In “Kill All Others” (the TV adaptation of Philip K. Dick’s story, The Hanging Stranger), it describes a world in which an ambiguous “other” group of people is used as a political tool to make people mistrust each other. Forcing everyone to question, who are the others? How can I tell if someone is an “other?” The nation devolves into a chaos of confusion and violence. At the climax the protagonist, Philbert Noyce, declares, “We are ALL others!”
While the story is sci-fi, the statement is true — we are all outliers in some way. I’m an outlier in that I’m a raw fiber artist. You can probably think of several points where you’re an outlier—can sing the periodic table song from memory, rare orchid aficionado, dulcimer savant.
The problem with outliers is that our puny human brains don’t know what to make of them. To simplify things, we toss them out or cram them in. Data science, stats, and analytics try to find ways to get a singular view of the majority. We strive towards universal answers and panic when things fall outside the bucket we’re holding.
What we need is Computational Inclusion — a way to accommodate the outliers without diminishing them. A way to relish their uniqueness and even elevate the mundaneness of the masses. We only see the mundane because we collect the most banal attributes of ourselves — income, zip code, family size, age. If we begin to capture the richness of ourselves, we can design for the richness of ourselves.
Data is the archaeology of modern living. It allows us to recover our possible pasts — reconstructing images of ourselves based on invisible artifacts. The design we apply to data looks forward to our possible futures.
Let’s use GDP as an example. GDP, an aggregate number, was designed in the 1930s as a way to measure production in the manufacturing age. Specifically, it equates to:
private consumption + business investment + government spending and investment + (exports — imports)
…which is an unfortunately simplistic way of valuing a country. It doesn’t measure our quality of life, infant mortality, or pollution levels. Basically, it’s about how much stuff we can thrust upon the world regardless for how good or bad for us it is — and if we’re thrusting ours around the globe more than we’re taking in, all the better.
And money follows the GDP — the ideal range of growth being 2–3%. Too much growth and we fear inflation; too little and we fear a recession. We’ve become so mired in the tight margins of this measurement that it becomes nearly impossible to adjust for things like climate change mitigation, affordable healthcare, and livable wages.
However, the data used to design the GDP in inherently biased. Invisibly embedded in it are the disparities of income, favorability of work outside the home, and rejection of resourcefulness. It’s a number that was designed at a time when data collection was still done manually at when we valued certain people’s work over others. There was no incentive, nor was it technically feasible, to be computationally inclusive. It’s 2018, and that is no longer true.
In 2008, led by Nobel economist Joseph Stiglitz, France conducted a study on the healthiness of the GDP and concluded that it “mismeasures our lives.” In essence, we’re conflating economic growth (in broad terms) with our well-being. Then French President Sarkozy proclaimed that we’re creating a “gulf of incomprehension between the expert certain in his knowledge and the citizen whose experience of life is completely out of sync with the story told by the data.”
Singular numbers, like scores and ratings, makes it easy for people to chart and assess. There’s comfort in a singular number to represent things that are complex and hard to quantify. It’s easy to accept at face value and embrace its simplicity. We have to start dismantling the singular values that hide the nuance that affects so many people.
We have an opportunity, and likely soon an imperative, to take advantage of near infinite computing power and our data hoarding proclivities. Our data artifacts hold a richness — details we’ve ignored because we want quick, cheap answers. We take the fat part of the bell curve and assume it will give us the fattest opportunities. But if we generally believe in the 80/20 rule, we should recognize that the head and tail of the curve is what really drives the bulk of change.
By embracing the outliers we can be more computationally inclusive. By looking deeper into the human experience to find what we truly value collectively we can begin to craft better formulas. It will embolden us to see more intricate data images of ourselves and hasten experiences that speak to each of us as individuals.