Recalibration, findings

[solarbird]

The problem with self-selecting surveys are that they come pre-loaded with a substantial bias that you may and may not be able to analyse, in part or in entirety. In the case of the survey I posted on the 25th, there are at least two biases which seemed obvious to me even at poll creation:

1. People who participate are more likely, rather than less, to engage me than average, biasing towards approachability
2. People who participate are more likely, rather than less, to find me emotionally approachable, since those with no emotional engagement at all are unlikely to spend time on surveys like this

There were a total of 21 responses (polling software says 22, but that includes my null response set so I could see the results directly), or 7.3% of average unique post views. (286 unique IPs, discounting bots, as reported by viewing statistics).

Even given these biases, a third of those responding find me intimidating enough that they are less likely to engage with me than they would be otherwise; of those specifically finding me intimidating, that number was a clear majority at 54%. The most common reason for intimidation cited by those who feel intimidated is intelligence, by half the respondents (50%) across the two questions.

This analysis probably isn't helping. XD

Also common were a collection of words around intensity, drive, and busyness; also mentioned were creativity, variations on "high partial continual attention," and an assortment of single-occurance answers.

Perceived egotism did not include a significant finding beyond a lack of widespread perception of substantial egotism (mean 3.0, median 3.5, mode 5.0). I do worry a bit about the clarity of the scale, but that's purely hypothetical at this point.

Emotional opacity provided a clear response: of 21 respondents (presumably biased towards engagement, by self-selection in poll), 10 found me emotionally confusing or outright opaque. Discounting two who stated they didn't attempt at all, that would lead to 10 of 19, or 53%. That's disappointing, since I've worked so very hard on this, but good to know. This does not carry over as strongly to my music, however, where only 3 of 18 (discounting three respondents who specifically stated not listening), or 17%, found my music emotionally opaque or confusing.

The last question diverged sharply from all recent experience, wherein of 20 respondents, only 1 (5%) said they might ignore the email, and they noted they might also respond, depending upon their state at the time. Over the last several years, this non-response/ignoring-the-issue response has been the standard response when I've raised an issue over social treatment; I've come to expect it as mostly the norm. Of course, as was pointed out in comments, "...the people who [would] say 'ignore the mail entirely' will not respond to this." That is a confirmed finding. XD

So, of those who responded, a substantial minority find me intimidating, and around half find me emotional confusing or opaque. I hypothesise that those who did not respond (eta: on average) find me more so, not less, due to the nature of the questions and the self-selecting response set. The survey is now closed; thanks to everyone who responded!

Comments

[lyonesse.livejournal.com]

er. you can't say "those who don't respond find me more...." you can't say anything about the non-responding. the math won't have it. sorry. the essence of a self-selected survey is that it only answers questions about the answerers. you don't know what distinguishes responders from non-responders. that's the nature of the beast.

i almost didn't respond to the survey because i was feeling busy and alienated myself, and i don't find you either opaque or confusing at all.

[solarbird]

Obviously this is true for both specific individuals and cases in general, but you don't think it's a reasonable approximation that, on average, people who respond to a survey of this specific type from a specific person are more likely to respond if they are more heavily invested in that person than not?

(I mean, don't get me wrong, I wouldn't publish based on that! But do you really think that's not a reasonable assumption?)

[007zed.livejournal.com]

<3

[solarbird]

^_^

[silussa.livejournal.com]

Emotionally confusing?

Okay, I'm at a loss to understand that one. I've generally thought your emotions were fairly reasonable; perhaps a little swinging at times, but you'd hardly be the first person I've known that way (mine have done a few odd moves over time themselves)

Ah, well, no accounting for tastes and flavors, I suppose.

[trickofthedark.livejournal.com]

I hypothesise that those who did not respond (eta: on average) find me more so

I didn't respond because I was under the impression the poll was for those who have actually met you personally. I know only what I've read on your blog, and while you do indeed seem awesome, I don't think you seem scary or opaque or confusing via the internets.

Also I think the line between intimidating and impressive is very subjective, and creatives should not be worrying about coming off as impressive, and this POV was not a ticky available in your poll. ;)

[403.livejournal.com]

Yay statistics!

[lyonesse.livejournal.com]

no, i don't. see above about me, for instance, and remember that n is small.

[solarbird]

Indeed, I do remember you, and I can name at least two other supporting examples! But from previous experience, I don't think they're representative examples. But I'll keep it in mind.

[lyonesse.livejournal.com]

honestly, i think both the questionnaire design and the above statement reveal confirmation bias.

[xpioti.livejournal.com]

*blink*bloink* You had a poll up? :)

*hugs* You're not unapproachable or intimidating, you're just way too cool for words. :)

[rozasharn.livejournal.com]

Seconded. I didn't really see how a person could find you intimidating through a blog, so I figured the poll was for People Who Are Not Me.

[solarbird]

Not surprising - I mean, I'm pretty sure there's no way to get this kind of activity up to the level of actual science regardless. That's why there's no standard deviation numbers, no confidence interval, no attempt a control population, and for that matter, no use of my "science" tag. I don't have any idea how I'd try to do this as actual science and really I'd absolutely have to hire it out even if I did, because hi, experimenter bias. But this is mostly just me trying to get some sort of ideas about, you know, stuff.

[solarbird]

Of a sort. See above in the conversation with lyonesse - there are reasons I didn't post more numbers.

[lyonesse.livejournal.com]

i know. the thing i worry about is that the methods tend to reinforce existing ideas whether they are correct or not. which is fine if that's what you want, but it's not going to do more than that. if anything, it's likelier to be productive to seek disproof, but of course that's also harder.

[solarbird]

Emotionally confusing?
Hooooooo, long story. The short form is that I regularly have issues with people interpreting emotional intent in wildly divergent ways from the authorial emotional intent. When they do, the results can be explosive. (e.g., people deciding I hate them and have always secretly hated them - if you read Skin Horse, the recent Hot Monster business? That.)

This happens actually more in person than online, and I'm pretty certain it has to do with my subconscious physical emotion cues being misaligned with human subconscious physical emotion cues. I've done a lot of work to translate as well as I can, given everything, and it's reduced - but not eliminated, by any means - the explosion count. But I suspect that I'm doing so much translation at any given time that the act of translation gets in the way of communication as well.

[solarbird]

I don't get the intimidation thing either, but I wanted to ask about it, because it's come up a lot. There were a lot of options not in the poll - I didn't want to try to anticipate all options so I tried to leave as many fill-it-in-yourself options as I thought I reasonably could.

[solarbird]

I didn't really see how a person could find you intimidating through a blog, so I figured the poll was for People Who Are Not Me.
So noted; it does happen, though. And as I say below, I don't get the intimidation thing either, but I wanted to ask about it, because it's come up a lot.

[maellenkleth]

Uh, how **would** one obtain an appropriate control population, given the circumstances of a self-selecting group of respondents?

[maellenkleth]

Gah-damn, this goes right back to our conversation about experimental design at lunch yesterday. ^_^

I learned, a long time ago in fieldwork, that it **always** pays to explicitly look for outliers that, if found to be extent, would utterly disprove your working hypotheses.

In experience, one tends to find more of value by standing back and deliberately looking for 'near-misses' to the question at hand.

[solarbird]

Yes, that being my point.

[solarbird]

Aw, thanks. *^_^* I'd like to think people thought that more often! But mostly I feel like people find me either uninteresting or too much work to be worth whatever reward they might get out of interacting with me. (I've had a lot of discussions of the transactional nature of human relationships with my therapist.)

[silussa.livejournal.com]

Given that I have female friends that I have given advice to on understanding OTHER females they deal with, plus have spent a long time studying people, it may well be that my reaction is far from typical.