You are overstating the results of such studies or use your datapoint out of context. Just because A is on average better than B, does not mean that A is better

You are overstating the results of such studies or use your datapoint out of context. Just because A is on average better than B, does not mean that A is better for everyone. If you have a study that shows that most people do(n't) like pineapple on pizza (maybe make a blog post "is pineapple on pizza such a good idea?"), that is rather unrelated to someone likeing pineapple on pizza or not and is not claiming that all people don't like pineapple on pizza. Darkmode is a nice thing! It feels like some people hype it as the 'one true better solution to everything' or a strictly superior version, when it boils down to preferance and minor differences. In the end you probably want to have both options (if you have the capacity to support them).