Even if you’re not a statistics whiz, everyone in public relations today needs to understand the big picture challenges of looking at data. As PR pros, part of our job is to analyze the success of strategies and tactics (our own and those of others) and make recommendations to our clients. Among other things, it’s also our responsibility to spot trends and position our clients within them.
This can be more difficult than it seems, since there’s so much false information and junk science out there – knowing the difference can be critical to success. Interestingly, the problem with false data interpretation isn’t limited to the corporate/marketing realm. In 2005, medical researchers concluded that most published research findings are likely false, and it’s also been said that the results of most medical trials can’t be replicated.
During his talk at South by Southwest (SXSW), Nate Silver, of FiveThirtyEight.com and author of The Signal and the Noise: Why So Many Predictions Fail — but Some Don't, explained that much of the problem comes from the fact that we don’t understand our own brains, which have yet to evolve to handle the amount of information we’re exposed to each day.
The instinctual shortcut we take when we have ‘too much information’ is to engage with it selectively, picking out parts we like and ignoring the remainder. But this limitation can be overcome, if we learn to control our biases.
Mo’ Data, Mo’ Problems
Here’s a mind-blowing stat: 90 percent of all the data we have access to was created in the past two years. As the signal-to-noise ratio continues to worsen, information overwhelm is a large stumbling block – there are ever-more more distractions from the signal we’re trying to find.
According to Silver, we still have cavemen brains that tend to oversimplify. This monkey brain was adapted for quick decisions based on very little data. For example, to a caveman, a rustling in the bushes can be simply translated into “uh oh, a lion is coming to get my family.”
But with all the information we’re exposed to today, Silver noted it’s hard for us to know what a real signal is, and our brains tend to overestimate random patterns. With rise of big data, we get more true signals, but also more spurious relationships.
Crisis of Science
We have a “crisis of science,” said Silver, which in large part is due to our biased interpretation of data and analysis. In short, we cherry pick the facts that tell the stories we'd like to hear.
This cherry picking is why so many projections of the 2012 elections failed. He noted that the political predictions he’s done at FiveThirtyEight.com “aren’t very hard – just averaging polls,” and the fact that that his success seems so remarkable shows how long we have to go in understanding math and science.
Rather than trying to make the data prove a hypothesis, science should be finding what the data says to us without bias, Silver explained. In fact, he makes a compelling argument that the very future of humanity and the world depends on it.
Change your mindset
Now that we have unprecedented access to big data, it’s important to recognize that it often prioritizes correlation over causality, which can lead to false conclusions. Just understanding this is a huge part of the battle – and as PR pros, it’s our job to change our attitudes and approach so we look at data, studies, survey results and the like with a healthy dose of skepticism.
Silver noted that the importance of data doesn’t render intuition useless. For example, we don’t just run job candidate resumes through a computer – an in-person interview allows us to pick up on things about what kind of employee a person will be that can’t be captured in a computer model. Use analytics as a baseline, and then merge the subjective and objective – sometimes we still need to “think slowly.”
The key is to refine your method as you go. “A lot of the time we’re hesitant to test our models and ideas, but the more you do, the better you can find the secret sauce,” Silver explained.
He also noted that in business, if you have your “stat geeks locked in a closet somewhere and the suits are somewhere else – that’s not going to go well. When you have business people speaking the language of data, that’s where you see companies grow.”
This was a mind-expanding session, and some of the pearls of wisdom from Silver were outside the topic of this post (did you know there is a link between “expert” overconfidence and false predictions?), but allow me to close with one more: “bureaucracy is the opposite of imagination.” How cool is that?
Photo credit: David Castillo Dominici