We've all been asked bad survey questions, in magazines or by telephone pollsters, where it wasn't even remotely clear what they wanted to know or, worse, there was no answer that accurately reflected our knowledge or beliefs. Good scientists, who are knowledgeable about the pitfalls of survey research, put in a lot of work to avoid these problems. One way they do so is by creating validated instruments or by using instruments that have been validated by other scientists.
How are instruments validated? It can be done in any number of ways. One good method is to look for biomarkers, measurable ways that confirm someone is telling the truth. Unfortunately, not everything can be measured with a swab or a blood test, so sometimes validating a survey instrument means giving people the survey and then asking them about it. It can also mean asking questions in many different ways until you find ones that give internally consistent results.
Validation isn't a one time thing, it's an ongoing process. As culture changes, so do people's understanding of words and meanings. Even more worrisome, not just how you ask a question, but where in a survey you ask it, affects how people respond. Just as there are no simple answers in science, there are also no simple questions.