Dave is still crusading over his exclusion from the Twitter SUL. Today, Dave would like to teach you about a few different statistical concepts:
It’s pretty clear something happened in July.
We know this much — TechCrunch was dropped from the Suggested User List, right around the time their follower count started heading down. As to why, we can only speculate that it was because they ran a piece that Twitter didn’t like.
7/16/09: Twitter’s Internal Strategy Laid Bare: To Be “The Pulse Of The Planet.”
People have always questioned whether there was a connection between being on the list and not being too critical of Twitter. At this point, there isn’t much doubt that the connection is there.
Here are the concepts that Dave would like to teach you about:
Sample Size — It is commonly believed that you need to have a large sample size to draw any meaningful conclusions. Dave disagrees. In his world, one data point is sufficient.
Causation — In Dave’s world, post hoc ergo propter hoc isn’t a logical fallacy, but a natural law. TechCrunch was removed from the SUL after that article, therefore that article was the cause of the removal.
Correlation — With the Dave Winer statistics model, you don’t need to consider evidence that you don’t like. Have any of the folks on the SUL written critically of Twitter? Who cares? That doesn’t matter! As long as the data fits with your theory, accept it, otherwise ignore it!
Science just got a whole lot easier.
