Impact factor is kind of like this:
Like a giant gorilla, you may love it, you may hate it, but you cannot ignore it.
This was particularly driven home to me last month listening to a panel of editors at The Crustacean Society meeting. They all kept talking about impact factors. You can see the same obsession with publishers quick to crow over any increase in impact factor, regardless of how trivially tiny it is.
Woo! Up a whole 1.5%!
One of the well-known flaws is that it’s an average. It doesn’t tell you anything about the individual articles.
In science, we report averages all the time. But one of the fastest ways to draw a criticism is to put an average value on a graph with no error bars.
Impact factor is an average with no error bars.
Impact factor would be improved by reporting some other measure of distribution (e.g., standard deviation). Then you would get some idea if the impact factor for a journal was being driven by a few powerhouse papers, with most sitting getting little attention, or whether the journal was consistently publishing solid performing papers.
Journal bragging pic by Functional Neurogenesis on Flickr; gorilla by Lady/Bird on Flickr; used under a Creative Commons license.