For Better Policy, Check Your Confirmation Bias

Charlie Arlinghaus

January 7, 2014

As originally published in the New Hampshire Union Leader

Don’t believe everything you read. Present company excepted, of course. There are many things we think we know but don’t. All too often someone posts a pointed anecdote or tidbit on Facebook or Twitter. The stories often aren’t quite true, but no one bothers fact checking when it sounds true and it’s what we want to believe.

Confirmation bias is the tendency to accept, with little skepticism, things that confirm what you already believe. It’s something we all suffer from and have trouble correcting for. It is particularly endemic in the public policy process. The things I think are a good idea have no negative consequences. Statistics that show otherwise are misleading or overstate the magnitude of the problem.

Sometimes the problem comes from general perception. One good example is immigration. We know, or think we do, that Republicans are mean and want to deport everyone. Or perhaps we think that President Obama just doesn’t care about immigration laws and wants lots of immigrants, legal or illegal.Alex Nowrasteh at the Cato Institute has studied actual data. He separates deportations into interior and border deportations to measure intensity of immigration enforcement. He found that border deportations are up under President Obama compared to President Bush, and interior deportations are up a great deal. It is important to note that this is both in numbers and as a percentage of the estimated illegal or unauthorized population.For example, President Bush averaged 276,000 total deportations annually while President Obama has averaged 405,000. What this means for your worldview, I’m not sure. You can use it to attack or support the current or former President, depending to some extent on your view of immigration. Regardless of how you feel about those two people or that issue, I suspect the data are not precisely as your Facebook friends might have predicted.

In our world today, we tend to find an anecdote that supports our worldview and cease any further investigation. Anecdotes are fun and not generally used as a jumping off point for further investigation, but rather as an end in and of themselves.

I was in a meeting at the end of last year when someone decided to make a point by comparing New Hampshire to Detroit — Detroit in this case being a stand-in for the anarchical state of nature that returns when civilization collapses. I am, I’m sure, oversensitive to attacks on Detroit. I grew up and went to school in and around Detroit, still root for the Lions during their inevitable playoff losses, and like to read the Detroit News.

Nonetheless, Detroit is a useful whipping boy. We know, or think we know, that the city is a shell, a bankrupt shell, and the economy has collapsed. The city did become the largest of the 600 municipalities that have declared bankruptcy in the last 75 years. But the city restructured (that’s what bankruptcy is), eliminated $7 billion in unfunded liabilities it had no hope of ever paying, went through two years of emergency management, and has ended bankruptcy and restored normal government. The downtown is seeing billions in new investment, and visitors are impressed with the economic changes, at least downtown.

The moribund auto industry? Statistics from the Union Leader this week show that GM, the largest automaker in the world (and the lynchpin of the Detroit economy as well), had a 19 percent increase in sales, for its biggest December in seven years. Domestic auto sales are now back to pre-recession levels and 60 percent higher than the 2009 low — good for Detroit the city, great for “Detroit” the industry.

Understanding and accounting for confirmation bias in data analysis can improve the public policy debates in New Hampshire. There is broad agreement that our economy is stagnant. Job growth isn’t what it was in the 1970s, ’80s and ’90s. However, in a recent gathering of business and policy leaders, I noticed a general consensus but disagreements over specifics.

Everyone in the room agreed with about half of what everyone else said. Even businesses disagree about what businesses want; some care more about tax rates, some are obsessed with energy costs, some relocate easily, others can’t realistically move.

Controlling for your own bias will be important in solving the state’s problems. But human perception being what it is, you probably think confirmation bias is something everyone else has.