In my early years as a discrimination investigator, I learned that a seemingly direct answer can be highly misleading. It was possible to see this exchange:
"Did Mary Doe discriminate against anyone on the basis of race?"
"Did Mary Doe ever treat the black employees differently than she did the white ones?"
"Oh, yes. I'd say she treated them differently."
"Can you describe an incident in which she treated them differently?"
"Well, she never gave them the better assignments. They always got the worst ones."
That's one reason why I am very suspicious of surveys. A general survey question on whether a person supports a particular program, for example, might get a negative response while failing to ascertain whether the person believes there is any alternative to the current policy or whether the person favors an increase in the program's funding or resources. Asking a general question is inadequate. The details flesh out the real position.
If a person denies supporting Position A, but favors the combination of all of the ingredients of Position A, then we can conclude that the person favors Position A. In order to sort that out, we need to go into the details.
Consider that the next time you read an employee attitude survey that appears to indicate that all is well or terrible. Far too many surveys ask broad questions that fail to go beneath the surface.