welian
#BlackLivesMatter
And it was a good thing to say! Thoughtful and relevant to the discussion!
Gender & Genre (Survey inside!)
- Thread starter welian
- Start date
Goddess Silv
Goddess of Silver Light
I do agree with Ghost in a lot of things. And going back to a bunch of my studies (I'm a social worker, yo), it really seems the case that when anonymity comes up, people are a lot more open than if they were in public. Or when it is super private, people open up and drop the whole mask of public roles. So the whole idea that men like x thing and women like x thing is dropped when a person can be open without fear of backlash.
Tangent, but it is why privacy laws in fields like I am in are super important unless there is a risk of harm. If you have someone spilling all their little secrets at you, it means that they trust you a lot. And I think the internet helps with some people in opening up as well and exploring their feelings and ideas from the safety of their own home. No one knows who they really are and they are not likely to know them irl so they can drop the whole 'Well I need to be like this so I can be acceptable' role and take up a different one. Maybe even completely changing their identity to something they feel like is more them. Which I believe that roleplaying helps with that. They can objectively use their character(s) in a safe way and explore themselves with it without harming no one.
Course this is a bit of a double edge sword in that a person can open up with harmful attitudes to others but, keeping it to roleplaying, a person is able to look at a different way of thinking and writing easier than if they were being watched by their rl friends than people on a forum that don't actually know how they act, or are, irl.
Honestly, it's something I would like to study more in-depth about. How writing and roleplaying online can help a person can explore themselves and their roles in society (including gender and identity roles) but that would require obscene amounts of data and people willing to have personal questions asked about them xD
Still, it is interesting, in the end.
Tangent, but it is why privacy laws in fields like I am in are super important unless there is a risk of harm. If you have someone spilling all their little secrets at you, it means that they trust you a lot. And I think the internet helps with some people in opening up as well and exploring their feelings and ideas from the safety of their own home. No one knows who they really are and they are not likely to know them irl so they can drop the whole 'Well I need to be like this so I can be acceptable' role and take up a different one. Maybe even completely changing their identity to something they feel like is more them. Which I believe that roleplaying helps with that. They can objectively use their character(s) in a safe way and explore themselves with it without harming no one.
Course this is a bit of a double edge sword in that a person can open up with harmful attitudes to others but, keeping it to roleplaying, a person is able to look at a different way of thinking and writing easier than if they were being watched by their rl friends than people on a forum that don't actually know how they act, or are, irl.
Honestly, it's something I would like to study more in-depth about. How writing and roleplaying online can help a person can explore themselves and their roles in society (including gender and identity roles) but that would require obscene amounts of data and people willing to have personal questions asked about them xD
Still, it is interesting, in the end.
welian
#BlackLivesMatter
[QUOTE="Goddess Silv]How writing and roleplaying online can help a person can explore themselves and their roles in society (including gender and identity roles)
[/QUOTE]
I've heard a number of anecdotes about people who get attached to playing characters that allow them to express something they don't otherwise express. I think it's why I enjoy playing arrogant pricks and mid-level villains. I get to be a bitch without hurting real people. xD
[/QUOTE]
I've heard a number of anecdotes about people who get attached to playing characters that allow them to express something they don't otherwise express. I think it's why I enjoy playing arrogant pricks and mid-level villains. I get to be a bitch without hurting real people. xD
Zerataku
Junior Member
I'm surprised that more females weren't interested in romance.
welian
#BlackLivesMatter
Huh, you're right. I guess romance is overrated or something.HawkPrincess said:
@Shining Lotus Sage
Thirty-five responses. How about that science, then?
Shining Lotus Sage
Avatar of Vanity
Statistics, at least. So is the spreadsheet you have on the OP the whole data set?
welian
#BlackLivesMatter
[QUOTE="Shining Lotus Sage]Statistics, at least. So is the spreadsheet you have on the OP the whole data set?
[/QUOTE]
Yep! That's the automatic sheet google creates when a survey is submitted. I've set conditional formatting on it for the font colors.
[/QUOTE]
Yep! That's the automatic sheet google creates when a survey is submitted. I've set conditional formatting on it for the font colors.
Shining Lotus Sage
Avatar of Vanity
Okay. Give me a while. I'll gawp at it over the weekend.
Shining Lotus Sage
Avatar of Vanity
Did I say ANOVA? That's for scalars. I should have said Chi-squared and crosstabs. I actually dug into SagePub and learned the applications of a couple new statistical methods, just to make sure I couldn't wring more meaning out of what was available. All stuff I'll use in the line of work, but nothing I'd had a need for yet.
Significant Correlations among genres:
[1]- in this sample. I have no way of estimating what larger group it might be representative of, and I doubt even the whole site.
- A person's preferred roleplay category predicts whether they'll favor the Action genre: Sci-fi is a strong yes, Fantasy is a moderate no. That's it for category/genre pairings, though.
- I can tell you with a lot of confidence, now, that gender has bupkis to do with favored genre[1]. I'm one of the results predicting that it would do so, so honestly, I'm as surprised as the rest of that set. But the predictive power of either variable on the other was generally less than .2 (of 1.0), and very frequently 0. No dice on gender and genre.
- Gender also has nothing to do with preferred roleplay category. Seriously, 10k-sample Monte Carlo significance of .213 here. Nothing.
- Gender could potentially be correlated with believing that gender is correlated with favorite genres. The significance here gets as tight as .095, which might be the best we'll get from this data set. So congrats, boys. You were mostly right on this one, and possibly for non-random reasons! Of course, the effect of preferred category is just as significant. So as often as boys will get this right, Modern or Realistic fans will get it wrong. And hey, I put my wrong answer in Fantasy, so I tried to help you out.
Significant Correlations among genres:
- Action is opposed to Mystery
- Crime is aligned with Horror (but also this is on the basis of a single person, so remember, none of this is very meaningful)
[1]- in this sample. I have no way of estimating what larger group it might be representative of, and I doubt even the whole site.
Bone2pick
Minority of One
[QUOTE="Shining Lotus Sage]
[1]- in this sample. I have no way of estimating what larger group it might be representative of, and I doubt even the whole site.
[/QUOTE]
How much confidence do you have in self-reporting as a method for determining genre preference?
- I can tell you with a lot of confidence, now, that gender has bupkis to do with favored genre[1]...
[1]- in this sample. I have no way of estimating what larger group it might be representative of, and I doubt even the whole site.
[/QUOTE]
How much confidence do you have in self-reporting as a method for determining genre preference?
Shining Lotus Sage
Avatar of Vanity
I see no reason for people to hide their preferred genres, or plausible reason they might not know. It's not a particularly stigmatized or sensitive topic. Definitely not a representative sample, and definitely has self-selection bias, of course. Do you have a reason to doubt self-report of this item?
welian
#BlackLivesMatter
//listens eagerly
I'm just really excited to have this sample size at all. I thought maybe ten people would answer, tops!
I'm just really excited to have this sample size at all. I thought maybe ten people would answer, tops!
Girouette
Lizard knight Overlord
Is mystery comedy really that not liked? I need to fix that!
Bone2pick
Minority of One
I pretty much doubt the accuracy of all self-reporting when it comes to preference. From what I've seen folks don't describe their preferences very accurately. Whether it's music genre, film genre, literary fiction genre, foods they prefer, romantic qualities they're drawn to, or habits that they indulge in — I witness huge inconsistencies in what they do vs what they say.
Shining Lotus Sage
Avatar of Vanity
Then it sounds as though you have already reached your own conclusions about whether the questions were legitimate to ask, ones that my analysis does not address. I'm not sure what would persuade you that, for instance, I am most likely to sign up for a game that features Horror and Drama themes, and I don't care to go looking for something. Your relationship with the principle of parsimony is none of my business.
welian
#BlackLivesMatter
Befooooore this possibly turns into a argument, I'd like both your opinions on something.
Were I to make a second version of this survey, what would you have me change about it in order to include more/more specific information?
Were I to make a second version of this survey, what would you have me change about it in order to include more/more specific information?
Bone2pick
Minority of One
@Shining Lotus Sage Your roleplay participation history would persuade me of your genre preference better than any completed survey. Actions speak louder and truer than words from my experience.
I can't comment on your last sentence because I don't understand it.
@welian Please don't adjust on my account, I just wanted to share my perspective on the accuracy of preference surveys. I find them interesting, but largely inconsistent with reality.
I can't comment on your last sentence because I don't understand it.
@welian Please don't adjust on my account, I just wanted to share my perspective on the accuracy of preference surveys. I find them interesting, but largely inconsistent with reality.
Shining Lotus Sage
Avatar of Vanity
I promise not to go any further. It is really, really none of my business.
If you wanted to make this survey more rigorous, you'd first have to decide on a sampling frame. What is the population you want information about? I don't think that any clever design will let you go broader than "The RPNation user base", so let's look at that.
The gold standard for sampling that population would involve a RNG, a list of user ids to randomize through, multiple waves of survey invitations, and some way to prevent multiple submissions and track who has not responded. All of this, preferably, done through a layer of obfuscation that ensures you don't know who anyone is beyond some reference ID.
I'm really not sure what chunks you might use, for a less rigorous sampling paradigm, so you'll probably wind up with the same self-selection bias you have here. There's just no way to know whether the people who volunteer their information are somehow atypical, relative to the rest of the population of interest.
For your survey design, it would be useful to introduce finer gradations, if you can do it without manufacturing specificity. Asking people to rank the categories instead of selecting just one would provide more details to work with. Rating each genre on a scale of 1-4, from strong dislike to strong like, would be better than a 9C2 of simple upvotes. Basic demographic information like gender and age, possibly with age grouped into several-year spans to smooth the data, since analysis will probably involve binning it anyway. And finally, if you are using an obfuscation method that prevents you and your team from identifying any individual responses, include a statement at the beginning assuring potential respondents of their confidentiality. It helps to minimize the weaknesses of self-reported data, which is valuable no matter how banal the information you gather may be.
As for whether you can ever use self-report for anything ever, I'd need to dig into the literature for longer than the free time I'm donating here to make my case. It is definitely a topic that can be and has been challenged, and the results are generally that the correlation between reported opinions and behavior that drives from those opinions depends on the topic and the work put in to minimize several well-documented biases that surveys are particularly susceptible to.
Explanation of any differences observed is even more situational. For instance, I might love horror and drama to death, but have done most of my searches for games during times when mostly romance and comedy were on offer. Then, a review of my playing history might differ wildly from my stated preference, and which one you value, as a researcher, depends on what you're using that information for. If you're doing ad-targeting research, you don't care what I want; you may want to know which pages to show buy shares on. My acted preferences are more meaningful there. But if you want to guage what sort of genre to create a game in to make sure it gets attention, you're probably better served going with what I say I enjoy, *especially* if I haven't been getting much of that.
But if I felt that liking comedy was for shallow, immature people(i do not), then you're right, I'd suppress my own appreciation for comedy in my responses, whatever the "real" level of that might be. But again, which reality do you care about? If I have that opinion, I probably won't let myself play in comedy games, either.
So generally, if you're going to question the results of a survey, it's most useful to have a reason that is specific to the survey you doubt. A blanket distrust of what people say about themselves is not particularly useful in market research, social sciences, or really anything but reading the newspaper. Because wow, are those surveys, on average, designed to get a particular result.
If you wanted to make this survey more rigorous, you'd first have to decide on a sampling frame. What is the population you want information about? I don't think that any clever design will let you go broader than "The RPNation user base", so let's look at that.
The gold standard for sampling that population would involve a RNG, a list of user ids to randomize through, multiple waves of survey invitations, and some way to prevent multiple submissions and track who has not responded. All of this, preferably, done through a layer of obfuscation that ensures you don't know who anyone is beyond some reference ID.
I'm really not sure what chunks you might use, for a less rigorous sampling paradigm, so you'll probably wind up with the same self-selection bias you have here. There's just no way to know whether the people who volunteer their information are somehow atypical, relative to the rest of the population of interest.
For your survey design, it would be useful to introduce finer gradations, if you can do it without manufacturing specificity. Asking people to rank the categories instead of selecting just one would provide more details to work with. Rating each genre on a scale of 1-4, from strong dislike to strong like, would be better than a 9C2 of simple upvotes. Basic demographic information like gender and age, possibly with age grouped into several-year spans to smooth the data, since analysis will probably involve binning it anyway. And finally, if you are using an obfuscation method that prevents you and your team from identifying any individual responses, include a statement at the beginning assuring potential respondents of their confidentiality. It helps to minimize the weaknesses of self-reported data, which is valuable no matter how banal the information you gather may be.
I probably misread your tone, then. My sincere opinion about self-reported preferences is that you are limited by two things: how well people know themselves, and how comfortable they are acknowledging a given detail in the survey setting. Recent research on using Internet administration to capture genuinely delicate things like sexual behaviors shows that it is at least as accurate, valid, and reliable as surveys conducted in a controlled environment.Bone2pick said:
As for whether you can ever use self-report for anything ever, I'd need to dig into the literature for longer than the free time I'm donating here to make my case. It is definitely a topic that can be and has been challenged, and the results are generally that the correlation between reported opinions and behavior that drives from those opinions depends on the topic and the work put in to minimize several well-documented biases that surveys are particularly susceptible to.
Explanation of any differences observed is even more situational. For instance, I might love horror and drama to death, but have done most of my searches for games during times when mostly romance and comedy were on offer. Then, a review of my playing history might differ wildly from my stated preference, and which one you value, as a researcher, depends on what you're using that information for. If you're doing ad-targeting research, you don't care what I want; you may want to know which pages to show buy shares on. My acted preferences are more meaningful there. But if you want to guage what sort of genre to create a game in to make sure it gets attention, you're probably better served going with what I say I enjoy, *especially* if I haven't been getting much of that.
But if I felt that liking comedy was for shallow, immature people(i do not), then you're right, I'd suppress my own appreciation for comedy in my responses, whatever the "real" level of that might be. But again, which reality do you care about? If I have that opinion, I probably won't let myself play in comedy games, either.
So generally, if you're going to question the results of a survey, it's most useful to have a reason that is specific to the survey you doubt. A blanket distrust of what people say about themselves is not particularly useful in market research, social sciences, or really anything but reading the newspaper. Because wow, are those surveys, on average, designed to get a particular result.
welian
#BlackLivesMatter
I was inspired by White Masquerade's poll on favorite genre, in which there was a discussion about where people roleplay things because it's what the love the most, or because it's "good enough" and what's available. Thus, although post history can be a really useful tool, I don't think it's what I'm interested in measuring here.
I believed @Grey mentioned wanting an alternate method of choosing genre. i don't believe I can have a "sort by rank" function on Google Forms, but I could easily split it into two questions, asking for favorite and second favorite respectively.
I'm also thinking I could split the survey based on gender, have three instances of the genre-related questions based on what gender people choose.
I could ask for username at the end as a loose kind of confirmation, and then hide the column in the spreadsheet, but I rather like having an anonymous survey. Especially since I plan to make the next one non-RPN specific and then throw it on Reddit.
I believed @Grey mentioned wanting an alternate method of choosing genre. i don't believe I can have a "sort by rank" function on Google Forms, but I could easily split it into two questions, asking for favorite and second favorite respectively.
I'm also thinking I could split the survey based on gender, have three instances of the genre-related questions based on what gender people choose.
I could ask for username at the end as a loose kind of confirmation, and then hide the column in the spreadsheet, but I rather like having an anonymous survey. Especially since I plan to make the next one non-RPN specific and then throw it on Reddit.
Shining Lotus Sage
Avatar of Vanity
And also I'm sorry I assumed you were being confrontational, bone.
AlexSilverX
The Blind Idiot God
Personally, I'm an "Action Adventure" type, because I feel Action and Adventure are settings which can be sprinkled with the others and are easier to drive forward. But that's personal opinion, not fact.
I'm male, and about (this is inaccurate as it is an estimation) 66% of my characters are male. The reason for this is that - as an RPer and writer - I've observed it's easier for me to write a character I relate to on some level. I'm not the type to shy away from a challenge, but the less the character resembles me, the more trouble I have getting into the right mindset to write them since I'm a bit of a "method writer" (and so my posting rate suffers, which may piss off other RPers and I want to avoid when possible). If they don't think like me or act like me, they need to at least resemble me to a degree (being male is usually the simplest way in which I get them to do that) for there to be a certain ease in RPing them. I usually try to keep characters in the 2-5 steps from Alex zone (i.e.: They are distinct from me in at least 2-5 ways, minor or major, in appearance, personality, and ideals, for a total of 6-15 differences.)
So if I ever play:
A smaller than average, pudgy, dark-skinned, non-cis female with long, light, curly hair either younger than 14 or older than 40, with a fondness for brightly colored or otherwise light clothing who's extroverted and outdoorsy with an inferiority complex, is an anti-smoker and drinker with a very organized mindset, strong ideas on what is right and wrong (as opposed to pragmatic), is opinionated and bigoted, has a fondness for small children and animals, a tendency for over sharing, is coy (as opposed to flirtatious), enjoys consistency in her day to day life, smiles easily but doesn't enjoy dark or otherwise "tasteless" humor, has some degree of hyperactivity and is attracted to rugged, burly men. Befriends and trusts people easily and implicitly. Doesn't enjoy reading, writing, videogames or horror movies but loves playing pool, basketball and going out clubbing, has a fear of crows, lightning, snakes and/or graveyards, dislikes "violent" sports, is very narrow-minded in terms of "relationship norms", is fervently religious (like a good Christian should be), a third-wave feminist (read: feminist extremist, I don't actually have anything against "sane" feminism), and bares dreams of- Well, "baring" many children in the future... Oh, and dismisses people for being outside the norm or otherwise disagreeing with her...
Odds are: A) I'm off my meds. B) Someone needs to take away my scotch C) I'm really testing the boundaries of what I refer to as my "character comfort zone".
Either way, I apologize in advance for my lack of activity in whatever RP I attempt this in as chances are, posting for that character will be really difficult for me. ._.'
EDIT: For those curious, the character depicted above differs from my personality, appearance and personal views in 38 different ways, making it more than double the number of differences I am typically used to.
I'm male, and about (this is inaccurate as it is an estimation) 66% of my characters are male. The reason for this is that - as an RPer and writer - I've observed it's easier for me to write a character I relate to on some level. I'm not the type to shy away from a challenge, but the less the character resembles me, the more trouble I have getting into the right mindset to write them since I'm a bit of a "method writer" (and so my posting rate suffers, which may piss off other RPers and I want to avoid when possible). If they don't think like me or act like me, they need to at least resemble me to a degree (being male is usually the simplest way in which I get them to do that) for there to be a certain ease in RPing them. I usually try to keep characters in the 2-5 steps from Alex zone (i.e.: They are distinct from me in at least 2-5 ways, minor or major, in appearance, personality, and ideals, for a total of 6-15 differences.)
So if I ever play:
A smaller than average, pudgy, dark-skinned, non-cis female with long, light, curly hair either younger than 14 or older than 40, with a fondness for brightly colored or otherwise light clothing who's extroverted and outdoorsy with an inferiority complex, is an anti-smoker and drinker with a very organized mindset, strong ideas on what is right and wrong (as opposed to pragmatic), is opinionated and bigoted, has a fondness for small children and animals, a tendency for over sharing, is coy (as opposed to flirtatious), enjoys consistency in her day to day life, smiles easily but doesn't enjoy dark or otherwise "tasteless" humor, has some degree of hyperactivity and is attracted to rugged, burly men. Befriends and trusts people easily and implicitly. Doesn't enjoy reading, writing, videogames or horror movies but loves playing pool, basketball and going out clubbing, has a fear of crows, lightning, snakes and/or graveyards, dislikes "violent" sports, is very narrow-minded in terms of "relationship norms", is fervently religious (like a good Christian should be), a third-wave feminist (read: feminist extremist, I don't actually have anything against "sane" feminism), and bares dreams of- Well, "baring" many children in the future... Oh, and dismisses people for being outside the norm or otherwise disagreeing with her...
Odds are: A) I'm off my meds. B) Someone needs to take away my scotch C) I'm really testing the boundaries of what I refer to as my "character comfort zone".
Either way, I apologize in advance for my lack of activity in whatever RP I attempt this in as chances are, posting for that character will be really difficult for me. ._.'
EDIT: For those curious, the character depicted above differs from my personality, appearance and personal views in 38 different ways, making it more than double the number of differences I am typically used to.
Last edited by a moderator:
