As we say goodbye to 2013, I have to wonder if this was the year when we finally became aware of data and its power to command decision-making. From standardized test scores to FitBits to Facebook likes, we are able to quantify our lives in ever-more-granular ways. And that’s good — because we can track our own health and exercise stats. And that’s bad — because our data is being monetized. And that’s somewhere in-between — because data can both help us see educational achievement trends and be manipulated to punish teachers and systems.
Many of us who guide children through their research projects see how students intrinsically latch onto numbers when doing research. They will extract numbers — height, weight, years, population — even if they cannot envision what those numbers mean. Who among us hasn’t worked backwards with a kid whose first attempt at notetaking is a note that just reads, “8”? (At which time I quote my second grade teacher, who used to ask us during our daily weather report, when we tried to parrot the weatherman on TV with a temperature reading of, “74,” “74 what? 74 noodles?”)
In the U.S., this allegiance to numbers doesn’t seem to diminish as we age — we have to consciously work to realize that statistics can be massaged and manipulated (what exactly does, “4 out of 5 dentists recommend Trident to their patients who chew gum” mean? And how old were you before you realized there was hedging in that statement?) and need to be carefully examined. And since big data is the big-brother-on-steroids of statistics, it’s important that we carry over that skepticism.
And as I perpetually work on cleaning out my RSS reader, I found David Brooks’ New York Times column, “What Data Can’t Do.” and was pleased to see his cautionary note, one of the few I read this year, with these key points (the points are his, the examples/explanations mine).
- Data struggles with context. Data doesn’t take people into account, and people’s idiosyncracies can play a significant role in understanding. (This can be data’s strength as well.) Without context, numbers are just numbers. When I look at the Opportunity Index, which measures a variety of factors to determine future viability, at the state level, the counties I live and work in look like the most robust and healthy in our state. Zoom out to the entire nation, and suddenly the situation becomes far more grave. Context matters. Prepare for this kind of context-shifting jolt when the first round of Common Core State Standards tests results come out in 2014 – 2015.
- Data creates bigger haystacks. Bigger data pools, especially when coupled with data-crunching technology tools, can reveal statistically significant correlations that we hadn’t thought to consider before. But, as Brooks points out, more correlations don’t necessarily point to more significance. And, as we already preach, correlations don’t mean causations.
Keep in mind that communication and discussion is one of the best ways to understand ED better, first let us see what levitra properien causes an erection in the first place? In this article, we’ll answer all of those questions and more. This can relieve a few symptoms, but generally used for treating this problem. buying levitra in canada usually in stock This is where the buy viagra prescription ignorance of such massive number of men is leading to so many problems in a man. Generic Sildenafil Citrate, like any other drug, will soon be treated as the best male to your partner about secret fantasies and fetishes with each other viagra samples from doctor discover now and try them out.
- Big data has trouble with big problems. This is a variation on #2. Brooks points out that it’s easy to use data to crunch numbers on experimental versus control groups. But heaps of data — insane quantities of it — don’t necessarily point to solutions to complex problems. Maybe you just end up with really complex data sets. Magnification doesn’t necessarily bring clarity (though it can).
- Data favors memes over masterpieces. A new toy on the market might get a huge number of “likes,” but does that popularity blip translate into staying power with the kids who receive that toy for Christmas? Instant data measures instant success/popularity/sales … but instant isn’t always what we value in the long run.
- Data obscures values. Fans of data sets will say that numbers don’t lie. But data is structured, stripped, smushed together, and sorted according to algorithms set by humans. So even when data seems to be out of context … it always is.
Brooks concludes his essay with, “This is not to argue that big data isn’t a great tool. It’s just that, like any tool, it’s good at some things and not at others.” And it’s our job to know that. So how do we stay a step ahead of data? How do we know how it’s being organized? And how do we teach that to today’s students?