Hacker News new | past | comments | ask | show | jobs | submit login
Investors Embrace Gender Diversity, but Not Female CEOs (arxiv.org)
6 points by surbas 22 days ago | hide | past | favorite | 7 comments

To quote the abstract

  "Further, we construct machine learning models to predict whether startups will reach an equity round, revealing the surprising finding that the CEO's gender is the primary determining factor for attaining funding"
This caused me to raise my eyebrows a little bit. Throwing a bunch of data into an ML algorithm, hoping you captured all the factors and then boom suddenly it turns out the most important thing that determines whether or not you get funding is what kind of reproductive organs you have? When that happens to be what your thesis is to begin with?

I'd definitely love to see more of this ML model and what factors they fed it.

EDIT: After reading the paper, they dropped everything down to 8 variables, all of which were based solely on the founder, (I feel like including industry, and other factors would also be very influential factors) and although under their ML models they did manage to arrive at the conclusion that gender was the most predictive factor it less than 1% more predictive than the next highest factor and less than 7% more predictive than the 4th lowest factor.

They through very limited data a a bunch of different ml algorithms, but also did "normal" statistical analysis. "less than 1% more predictive" in this paper. What part are you referencing?

Table 4. Predictive model feature importance on US data.According to boththe decision tree and random forest models, the most important feature for reaching apriced round in the US is whether the company consists of onlymale founders

Specifically the decision tree analysis column. There is a great difference in the Random Forest column, which should be noted. But I am not enough of an ML expert to know which one is more accurate.

Basically if the women is the CEO, their startups will not get as much funding by a factor of about 50.

Came across this paper while reading this below article from Forbes:


PDF of paper: https://arxiv.org/pdf/2101.12008.pdf

Clearly the solution is to force investors to invest in female-led companies at gunpoint.

Not sure why that's your conclusion. Doesn't mean the bias doesn't exist.

this kinda of stats is really biaised by the nonlinearity of investments

Applications are open for YC Summer 2021

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact