The Distortion Machine

Yesterday, I was asked to participate on a panel at the SSP's 40th Anniversary Annual Meeting on the topic of measuring the real world impact of research and scholarship. It was a great panel, and everyone delivered the goods. My portion (proudly delivered without slides) dealt with what I think is a main issue that too often is accepted or ignored -- the fact that the distributors of content are no longer neutral, and in fact have real effects on who sees what and when. These are the Facebooks, Googles, and Twitters of the world.

Below are my notes from the session, presented without edits. 

The Internet inherently brings amplification and persistence effects to information. Small glitches in science are amplified and preserved far beyond reason. The vaccines and autism scandal of the late-1990s was a mistake in the scientific literature that, because of the amplification and persistence effects of the Internet, is still being used to create confusion, a situation that continues to cause unnecessary illness and death.

There is an important and often-overlooked factor making it difficult for good, well-intentioned organizations to reach their audience or the general public.

This factor affects our metrics, our citations, article discovery, and public awareness of new and good science.

This factor is huge, well-funded, and pervasive.

Because of this factor, our current era may well become known as information’s equivalent to the “juicing era” in baseball, when homerun hitters like Mark McGuire, Barry Bonds, and Sammy Sosa set records juiced on steroids – requiring all the records to be corrected when the true and illicit source of their productivity was revealed.

The elephant in the room is that our distributors are no longer trustworthy.

Postal service was neutral as a delivery service.

Email is neutral as a delivery service.

Social media, YouTube, and major search engines are not neutral distributors, and their biases have outsized and persistent effects.

“Many of them have the powers that we used to associate with governments.” – Farhad Manjoo, the NYTimes.

The US distribution of Facebook is larger than the circulation of all newspapers combined.

80% of Altmetric’s measurements come from Twitter activity.

Google is every journals dominant search engine source.

Social media companies exploit outdated legal concepts describing them as platforms, not media companies.

They have been tacit participants in the emergence of global terrorism.

They reflect biases and beliefs about the world that most of us would easily identify as wrong, racist or sexist, or just inadequate and simple-minded.

Their business model favors extreme information, which often means untrustworthy or shocking information, which often equates to misinformation.

Until the intermediaries and distributors are made neutral again, which means they will serve the content community (or at least behave according to the same social norms), this problem will be impossible to address.

Until then, any metric, citation, or search you run has to be viewed as if it is juiced by some algorithm – or the after-effect of some algorithm – an algorithm you don’t understand.