Fair enough.
One thing I learned from Ty's survey and mine done at the same time - long form answers feel really good for respondents to fill out...but are REALLY hard to parse. In Ty's survey, for example, you can see overall satisfactions and stuff (the results are on that link towards the bottom of the post), but then there's the long form answers. While there are some things that seem to be repeated (a LOT of people want Energy Drain decoupled from AF, for example), the qualified (as opposed to quantified - numbers) answers are way harder to parse. "Oh, you say you want more DPS buttons but you want to heal more...what...what does that look like?" Same with the Reddit polls we ran in conjunction with it, where it was often hard to classify people's answers when you get something like "Do you like the current state of healers? I hate how simple WHM is. I find AST is too busy. I think AF should be decoupled from Energy Drain, but I also want Miasma 2 back but I also want encounters to do more healing and I don't like Jobs being braindead but WAR is my favorite healer". There are a lot of mixed signals making it WAY harder to do anything useful with that feedback outside of identify some keywords/phrases here and there that you see repeat a lot.
...which is harder than it seems since respondents write their responses in all different ways. One might list WHM then SCH then AST then SGE, another might list SGE then WHM than AST and leave off SCH completely. Some might mention their views on DPS first, others healing first, others encounters first, and leaving off some of the others. Some might write extremely short "I hate it" while others may go into extreme length.
It's why so many surveys use multiple choice but not long-form answers, like as not. But that limits the survey quite a bit and you have to make sure to offer a lot of options so that you cover all the major bases, and you might still miss something.
.
TL;DR: Polls are hard.
.
EDIT:
To be fair, this might have more to do with something else, like trying to figure out what kinds of zones/environments people like/would like to see more of and why.
Though not ALWAYS, I GENERALLY dislike that. I think a scale from 1-5 or 0-10 is better, since it allows more nuance. "I kind of am satisfied, but am I VERY satisfied or more neutral? But there's not a neutral option. So do I answer no opinion, satisfied, or dissatisfied?" It's also why I dislike the "dislike, slightly like, somewhat like, really like" "scale" that some pollsters like to use. "Does slightly like get counted with somewhat, or with dislike? Is it three versions of like or is the slightly counted as negative? What if I only kinda like it but I DO like it and don't want to get counted as on the negative side?"
With a 1-2-3-4-5 (1 hate, 2 dislike, 3 neutral, 4 like, 5 love) or a 0-10 (0 being hate, 10 being love, 5 being neutral, and giving people 1-2-3-4 and 6-7-8-9 flavors of dislike/like to really express their views) being far better, I think.



Reply With Quote

