Less Wrong Study Hall—Year Two

The Less Wrong Study Hall is still going, two years after its creation. I’m trying to make last year’s survey an annual thing, so I did it again. We’ve been growing slowly but steadily. According to this year’s census, we have 60+ users. By comparison to Yvain’s Less Wrong survey, that gives us just under 4% penetration.

For anyone who isn’t familiar with the Hall: We’re a collection of (mostly) Less Wrong users who gather to work in a video chat room. The idea is not so much to collaborate on projects, as to have visible companions in the effort to get things done. There’s a thorough description here, and its description of our efforts is still more or less accurate. The room is hosted by Tinychat, and we work according to the Pomodoro Technique.

I’ll briefly reiterate our social norms for anyone who doesn’t want to read the whole prior post:

  • Say hello and ask for the current time when entering, if needed.

  • Don’t talk during pomos.

  • Do talk during breaks.

  • Talking about work is encouraged.

  • Bragging about work is encouraged.

  • Don’t turn your mic on.

  • But do turn your camera or desktop view on if you want

The tinychat room is here, and the Complice frontend is here. Either will work. The password is ‘lw’. The password exists not as a security measure but as a roadblock for random walk-ins.

Aside from our growing population, the most significant change of the last year was malcolmocean succesfully integrating the room into his Complice application. This doesn’t give us control of the room, but does give us (by which I mean him) the ability to bolt on features, including a unified pomodoro timer with audible dings. It’s achieved fixation remarkably quickly; most of our participants now use it.

Last year’s top suggestion box item was better enforcement of pomodoros; break overruns were extremely common. The public timer has rapidly and all-but-completely solved the problem. No technological enforcement was necessary, which surprised me. There was some concern that users who weren’t inclined to use Complice would be left in the dark, but that doesn’t seem to have happened. I would like to emphasize that if you are joining via Tinychat rather than Complice, it is still okay to explicitly ask for the time. We know that not everyone can see the timer, and we won’t bite you.

We do have some challenges to deal with in the upcoming year. Most of them involve our continuing dependence on Tinychat. Our Complice frontend is definitely not Tinychat-approved, and I suspect they will try to step on it if they notice. With our population growing, we have begun to bump against Tinychat’s twelve-camera limit. And, of course, as long as we’re dependent on Tinychat we can’t develop the Hall into a single unified application.

There may be some indication of light at the end of that tunnel; we’ll see. In the meantime, we do still have room for growth, and we still welcome new people.

Personal Reflections

As I write this, I find that we have recovered our original Creator from the darkness of non-usage. That made my day.

I actually use the Hall less than I did last year. I changed jobs, and my new schedule leaves me less time alone that I would normally use to join the chat. So, my comments may be less reflective than otherwise.

The numbers say that we are bigger, but it doesn’t really feel that way. We have run close to the video limit, but not very often; most of the time that I go (usually on the weekends), there are three to six cameras active. I’ve talked to a few others about it, and it sounds like the increase may be coming from more users during lower-use times than an increase in peak usage.

I mostly use the Hall for handling personal chores and writing. I occasionally program, but this is one task for which I’ve found the Hall counterproductive. Pomodoro breaks flush my mental state, screwing up the symbolic juggling that any programmer will be familiar with.

If anyone else has the same problem, but still wants to use the Hall while coding, I suggest ignoring breaks and turning the volume down so that timer dings and people talking doesn’t pull you out of hack mode. That’s what I do. In the future, it would be nice to have rooms with different pomo periods to accommodate things like this. Right now I don’t think we have enough people to fill multiple rooms, but with any luck that will change.

One of the most interesting responses in the survey below came from a former user, who used the comment box to describe why they’d left. They cited a similar issue; they needed help with larger tasks, not the chunkable ones that suit pomodoros best. They also cited cultural fit; apparently they were put off by the silly atmosphere of many of the breaks.

I want to address that specifically, because there was some concern about it last year that did not seem to hold up when the survey asked about it. Here we have someone for whom it really was an issue. So it is, indeed, a real thing. On the other hand, other respondents noted the same elements as a plus, and I myself would be sad to see it go away. I think this is something that we are just going to have to accept as an aspect of our culture that inevitably will not suit everyone.

The respondent in question seemed to agree. That said: Whoever you are, sorry you left us and thank you for telling us why.

2015 Census and Survey Results

Numbers are fun, so here there be numbers. Once again, credit to Yvain for all the questions I pilfered from his Less Wrong survey.

There were 63 respondents, up from last year’s 23. I’m unsure how much of that increase is real and how much is a selection effect. This year, the Complice room automatically advertised the survey during breaks, so it’s likely a higher percentage of users took it.

Last year I took the Google Forms statistics verbatim, but this year I couldn’t do that for Reasons. The script I used to generate these numbers is possibly the worst code I’ve written in the last decade. As far as I can tell they’re still correct, but if anyone notices something obviously wrong, please let me know.

The text of the individual questions can be found here. My comments are in brackets.

Population

Population Count

Yes 60 92.31%
I used to but don’t anymore 3 4.62%
No, I’m just here to mess with your statistics 2 3.08

[[ People who were here to screw up the stats were left out of the stats. Har. ]]

Survey History

Yes 15 23.81%
No 48 76.19%

[[ This was an attempt to get an idea of the proportion of users who stick around. 15 of last year’s 23 survey takers took this year’s survey. ]]

Technology

Web Browser

Chrome 45 71.43%
Firefox 17 26.98%
Other 1 1.59%

[[ This question was intended to provide useful information for developing a TC alternative. What’s interesting about this: The best alternative I’ve found to Tinychat, that we can host ourselves, is called Jitsimeet. It does not support Firefox yet (or rather, Firefox does not yet support the tech it uses, although it’s in the pipeline), and so I never put *that* much effort into investigating it. That our users so heavily favor Chrome suggests that it may be worth a second look. ]]

Complice

Yes 55 87.30%
No 8 12.70%

[[ Almost certainly an overestimate. Complice automatically advertised the survey during breaks, so complice users would have been disproportionately aware of it. ]]

Demographics

Age

n 61
mean 25.20
stdev 6.86
min 15
q1 22
q2 24
q3 27
max 60

Country

United States 29 46.77%
Germany 9 14.52%
United Kingdom 7 11.29%
Canada 5 8.06%
Denmark 2 3.23%
Poland 2 3.23%
Other 8 12.90%
No Answer 1 1.61%

[[ Last year Germany held a plurality of our users; most of our growth since then has been in the U.S. ]]

Race

White (non-Hispanic) 49 77.78%
Asian (East Asian) 4 6.35%
White (Hispanic) 4 6.35%
Other 6 9.52%

[[ Yay, we’re not 100% White anymore. Since my script lumps any answer that got exactly one response under “other”, I’ll note that the six responses here represent one Middle Eastern, two mixed, one Asian (Indian Subcontinent), one Aboriginal, and one “none.” ]]

Sex, Gender, Relationships

Sex

Male 44 69.84%
Female 19 30.16%

Gender

Male (cisgender) 40 64.52%
Female (cisgender) 15 24.19%
Other 7 11.29%
No Answer 1 1.61%

[[ Surprisingly close to balanced, here, especially by comparison to Less Wrong proper. This doesn’t match my experience in-room, which is more heavily male; I’m wondering if our female/​other participants don’t spend as much time in the room as the men, or if they just don’t run their cameras as much. (also interesting: Despite transgender options being listed, only one person actually used them. All the other “Others” were write-ins.) ]]

Sexual Orientation

Heterosexual 37 59.68%
Bisexual 14 22.58%
Asexual 5 8.06%
Heteroflexible 3 4.84%
Other 3 4.84%
No Answer 1 1.61%

[[ I am amused that Heteroflexible escaped the Other ghetto despite not actually being provided as an answer. I’m also not sure why Less Wrong has such an overrepresented bisexual contingent. I noticed that in Yvain’s survey and it shows up here, too. ]]

Relationship Style

Uncertain /​ no prefrence 24 39.34%
Prefer polyamorous 18 29.51%
Prefer monogamous 17 27.87%
Other 2 3.28%
No Answer 2 3.28%

Number of Current Partners

0 25 42.37%
1 25 42.37%
2 4 6.78%
3 4 6.78%
Other 1 1.69%
No Answer 4 6.78%

Relationship Goals

…not looking, but open to the possibility 39 62.90%
…and currently looking for more relationship partners 13 20.97%
…and currently not looking for more relationship partners 10 16.13%
No Answer 1 1.61%

[[ It’s funny to compare this to last year’s survey, which lacked the “not looking, but open” option. Apparently most people round that up to “looking”. ]]

Relationship Status

Single 28 45.90%
Relationship 20 32.79%
Married 9 14.75%
Other 4 6.56%
No Answer 2 3.28%

Children

0 55 91.67%
1 2 3.33%
2 3 5.00%
No Answer 3 5.00%

[[ Not pictured: One user with minus 37 children, which I unilaterally rounded up to zero. ]]

More Children

Yes 15 24.19%
No 24 38.71%
Uncertain 23 37.10%
No Answer 1 1.61%

Work and Education

Work Status

Student 38 61.29%
For-profit work 11 17.74%
Self-employed 3 4.84%
Unemployed 3 4.84%
Homemaking 2 3.23%
Academics (on the teaching side) 2 3.23%
Other 3 4.84%
No Answer 1 1.61%

[[ We’re as student-centric as ever ]]

Profession

Computers (practical: IT, programming, etc.) 17 32.69%
Mathematics 5 9.62%
Biology 5 9.62%
Neuroscience 3 5.77%
Computers (AI) 3 5.77%
Medicine 3 5.77%
Physics 3 5.77%
Engineering 3 5.77%
Statistics 2 3.85%
Other “social science” 2 3.85%
Philosophy 2 3.85%
Other 4 7.69%
No Answer 4 7.69%

Degree

High School 13 22.03%
Bachelor’s 23 38.98%
Master’s 12 20.34%
PH D. 3 5.08%
MD/​JD/​other professional degree 3 5.08%
None 5 8.47%
No Answer 1 1.69%

Less Wrong

Less Wrong Use

I don’t use Less Wrong at all. (skip the rest of this section) 2 3.17%
I lurk, but never registered an account. 9 14.29%
I’ve registered an account, but never posted. 9 14.29%
I’ve posted a comment, but never a top-level post. 18 28.57%
I’ve posted in Discussion, but not Main. 17 26.98%
I’ve posted in Main. 8 12.70%

[[ More Main posters than I ever expected. Do we get status points for this? ]]

Time in LW Community

n 61
mean 3.09
stdev 1.79
min 0.15
q1 2
q2 3
q3 4
max 8

Karma Score

n 59
mean 345.41
stdev 624.84
min 0
q1 0
q2 64
q3 384
max 2978

Meetups

No. 25 40.98%
Yes, once or a few times. 23 37.70%
Yes, regularly. 13 21.31%
No Answer 2 3.28%

Community

No. 34 55.74%
Yes, sometimes. 11 18.03%
Yes, all the time. 16 26.23%
No Answer 2 3.28%

Survey (Yvain’s)

Yes. 38 62.30%
No 23 37.70%
No Answer 2 3.28%

Less Wrong Study Hall and You

Time in Community (LWSH)

Less than a month. 15 23.81%
1 − 6 months 15 23.81%
6 − 12 months 10 15.87%
1 − 2 years 14 22.22%
Since the beginning (March 2013) 9 14.29%

[[ We had an influx of newbies after the Complice room was announced; I’m not sure how much that is represented here. ]]

Frequency

Every day 7 11.11%
Several times a week 13 20.63%
Once or twice a week 16 25.40%
Less than once a week 5 7.94%
It varies 17 26.98%
I haven’t been here long enough to form a pattern 5 7.94%

[[ While we have more users, they don’t come to the Hall as often. I think this explains why it doesn’t feel that much busier even though our population is much higher. ]]

Time in the Hall

n 62
mean 181.18
stdev 108.62
min 3
q1 120
q2 180
q3 200
max 600

[[ On the other hand, time-per-visit is pretty much the same. ]]

Usage

Academic studies 48 76.19%
Personal projects 45 71.43%
Chores/​Paperwork/​Necessities 38 60.32%
Deliberate practice (e.g. learning guitar) 15 23.81%
Work for an employer 12 19.05%
Other 2 3.17%

[[ If those percentages look funny, it’s because this question permitted multiple answers. This pair of questions and the next asked for all answers that applied, followed by the most important answer. ]]

Usage 2

Academic studies 36 57.14%
Personal projects 12 19.05%
Chores/​Paperwork/​Necessities 7 11.11%
Deliberate practice (e.g. learning guitar) 2 3.17%
Work for an employer 6 9.52%

Draw

Social reinforcement for working. 58 93.55%
Distraction reduction via group pomodoros. 52 83.87%
Camera-induced self-consciousness when working. 39 62.90%
Social punishment for not working. 9 14.52%
Other 10 16.13%
No Answer 1 1.61%

Draw 2

Social reinforcement for working. 27 45.00%
Distraction reduction via group pomodoros. 16 26.67%
Camera-induced self-consciousness when working. 12 20.00%
Social punishment for not working. 2 3.33%
Other 3 5.00%
No Answer 3 5.00%

[[ The carrot still gets more love than the stick. ]]

Camera

Yes, always. 18 29.03%
Yes, sometimes. 36 58.06%
Rarely 3 4.84%
Never 5 8.06%
No Answer 1 1.61%

[[ Camera use is somewhat less frequent, apparently because of people dropping from “always” to “sometimes.” That’s a bit disappointing. I like seeing five or six faces at once. ]]

Desktop Sharing

Yes, always. 2 3.23%
Yes, sometimes. 2 3.23%
Rarely 5 8.06%
Never 53 85.48%
No Answer 1 1.61%

Time Zone

UTC-08:00 6 10.00%
UTC-07:00 2 3.33%
UTC-06:00 2 3.33%
UTC-05:00 19 31.67%
UTC-04:00 3 5.00%
UTC+00:00 6 10.00%
UTC+01:00 15 25.00%
UTC+02:00 2 3.33%
UTC+08:00 2 3.33%
Other 3 5.00%
No Answer 3 5.00%

[[ Mostly east and west coast U.S., and central Europe. I still need to find a way to take these numbers and the temoral-habits questions and work out what UTC times are most heavily populated. Malcolm, is there any chance Complice could, say, poll the number of active users and graph it over time? It wouldn’t be perfect (not everyone uses Complice) but it would probably be pretty close, and it would save me the trouble. ]]

Temporal Habits (Weekdays)

Mornings (6am-12pm) 12 19.35%
Afternoons (12pm-5pm) 29 46.77%
Evenings (5pm-10pm) 37 59.68%
Late night/​very early morning (10pm-6am) 12 19.35%
Too variable to say 14 22.58%
I don’t use the room during the week 2 3.23%
No Answer 1 1.61%

Temporal Habits (Weekends)

Mornings (6am-12pm) 8 12.90%
Afternoons (12pm-5pm) 26 41.94%
Evenings (5pm-10pm) 26 41.94%
Late night/​very early morning (10pm-6am) 7 11.29%
Too variable to say 27 43.55%
I don’t use the room on weekends 3 4.84%
No Answer 1 1.61%

Referrals

Other comments or posts on Less Wrong 22 35.48%
The initial announcement 22 35.48%
Referred by a friend or partner 12 19.35%
Other 6 9.68%
No Answer 1 1.61%

[[ My favorite “other” referral was someone who checked the URL on tinychat entirely be coincidence, before it was passworded. ]]

Interaction

Yes, regularly. 12 19.35%
Yes, sometimes. 19 30.65%
No. 31 50.00%
No Answer 1 1.61%

Interaction 2

Yes, regularly. 7 11.29%
Yes, sometimes. 10 16.13%
I’ve met a few people in person once or twice. 11 17.74%
No. 34 54.84%
No Answer 1 1.61%

Romance

Yes. 6 9.84%
I didn’t meet them through the Hall, but they come there now. 9 14.75%
No. 46 75.41%
No Answer 2 3.28%

LWSH Efficacy

Base Akrasia

1 0 0.00%
2 5 8.20%
3 6 9.84%
4 19 31.15%
5 20 32.79%
6 10 16.39%
7 1 1.64%
No Answer 2 3.28%

Akratic Impact

1 6 10.00%
2 26 43.33%
3 18 30.00%
4 7 11.67%
5 3 5.00%
6 0 0.00%
7 0 0.00%
No Answer 3 5.00%

Base Hedonia

1 3 4.92%
2 7 11.48%
3 13 21.31%
4 16 26.23%
5 15 24.59%
6 6 9.84%
7 1 1.64%
No Answer 2 3.28%

Hedonic Impact

1 0 0.00%
2 1 1.67%
3 3 5.00%
4 10 16.67%
5 23 38.33%
6 21 35.00%
7 2 3.33%
No Answer 3 5.00%

[[ Eyeballing it, it looks like the Hall removes akradons somewhat more efficiently than it produces hedons. Which is fair given its purpose. ]]

Distractions

n 58
mean 29.29
stdev 20.57
min 1
q1 15
q2 25
q3 40
max 85

Distraction Type

Spontaneous web browsing or other computer use. 49 83.05%
Digital interruptions (email or IM) 39 66.10%
In-person interruptions (family or friends wanting attention) 37 62.71%
People talking or otherwise drawing attention in the Hall during the pomo 15 25.42%
Other 4 6.78%
No Answer 4 6.78%

[[ I am optimistic about the “People talking or otherwise drawing attention” number dropping next year, thanks to the unified pomo timer. ]]

Distraction Cause

Spontaneous web browsing or other computer use. 34 59.65%
In-person interruptions (family or friends wanting attention) 10 17.54%
Digital interruptions (email or IM) 9 15.79%
Other 4 7.02%
No Answer 6 10.53%

Overwork

n 59
mean 30.54
stdev 27.49
min 0
q1 10
q2 20
q3 50
max 100

Accomplishments

Yes 24 41.38%
No 34 58.62%
No Answer 5 8.62%

[[ “Yes” answers are down from 65% last year. :-( I’m hoping that’s because we have a lot of new users who haven’t had time to do anything big yet. ]]

Accomplishment Examples

This was a freeform question. Lots of school related answers; one of our users studied finished their PhD at Oxford. Our HPMoR translator is still translating. Some personal projects, of course (I wrote and published some fanfiction, if I may pimp myself). And the Complice frontend, unsurprisingly, was written during Hall time.

Akrasia

Akrasia

Yes 21 35.00%
No 39 65.00%
No Answer 3 5.00%

[[ Down from 50%. I was speaking with Lachouette some time ago, and commented that I thought our community was self-selected for akratic problems; mostly on the grounds that people who didn’t need help to fight their own akrasia, wouldn’t come here. She said that from her perspective it looked like we were selected for unusually productive people. Does this number mean I win or that I lose? I don’t know what percentage of the general public would answer Yes. ]]

Akrasia: Current

Yes 5 8.33%
No 55 91.67%
No Answer 3 5.00%

[[ Very slightly down. ]]

I don’t have much to say about the rest of the Akrasia questions, and I’m thinking of dropping these questions next year. At least one person suggested that the survey was too long as it is, and I think I agree. I included this section in year 1 because it was in Yvain’s survey at the time, but I’m not sure we’re getting any value out of it—by use or amusement. He doesn’t seem to be using it anymore either.

Akrasia: Illness

None 28 51.85%
Depression 13 24.07%
ADHD 5 9.26%
Autism or autism spectrum disorder 3 5.56%
Other 5 9.26%
No Answer 9 16.67%

Akrasia: Medicines 1

No 39 68.42%
Modafinil 3 5.26%
Sertraline 2 3.51%
Bupropion 2 3.51%
Other 11 19.30%
No Answer 6 10.53%

Akrasia: Medicines 1.5

No 45 83.33%
Modafinil 4 7.41%
Other 5 9.26%
No Answer 9 16.67%

Akrasia: Medicines 2

1 6 28.57%
2 5 23.81%
3 5 23.81%
4 3 14.29%
5 2 9.52%
No Answer 42 200.00%

Akrasia: Supplements 1

No 23 42.59%
Vitamin B12 3 5.56%
Melatonin 2 3.70%
Vitamin D 2 3.70%
Multivitamin 2 3.70%
Other 22 40.74%
No Answer 9 16.67%

Akrasia: Supplements 1.5

No 43 84.31%
Modafinil 2 3.92%
Other 6 11.76%
No Answer 12 23.53%

Akrasia: Supplements 2

2 8 32.00%
3 5 20.00%
1 4 16.00%
5 4 16.00%
4 4 16.00%
No Answer 38 152.00%

Akrasia: Therapy 1

No 34 62.96%
CBT 2 3.70%
talk therapy 2 3.70%
Other 16 29.63%
No Answer 9 16.67%

Akrasia: Therapy 2

3 9 42.86%
2 5 23.81%
5 3 14.29%
1 2 9.52%
4 2 9.52%
No Answer 42 200.00%

Akrasia: Meditation 1

No 22 41.51%
Mindfulness Meditation 11 20.75%
Meditation 3 5.66%
Yoga 2 3.77%
Other 15 28.30%
No Answer 10 18.87%

Akrasia: Meditation 2

2 10 30.30%
3 9 27.27%
4 8 24.24%
1 5 15.15%
Other 1 3.03%
No Answer 30 90.91%

Akrasia: Elsewhat 1

No 28 54.90%
Other 23 45.10%
No Answer 12 23.53%

Akrasia: Elsewhat 2

2 12 50.00%
1 9 37.50%
3 2 8.33%
Other 1 4.17%
No Answer 39 162.50%

Akrasia: Communication

Yes 27 49.09%
No 28 50.91%
No Answer 8 14.55%

Silliness

Tinychat Hatred

1 (Thousand burning suns) 10 19.61%
2 10 19.61%
3 14 27.45%
4 9 17.65%
5 (Emperor Palpatine) 8 15.69%
No Answer 12 23.53%

Tinychat Screams (number of)

Slightly more information than last year: The sum, mean, max, and possibly the standard deviation were all variants of infinity. My favorite answers were e^tau and the first couple hundred digits of Pi without the decimal point.

Stuffies

No 18 31.03%
Yes, but only one 11 18.97%
Yes, more than one 20 34.48%
Tons 9 15.52%
No Answer 5 8.62%

[[ Stuffies! ]]

Stuffies on Camera

Yes 27 49.09%
No 28 50.91%
No Answer 8 14.55%

Owl (Eris/​Terry/​Levi)

Yes 4 7.02%
No 39 68.42%
I don’t know what you’re talking about. 14 24.56%
No Answer 6 10.53%

[[ I think something went wrong here, because I know I’ve seen more than four users with an Eris/​Terry/​Levi. These are the owl stuffies that some LWSH residents have. They’re sort of our de facto mascot. ]]

Owl Aquisition

Yes 35 62.50%
No 17 30.36%
I already have one. 4 7.14%
No Answer 7 12.50%

[[ I am amused at the three people who answered “I don’t know what you’re talking about” to the previous question, but wanted one anyway (you can buy them here if you know how to read Dutch and live somewhere they’ll deliver). I’m even more amused by the confused soul who didn’t know what the question was talking about but still owned one. ]]

Miscellany

Suggestion/​Comment/​Question boxes

These were freeform responses. As with last year, the most common requests were “replace Tinychat” and variants of “give us some feature that requires replacing Tinychat.” We’re getting there, sort of! There was also a demand for more crocodiles, and a love letter from someone who appreciated the existence of the Owl questions.

I was going to complain here about how running surveys like this is a lot more grueling than Yvain makes it look. Then I found something special in the comment box; I don’t know who said it but it makes efforts to bring attention to this project seem more than worthwhile. I’ll just leave it right here. To all those who show up in the Hall: you are in part responsible for this.

“Joining the Study Hall is probably the literal best thing that happened to me this decade. Thank you to the people who made this place exist.”