"What do you think?" gets you opinions. Opinions in a design review are mostly just taste with extra steps, and taste doesn't scale past the person giving it.
The question I use instead
"What would have to be true for this to be wrong?" It forces the reviewer to name an assumption instead of a preference. Half the time, naming the assumption is enough for them to answer their own objection. The other half, you've found a real gap in the evidence, which is a much more useful thing to walk out of a review with than a list of button-color opinions.
This note is short on purpose. Longer version, with examples from actual reviews, coming later.