The low number can be a product of how the poll questions are asked. If you simply ask “Do you see White Nationalism” you may elicit a low response. However if you precede the question with examples of White Nationalism (have you seen swastikas in the barracks, have you heard people joking about lynchings, do people speak about the holocaust as being fake) then ask the poll questions, this would be more likely to elicit a higher response.
Having taken way to many surveys for the Army they tend to be simplistic (and mandatory, and a chore) . Not a lot of conjoint analysis going on.