According to Survey Monkey, if you get about 100 responses per million people in a given population, you'll get answers that are within a 10% margin of error and around 1100 people per million to get answers within a 3% margin of error.

Ideally, you'd want a pool of totally random players, while Reddit is also going to skew toward people who have something to say rather than those who have nothing to say, and having something to say usually skews toward negative. So your sample size isn't going to be as "clean" as it should be for total accuracy. Unfortunately there's nothing you can really do about that. Live Service games will usually offer regular surveys that offer in-game rewards for completing them sometimes as frequently as monthly. Genshin, for example, typically has a survey at the end of every 6 week period if not more, which is their patch cycle, and they offer the equivalent of gil in a large sum for taking the survey.

It's probably something FFXIV should do to be fair, not just on job design, but also content and events. The forums are quite limited in the information they can offer. They seem to collection information somehow, but who knows exactly how they go about doing that.

Something you've brought up before which is true is that there are many more players who have little issues with healers, or any job really, out playing the game, and getting their input is something that would be difficult to collect since they are part of the largest group of players that are basically "go-with-the-flow" type players who tend to be fairly content with most things as long as they don't find changes to be a chore or frustrating to deal with.