When President Donald Trump and his surrogates said the media lied about how many people were at his inauguration, it was easy to tell our readers it was absurd. There were pictures that showed he was wrong — and virtually everyone who attended both events could attest to it.
But here’s the potential future scenario that worries me: The Bureau of Labor Statistics announces its monthly data showing unemployed rates — and Trump and his surrogates say it’s wrong.
They say that it’s based on a poll — and given how wrong they were during the election, these are also probably wrong.
They accuse the Bureau of Labor Statistics of using a deceitful way of calculating unemployment, since they only consider you unemployed if you’ve looked for a job in the past four weeks.
They accuse the Census Bureau, who conducts the survey it’s based on, of only talking to 60,000 households, which couldn’t possibly be representative of the entire US — especially not rural areas.
Oh, and here’s the kicker, they say: Did you know the numbers are “adjusted” based on which month it is? That doesn’t change whether or not a farm worker has a job or not, but we’re still saying he does.
Then they accuse the media of running with these numbers, which are misleading — and doesn’t show the full effect of how much President Trump’s policies have created jobs.
If you think this scenario is farfetched, Trump has already put the unemployment rate into question. Early in the primaries, he said the “real” unemployment rate was 42 percent. Then he said BLS’s reported rate of 5 percent was “one of the biggest hoaxes in American modern politics.”
Much of the way we judge this administration will be based on data. It will be based on unemployment rates, on economic growth, on uninsured rates, on crime rates. The best thing about data is that we’ve created smart ways to measure things that humans can’t see with their eyes. It helps us see invisible, but very important, things.
Now let’s get something straight: BLS does a good job at calculating unemployment rates. It talks to more than enough people. It talks to sample of households that are representative of the US population. And since unemployment is usually higher in winter months because farm workers aren’t working then, they adjust the data so that we can compare how unemployment has changed month-to-month. They are transparent about how they calculate this — and if you don’t like this measure, they also report what percentage of people are in the labor force. But this includes everyone who is 16 and over, including retired folks, stay-at-home parents, students, and others who have a good reason not to be working.
But that doesn’t change one very important thing: This is still something you can’t see with your naked eye. That’s why data people spend so much time devising smart ways to measure these things — and why people like me spend so much time finding the most honest way to visualize these things.
Imagine another scenario: Trump and his surrogates call into question the CDC’s data showing the number of people without health insurance, which is also based on a poll and uses these scientifically accepted tools to get a good estimate.
As absurd as it is to question a crowd size, it shows that Trump and his communications team are willing to ignore visible data. In this case, there were pictures; there were witnesses.
But in so many other cases, we’ll be wielding a spreadsheet, a chart, and a methodology document. I’ve tried to convince skeptical people about the efficacy of a dataset — about why this is the best way to measure something invisible, and why I’m very confident in its accuracy, but why I can’t say 100 percent accurate. With this administration, we’re learning that the small seam between 100 and 99-point-whatever is worrisome.