Another examination of (ab)use of statistics in media, in this case culminating in a pretty damning takedown.
I received one comment on my BLS post that said little more than “numbers lie”. Ironically, the BLS statistical manipulation (manipulation here has the data scientist definition, not the misunderstood public one) was completely valid, and the point about the analysis was more about politicians bloviating about it than it was about the value of seasonal correction (which as someone working in energy I can tell you is not only robust but also extremely important to make sense of a lot of data).
In this case though, the source article at Vox seems to be ignoring context to make a point that, when held to a fairly basic level of scrutiny, falls apart. Also key is that the new conclusions are much more interesting: future shock may not be a thing, but as communications advance at step-change levels, the socioeconomic forces actually behind adoption curves become more important to consider. This is a key and likely troubling conclusion in our current age of economic inequality. I could rant further, but I’m going to save that for an article on payday loans and check cashing services so I can become really angry at economic access issues.