YC Demographics

http://blog.ycombinator.com/yc-demographics

by austenallred • 12 years ago

77 66 12 years ago

blog.ycombinator.com

minimaxir3• 12 years ago

How large is the random sample? Since that you're asserting that YC does not discriminate because the actual proportion of female entrepreneurs funded is more than the number that applied, the conclusion is open to variance.

EDIT: Originally incorrectly said "sampling bias."

sama5• 12 years ago

for gender, it's the entire applicant pool (we ask for gender so that's easy to compute.)

for race, it's a bit harder. we don't ask for it, but we looked through hundreds of videos and made our best guess (it's also why we left the number as approximate, as of course there are cases where we got it wrong.)

minimaxir3• 12 years ago

Ah, that clears up the sampling bias.

I might have asked the wrong question. What I meant was "how many people are 5% of YC applicants" that were used to calculate a representative proportion? Let's say 2000 companies apply each with 2-3 founders, so 4000-6000 founders total. 5% of that is 200-300, which can result in a reasonable margin of error for the conclusion proportion but may not be perfect.

api• 12 years ago

Any sample size this small is open to sampling bias.

cronin101• 12 years ago

Small nitpick: What you are talking about is not sampling bias, it is something else.

You are alluding to how confidence in the presented results may be significantly lowered due to a high variance - associated with small sample sizes (it is a statistically significant possibility that you obtained the outcome through chance and not because it is representative)

Sampling bias is another thing altogether, when the method of taking measurements will exclude some of the population. For example, estimating ratio by taking note of the presumed gender of people leaving a particular room - not realising that it is in fact the female toilet.

Nonetheless, it is always a good idea to query the sample size when results are presented in such a manner :-)

minimaxir3• 12 years ago

You're correct. I double-checked my terminology now. :)

_pius2• 12 years ago

Great to see this data, thanks for releasing it.

williamstein• 12 years ago

I wonder what the age distribution is? (Not just gender and race...)

DodgyEggplant3• 12 years ago

Yes. Age is a fundamental figure for any demographics. Gender, ethnic group, and age. The fact that age is totally ignored is a huge blind spot.

bkessler100• 12 years ago

Exactly. They ask for age in their application and not gender or race.

swampthing4• 12 years ago

Not a YC partner, but I help review apps... when I look at age, it's not so that I can discriminate based on it, it's so that I can see how impressive a person's accomplishments are in light of how much time they've had to achieve them.

TheMakeA• 12 years ago

Small correction: The W15 applications did ask for gender.

sama5• 12 years ago

we've talked about this before. we fund founders of all ages above 18 (though i think the oldest founder in the current batch is 66). we haven't calculated it yet but the median age for this batch is probably in the late 20s, with most founders in their 20s, 30s, and 40s. the last time i looked, we'd funded many more founders in their 40s than in their teens.

there are a lot of promising 16 and 17 year olds that apply that we can't fund, though i think it's probably a good thing for them that we can't.

as we have also discussed before, the majority of the most valuable companies we have funded have been started by founders in their 20s (with exceptions like stripe on the lower side and zenefits on higher side).

DodgyEggplant3• 12 years ago

In some fields, especially science, you probably need to be 10-15 years in to start something, so 30's and up. The question: what is the age stats if you do not calculate science oriented startups.

chroma3• 12 years ago

If (as I think the case is) YC doesn't discriminate on age at all, one might still expect a lower acceptance rate for higher ages. Allow me to explain why.

As much as we hate to admit it, our brains (like the rest of our organs) start to decline in our 20's.[1] We're not quite as quick as we used to be. We can't work as hard. We take a little longer to learn new things. In almost every discipline, crystallized knowledge more than offsets this loss of fluid intelligence. Eventually though, age catches up with us. Combine that with YC's very high bar (accepting the top 3% of applicants), and even a slight dulling due to aging would cause a discrepancy.

1. http://lesswrong.com/lw/4gi/age_fluid_intelligence_and_intel...

frostmatthew2• 12 years ago

even a slight dulling due to aging would cause a discrepancy

Only if you completely discount the value of experience.

andrea_s• 12 years ago

That's blatant ageism (and a very common misconception, too).

adamzerner• 12 years ago

Why sample as opposed to looking at the whole population? The typical reason is, "because it's difficult to survey the entire population". But in this case I'd think that it'd be easy because I assume you have all the data.

klochner• 12 years ago

They likely don't record the team composition, so this would require hand-labeling of any data points.

Do we know total applicants? I'd like to know what n produced a 5% sample.

komrade• 12 years ago

It would be interesting to see the proportion of foreign entrepreneurs (applied from abroad).

yurisagalov• 12 years ago

Sam answers this in a different point in this thread -- https://news.ycombinator.com/item?id=8889333 (40% are born abroad)

staunch3• 12 years ago

Socioeconomic diversity is the elephant in the room.

jiggy20114• 12 years ago

How could YC even be able to calculate that?

staunch3• 12 years ago

It's pretty easy to figure out what class people are in if you actually care to know.

Competition is extremely high to enter Stanford, with an admit rate of 7.2 percent. And those who make it in are, as a group, financially better off than the rest of the American population. According to Director of Financial Aid Karen Cooper, the median family income at Stanford is approximately $125,000; by contrast, the median family income in the United States in 2008, the last year for which data are available, is $61,521.

http://www.stanforddaily.com/2010/08/12/diversity-remains-on...

http://en.wikipedia.org/wiki/Household_income_in_the_United_...

Now try to guess what class these people were born into: http://www.ycombinator.com/people/

I'm still a huge fan of YC (very much not a hater) but on this rather large issue I think they're on the wrong side of history. A truly good actor compared to most, but far short being egalitarian.

aaronz8• 12 years ago

How about the non-hispanic, non-black minorities?

Or even more interesting to me, how about the distribution of education level and background?

blazespin• 12 years ago

Yeah, I find a lot of software organizations it's becoming more and more a disadvantage of being white. Not sure I can complain, it's worked out for me, but I worry about my kids.

jmcohen• 12 years ago

All evidence points to the contrary. 34% of Facebook's employees are Asian, but only 19% of managers are.

NYT article: http://bits.blogs.nytimes.com/2014/06/25/facebook-mirrors-te...

Facebook's "Equal Employment Opportunity" report: https://fbnewsroomus.files.wordpress.com/2014/06/facebook-ee...

proveanegative• 12 years ago

Data on citizenship would also be interesting. How many founders are foreigners and how many are US citizens?

sama5• 12 years ago

for last batch, more than 40% of the founders were born outside the US.

sama5• 12 years ago

we didn't look at this here because we fund a lot of asian and indian founders, and we just don't have enough of a sample to look at e.g. native american founders.

on things like education level, country of birth, age, etc we've previously looked at and discussed the data and it seems very diverse.

atom-morgan• 12 years ago

What about someone who's mixed?

kolbe• 12 years ago

Michael, I think you should really ask yourself: what good can come of this? It doesn't really demonstrate anything about YC's fair-opportunity goals, and it opens the topic up for significant scrutiny--both by rights activists, and now (even worse?) by stats majors who are always looking for an opportunity to tear apart someone's oversimplified models.

logicallee• 12 years ago

Michael, I think you should really ask yourself: what good can come of this?

I got this one!

The good that comes is you solve the current status quo, where on a thread like "Why silicon valley works" people make comments and genuinely feel like this: https://news.ycombinator.com/item?id=8552487

If you are a white male, in your early 20's and went to Stanford, go for it. If you're not, don't even try to play the game; it's rigged to use you, not help you.

or up-thread someone else

Complete and utter bull. Maybe in [YC's] world, sure! White? 20-something? No-kids? Harvard/Stanford? B2C? Let’s chat! > Early-40's/Late-30's? Not-white? Married w/ kids? Not-Harvard/Stanford? B2B? Nope. (Don't bother to apply to YC either).

Theirs is a strong argument when 0.0% of founders are black. When it's 4%, man, 1 in 25 is better odds than 80% of the things you're doing to build a $100M exit anyway. You'll go for it and apply.

I gave up as I'm in a hurry, but search 'Hackernews search' (hn.algolia.com) for white/asian, "bother to apply" and similar words - there are a ton of comments by people who think they shouldn't because it's 0%. They should!!!

This blog post is totally great, and a bit surprising. Kudos.

unlaut• 12 years ago

The question is, would this post still be made publicly, in the same manner, if the data actually found serious discrimination happening?

Who knows, maybe YC is actually divinely meritocratic and there are no biases. Maybe they are transparent enough to make their biases public if they were discovered, and I absolutely admire that YC takes the initiative to check.

But if we were to be scientific about it, having YC take a look at discrimination within YC wouldn't be considered a reliable experiment that you could draw any conclusions from.

maceo2• 12 years ago

"The good news is that there is no disadvantage to applying to YC as a female or minority founder."

I don't understand why that's good news.

It'd be good news if YC were actively preferring female and minority founders. Everyone acknowledges the glaring diversity problem in tech, but apparently YC is happy standing on the sidelines.

john_b• 12 years ago

From the OP:

"We want to be sure we’re not discriminating against founders in our funding decisions at YC"

This is not an affirmative action program. YC is an accelerator, not a charity. It primarily seeks long term positive returns, and to do that you have to evaluate people and their ideas based on merit, not their gender. Having more female founders would probably be better for the startup ecosystem as a whole, but "actively preferring" a class of candidates based on attributes poorly correlated with their probability of success would be bad for YC.

_pius2• 12 years ago

This is not an affirmative action program. YC is an accelerator, not a charity.

Let's not inject FUD about affirmative action into this conversation.

MIT, for instance, has a robust affirmative action program and I can assure you the Institute isn't in the business of picking unqualified applicants. One of the most famous (and successful) affirmative action programs in the world is implemented by the NFL — hardly a charity.

http://en.wikipedia.org/wiki/Rooney_Rule

Institutions that implement affirmative action do so selfishly to ensure that they're fielding the best proverbial team possible.

maceo2• 12 years ago

Look, whether we like it or not, VCs prefer white or asian founders over black or hispanic founders. That's fact, not conjecture [1]. I understand that YC is an accelerator. It's also a cultural institution. Blacks and hispanics are at a disadvantage when raising because VCs DON'T evaluate strictly on merit; they evaluate on a whole host of characteristics, many of which are subject to bias.

The YC stamp of approval is a valuable badge when fundraising. I'd like to see them find a way to use this to combat the inherent bias that women, blacks, and hispanics face when fundraising.

[1] https://www.cbinsights.com/blog/venture-capital-demographics...

adrianbordinc• 12 years ago

Why should they prefer different genders or races? What was that thing called... oh yes, equality.

shiftpgdn• 12 years ago

Who gives a shit? YC should fund good ideas with good founders regardless of skin color, gender or whatever else.

impendia• 12 years ago

I don't understand why that's good news.

This is because of "implicit bias", wherein people discriminate against women and/or minorities unconsciously and unintentionally.

For example, here is an article demonstrating that orchestra auditions were subject to bias against women:

http://scholar.harvard.edu/files/goldin/files/orchestrating_...

I suspect that everyone (or nearly everyone, anyway) evaluating the auditions wanted the best musicians regardless of gender, but the linked study determined that when orchestras began using blind auditions, more women were offered positions.

In YC it is impractical to conduct "blind interviews", so that it is good practice for the partners to ensure that they aren't accidentally disadvantaging women and minorities.

Whether YC should actively prefer female and minority founders is, I believe, much more subjective.

evanwarfel• 12 years ago

Good to know that YC doesn't seem to be affected by biased selection process. The low overall percentage of minority or female founders is a reflection of the current state of society, which, alas, isn't where it could be.

dimdimdim• 12 years ago

The post does not add any value -- "you are not at a disadvantage" means we are evaluating everyone with the same scale. Why should this be a surprise? this is the least we would expect anyway.

moisesvega• 12 years ago

It would be interesting to have some more insights, like how many of rhw minority groups ( black or hispanic as the article says ) were funded, also more meaningful like:

Race/Sector of interest Age/Sector of interest Gender/Sector of interest People with Higher Scholar degree/ Sector of interest Startups funded by sector

j_baker• 12 years ago

this will not be a zero-sum game

Is there some kind of rule at YC that mandates that blog posts must insist that nothing is a zero-sum game?

byEngineer• 12 years ago

yes! It is the same rule that says that we must be sad that 50% of all applicants aren't women.

auggierose• 12 years ago

It's not like the number of entire YC 2015 winter applications is super large. Why take a sample?

billconan• 12 years ago

I really want to see what percent introvert founders take.

I have been reading articles saying introversion is not good for entrepreneurship. I wonder if this is true.

if it is true, will VCs treat introvert founders like they do to minorities?

CmonDev• 12 years ago

Some types of discrimination are not "fashionable" enough. If you are not of right color/gender you are out of luck as a minority.

dirkdk• 12 years ago

the big question is of course about self selection: how many nonwhite, female and non-Stanford entrepreneurs would like to apply for YC but don't because they think their chances are basically zero?

YC Demographics

Related Stories