The Perils of Measuring Trust

 

The desire to measure trust is busting out all over. Some of it is due to management myths (“you can’t manage it if you can’t measure it”), and some of it is due to natural curiosity.

Do People Trust the Government More Under Republicans?

A great example is last Friday’s op-Ed in the New York Times, Imbalance of Trust, by Charles M. Blow. 

Says Mr. Blow:

…Americans seem to trust the government substantially more after a Republican president is elected than they do after a Democratic one is elected — at least at the outset.

Since 1976, the polls have occasionally included the following question: “How much of the time do you think you can trust the government in Washington to do what is right — just about always, most of the time, or only some of the time?”

The first poll taken in which this question was asked after Ronald Reagan assumed office found that 51 percent trusted the government in Washington to do the right thing just about always or most of the time. For George H.W. Bush, it was 44 percent, and for George W. Bush it was 55 percent.

Now compare that with the Democrats. In Jimmy Carter’s first poll, it was 35 percent. In Bill Clinton’s, it was 24 percent, and for Barack Obama’s, it was only 20 percent. (It should be noted that the first poll conducted during George W. Bush’s presidency came on the heels of 9/11).

The implicit assumption Mr. Blow makes is that trust changes quickly, and that polls reflect it; that the selection of a Democrat quickly results in low trust scores, while the selection of a Republican quickly results in high trust.

Or Do Democrat Administrations Build Trust in Government?

Let’s challenge Blow’s assumption.  Let’s assume that social trust–as many academics suggest–changes much more slowly than Mr. Blow assumes.  That in fact, questions like “do you trust the government” shift over a matter of many years–not a few months.  (See, for example, Professor Eric Uslaner, whose studies suggest that many forms of social trust evolve not only over years, but over generations).

Now let’s rewrite Blow’s paragraph—same facts, different implicit assumption:

…Americans seem to trust the government substantially more after a prolonged period of Democratic leadership than they do after Republicans have held the office—and the effect even carries over into the next administration for a few months.

Since 1976, the polls have occasionally included the following question: “How much of the time do you think you can trust the government in Washington to do what is right — just about always, most of the time, or only some of the time?”

The first poll taken in which this question was asked was when Carter had taken office, after eight years of Nixon and Ford.  In that poll, only 35 percent trusted the government in Washington to do the right thing just about always or most of the time.  Carter restored trust in government; when Reagan took over, that number tested at 51%.

However, after 12 years of Reagan/Bush, when Clinton had moved into the White House, it had been driven down all the way to 24% (Reagan did, after all, preach that government itself was the problem, not the solution).  By the end of Clinton’s two terms, that number had gone back up to 44%, of which George W. Bush was the beneficiary 8 months into his first term.

But with Republican Dubya at the helm for 8 years, trust in government dropped precipitously (Iraq, Katrina, et al); so far that the score early in Obama’s term was only 20%. 

Same facts: different assumptions. Who’s right? It depends. It depends on partly on how people interpreted that question, and even moreso on how long it takes people to shift their viewpoint on that particular question.

Trust is tricky. It’s not like measuring the temperature, or even political polls. The interpretation contains a lot more art and a lot less science than most simple surveys would suggest.

Interpreter beware.

1 reply

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *