The worst thing?
Suppose there is a referendum in, I don't know, Venezuela or something; that's adequately hypothetical. Imagine they put a referendum banning gay marriage on the ballots in Venezuela, and it passes with similar levels of turnout and support. There's a guy in your office from Venezuela who's reading about it, a little bit troubled by the news, and you say, "wow, I guess 60% of Venezuelans are literally horrible terrible people."
Do you think that he is going to be most upset with the precision and accuracy present in the percentage you used, and not, say, the idea that you just casually pass judgement on what might easily be 60% of his family, extended family, friends, former co-workers, schoolmates, acquaintances, and the like? Not with the way you dismiss entirely their value as moral actors, without any further attempts to seek understanding of them as human beings, pre-emptively closing the door on learning about their culture and their values, not even pretending that you've an interest in understanding their lives as flawed, imperfect human beings, implicitly disclaiming all need to sympathize with their troubles? Not with the way that you, hailing from a background of substantial privilege in the San Francisco Bay Area, college-educated, and now happily earning six figures in a software job, feel that it's right and just to look down on these people, some of whom are fighting to live somewhere better than a shack in the woods?
https://www.math.upenn.edu/~deturck/m170/wk4/lecture/case1.h...
The worst thing here is the suggestion that there was 100% turnout, instead it was closer to 30%, so only about 20% voted for it. The amendment was found unconstitutional a short time afterwards anyway, finding that marriage is a fundamental right for everyone.