" (among those in the chart, of course).