Facebook and smoothing our data: a master class in data manipulation

26 November 2021 – In this weekend’s edition of “Brain Droppings” (which you can access here) one of our most-read articles concerned the Facebook studies which show that thousands of vulnerable people are harmed by Facebook and Instagram but lost in Meta’s “average user” data. As noted by my Social Media Manager, Catarina Conti, it’s a subtle point, easily missed: small average changes can mask big variations.

Note: this is not smoothing in statistics and image processing where you smooth a data set is to create an approximating function that attempts to capture important patterns in the data, while leaving out noise or other fine-scale structures/rapid phenomena. For example, the Federal Reserve uses smoothing techniques to help show the economic trend in data. These are statistical manipulations referred to as “smoothing techniques” and are designed to reduce or eliminate short-term volatility in data. But they are explained in copious footnotes. A smoothed series is preferred to a non-smoothed one because it may capture changes in the direction of the economy better than the unadjusted series does.

The article about Facebook is “The Thousands of Vulnerable People Harmed by Facebook and Instagram Are Lost in Meta’s Average User Data” and I want to draw your attention to a few paragraphs:

“Consider a world in which Instagram has a rich-get-richer and poor-get-poorer effect on the well-being of users. A majority, those already doing well to begin with, find Instagram provides social affirmation and helps them stay connected to friends. A minority, those who are struggling with depression and loneliness, see these posts and wind up feeling worse. If you average them together in a study, you might not see much of a change over time.”

The write up also points out:

“The tendency to ignore harm on the margins isn’t unique to mental health or even the consequences of social media. Allowing the bulk of experience to obscure the fate of smaller groups is a common mistake, and I’d argue that these are often the people society should be most concerned about. It can also be a pernicious tactic. Tobacco companies and scientists alike once argued that premature death among some smokers was not a serious concern because most people who have smoked a cigarette do not die of lung cancer”.

I like the word “pernicious.” But the keeper is “cancer.” The idea is, it seems to me, that Facebook – sorry, Meta — is “cancer.” Cancer is a term for diseases in which abnormal cells divide without control and can invade nearby tissues. Cancer evokes a particularly sonorous word too: Malignancy.

Is Meta smoothing numbers the way the local baker applies icing to a so-so cake laced with a trendy substances like cannabutter and cannaoil? My hunch is that dumping outliers, curve fitting, and subsetting data are handy little tools Meta uses – just to help the data. Without telling anybody.