I don't understand this weird line of thinking. Don't do positive X because some people might respond to it in a weird way and do neutral Y. Cattling statistic

I don't understand this weird line of thinking. Don't do positive X because some people might respond to it in a weird way and do neutral Y. Cattling statistics like this is silly. If you showed fat people images of obesity and said that they'd be healthier if they lost their weight, you would also cattle the statistics and look at how many fat people decided to suicide versus how many took the advice (or didn't)? What a dumb way to construct advice.