I like your two framing questions. I also agree about the power of the loaded question (“Have you stopped beating your wife?”).
Now, while I am certainly trying to provoke responses in this forum, I don’t think the way I asked this question about spin forces the same discussion that a question of how low morale is or whether you’ve stopped abusing your spouse does.
Both of those questions assume an existing state (morale IS low, you DO abuse your spouse). I am asking a question that any of you can answer (and have) with a yes or a no and define a state that the question didn’t force.
Ok, I’ll stop being defensive.
Your BLS example is a good illustration of useful govt data without much frosting (low context). It’s great when that can happen.
For those parts of govt with missions that revolve around services and policies, however, low context isn’t always an option. Whether I’m a citizen or a senator, you’re probably going to have to give me context beyond pure data as to why I should care about your services or want to support/comply with your policies.
As for surveys of people (Fed. Empl. Viewpoint Survey), bring your mine-detection equipment: There is nothing just-the-facts about asking people for their feelings or opinions of things, and such survey results usually require a boatload of context.
Even asking people a factual question — do you own a lawn mower? — might not elicit a straight answer
In any case, the execution of govt services and policies should be based on some data that support the need for said services/policies, too. If nothing else, we in govt should be careful about forcing data to ask questions we are pre-disposed to answer.