"Target audience" size for the Less Wrong sequences

byLouie9y18th Nov 201088 comments


[Note: My last thread was poorly worded in places and gave people the wrong impression that I was interested in talking about growing and shaping the Less Wrong community.  I was really hoping to talk about something a bit different.  Here's my revision with a completely redone methodology.]

How many people would invest their time to read the LW sequences if they were introduced to them?

So in other words, I’m trying to estimate the theoretical upper-bound on the number of individuals world-wide who have the ability, desire, and time to read intellectual material online and who also have at least some pre-disposition to wanting to think rationally.

I’m not trying to evangelize to unprepared, “reach” candidates who maybe, possibly would like to read parts of the sequences.  I’m just looking for likely size of the core audience who already has the ability, the time, and doesn’t need to jump through any major hoops to stomach the sequences (like deconverting from religion or radically changing their habits -- like suddenly devoting more of their time to using computers or reading.)

The reason I’m investigating this is because I want to build more rationalists.  I know some smart people whose opinions I respect (like Michael Vassar) who contend we shouldn’t spend much time trying to reach more people with the sequences.  They think the majority of people smart enough to follow the sequences and who do weird, eccentric things like “read in their spare time”, are already here.  This is my second attempt to figure this out in the last couple days, and unlike my rough 2M person figure I got with my previous, hasty analysis, this more detailed analysis leaves me with a much lower world-wide target audience of only 17,000.


Total Population
Filters Away (%)
Speaks English + Internet Access
Believes in evolution | Atheist/Agnostic
“NT” (Rational) MBTI
IQ 130+ (SD 15; US/UK-Atheist-NT 108 IQ)
30 min/day reading or on computers

Yep, that’s right.  There are basically only a few thousand relatively bright people in the world who think reason makes sense and devote at least 2% of their day to arcane activities like “reading” and "using computers".

Considering we have 6,438 Less Wrong logins created and a daily readership of around 5,500 people between logged in and anonymous readers, I now actually find it believable that we may have already reached a very large fraction of all the people in the world who we could theoretically convince to read the sequences.

This actually matters because it makes me update in favor of different, more realistic growth strategies than buying AdWords or doing SEO to try and reach the small number of people left in our current target audience.  Like translating the sequences into Chinese.  Or creating an economic disaster that leaves most of the Westerner world unemployed (kidding!).  Or waiting until Eliezer publishes his rationality book so that we can reach the vast majority of our potential, future audience who currently still reads but doesn’t have time to do anti-social, low-prestige things like “reading blogs”.

For those of you who want to consider my methodology, here’s the rationale for each step that I used to disqualify potential sequence readers:

Doesn’t Speak English or have Internet Access:  The sequences are English-only (right now) and online-only (right now).  Don’t think there’s any contention here.  This figure is the largest of the 3 figures I've found but all were around 500,000,000.

Not Atheist/Agnostic: Not being an Atheist or Agnostic is a huge warning sign.  93% of LW is atheist/agnostic for a reason.  It’s probably a combo of  1) it’s hard to stomach reading the sequences if you’re a theist, and 2) you probably don’t use thinking to guide the formation of your beliefs anyway so lessons in rationality are a complete waste of time for you.  These people really needs to have the healing power of Dawkins come into their hearts before we can help them.  Also, note that even though it wasn't mentioned in Yvain's top-level survey post, the raw data showed that around 1/3rd of LW users who gave a reason for participating on LW cite "Atheism".

Evolution denialist: If you can’t be bothered to be moved to correct beliefs about the second most obvious conclusion in the world by the mountains of evidence in favor of it, you’re effectively saying you don’t think induction or science can work at all.  These people also need to go through Dawkins before we can help them.

Not “NT” on the Myers-Briggs typology: Lots of people complain about the MBTI.  But in this case, I don’t think it matters that the MBTI isn’t cleaving reality perfectly at the joints or that these types aren’t natural categories.  I realize Jung types aren’t made of quarks and aren’t fundamental.  But I’ve also met lots of people at the Less Wrong meet-ups.  There’s an even split of E/I and P/J in our community.  But there is a uniform, overwhelmingly strong disposition towards N and T.  And we shouldn’t be surprised by this at all.  People who are S instead of N take things at face value and resist using induction or intuition to extend their reasoning.  These people can guess the teacher’s password, but they're not doing the same thing that you call "thinking".  And if you’re not a T (Thinking), then that means you’re F (Feeling).  And if you’re using feelings to chose beliefs in lieu of thinking, there’s nothing we can do for you -- you’re permanently disqualified from enjoying the blessings of rationality.  Note:  I looked hard to see if I could find data suggesting that being NT and being Atheist correlated because I didn’t want to “double subtract” out the same people twice.  It turns out several studies have looked for this correlation with thousands of participants... and it doesn’t exist.

Lower than IQ 130: Another non-natural category that people like to argue about.  Plus, this feels super elitist, right?  Excluding people just because they're "not smart enough". But it’s really not asking that much when you consider that IQ 100 means you’re buying lottery tickets, installing malware on your computer, and spending most of your free time watching TV.  Those aren’t the “stupid people” who are way down on the other side of the Gaussian -- that’s what a normal 90 - 110 IQ looks like.  Real stupid is so non-functional that you never even see it... probably because you don’t hang out in prisons, asylums and homeless shelters.  Really.  And 130 isn’t all that “special“ once you find yourself being a white (+6IQ) college graduate (+5IQ) atheist (+4IQ) who's ”NT” on Myers-Briggs (+5IQ).  In Yvain’s survey, the average IQ on LW was 145.88.  And only 4 out of 68 LWers reported IQs below 130... the lowest being 120.  I find it inconceivable that EVERYONE lied on this survey.  I also find it highly unlikely that only the top 1/2 reported.  But even if everyone who didn’t report was as low as the lowest IQ reported by anyone on Less Wrong, the average IQ would still be over 130.  Note:   I took the IQ boost from being atheist and being MBTI-“N” into account when figuring out the proportion of 130+ IQ conditional on the other traits already being factored in.

Having no free time: So you speak English, you don’t hate science, you don’t hate reason, and you’re somewhat bright.  Seem like you’re a natural part of our target audience, right?  Nope... wrong!  There’s at least one more big hurdle: Having some free time.  Most people who are already awesome enough to have passed through all these filters are winning so hard at life (by American standards of success) that they are wayyy too busy to do boring, anti-social & low-prestige tasks like reading online forums in their spare time (which they don’t have much of).  In fact, it’s kind of like how knowing a bit about biases can hurt you and make you even more biased.  Being a bit rational can skyrocket you to such a high level of narrowly-defined American-style "success" that you become a constantly-busy, middle-class wage-slave who zaps away all your free time in exchange for a mortgage and a car payment. Nice job buddy. Thanks for increasing my GDP epsilon%... now you are left with whatever rationality you started out with minus the effects of your bias dragging you back down to average over the ensuing years.  The only ways I see out of this dilemma are 1) being in a relatively unstructured period of your life (ie, unemployed, college student, semi-retired, etc) or 2) having a completely broken motivation system which keeps you in a perpetually unstructured life against your will (akrasia) or perhaps 3) being a full-time computer professional who can multi-task and pass off reading online during your work day as actually working.  That said, if you're unlucky enough to have a full-time job or you’re married with children, you’ve already fallen out of the population of people who read or use computers at least 30 minutes / day.  This is because having a spouse cuts your time spent reading and using computers in half.  Having children cuts reading in half and reduces computer usage by 1/3rd.  And having a job similarly cuts both reading and computer usage in half.  Unfortunately, most people suffer from several of these afflictions.  I can’t find data that’s conditional on being an IQ 130+ Atheist but my educated guess is employment is probably much better than average due to being so much more capable and I’d speculate that relationships and children are about the same or perhaps a touch lower.  All things equal, I think applying statistics from the general US civilian population and extrapolating is an acceptable approximation in this situation even if it likely overestimates the number of people who truly have 30 minutes of free time / day (the average amount of time needed just to read LW according to Yvain’s survey).  83% of people are employed full-time so they’re gone.  Of the remaining 17% who are unemployed, 10% of the men and 50% of the women are married and have children so that’s another 5.1% off the top level leaving only 11.9% of people.  Of that 11.9% left, the AVERAGE person has 1 hour they spend reading and ”Playing games and computer use for leisure“.  Let’s be optimistic and assume they somehow devote half of their entire leisure budget to reading Less Wrong, that still only leaves 5.95%.  Note: These numbers are a bit rough.  If someone wants to go through the micro-data files of the US Time Use Survey for me and count the exact number of people who do more than 1 hour of "reading" and "Playing games and computer use for leisure", I welcome this help.



Anyone have thoughtful feedback on refinements or additional filters I could add to this?  Do you know of better sources of statistics for any of the things I cite?  And most importantly, do you have new, creative outreach strategies we could use now that we know this?