Google and the Centipede Fallacy
The recent study on the value of news may not say quite what they claim
For this week we will try something slightly different, in that I’ll be linking out to an article rather than you getting it right here.
Early this week I was drawn into a debate with Benedict Evans on a study that Google had published in which they claimed to have proven that news had no value to them.
I made the point that seeing something in an experiment run on 1% of a population was not the same as rolling it out to 100% as a policy. He accused me of being a sampling denier and of clinging to the past, which is pretty standard stuff.
This had piqued my interest though, so I did spend some time going through the study with the same rigour as the Google PR team had done. By that time
from had picked up on the same issue and so happily my analysis has ended up being published there.I think it does bear out my initial suspicion that while the science behind the experiment is sound, the context and the interpretation of it is more subjective. and The results should be treated with that in mind, particularly when trying to determine what value news has to the dominant search engine.
The piece can be read If you click here - enjoy!