I think it's more accurate to say "34% of respondents lied about race. Without more details, I'm skeptical that their methodology, as presented, was very rigor

I think it's more accurate to say "34% of respondents lied about race. Without more details, I'm skeptical that their methodology, as presented, was very rigorous to be able to extrapolate to the entire population. So this is very suggestive, but I'd like to see a more tightly controlled version conducted.