6 ms·

I found the title to be phrased very weakly compared to the abstract. > We study gender and race in high-impact entrepreneurship using a tightly controlled ran

I found the title to be phrased very weakly compared to the abstract. > We study gender and race in high-impact entrepreneurship using a tightly controlled randomized field experiment. We sent out 80,000 pitch emails introducing promising but fictitious start-ups to 28,000 venture capitalists and angels. Each email was sent by a fictitious entrepreneur with randomly assigned gender and race. Female entrepreneurs received 9% more interested replies than males pitching identical projects and Asians received 6% more than Whites. Our results suggest that investors do not discriminate against female or Asian entrepreneurs when evaluating unsolicited pitch emails and that future research on investor biases should focus on networks and in-person interactions.

nairboon 6y ago
>Our results suggest that investors do not discriminate against female or Asian entrepreneurs when evaluating unsolicited pitch emails... Are they implying that investors do discriminate against non-Asian males? Why write it the other way around.
- jhgb 6y ago
  > Why write it the other way around. Maybe that's how the hypothesis under test was phrased?
  - throwaway894345 6y ago
    Indeed, reading further into the article they seem to be trying to answer, “At what stage in the funding pipeline do VCs become biased against women?”. It’s interesting that they don’t entertain the possibility that the gap might be caused by something other than bias (at least so far... I’m only 6 pages in). For example: > Combating bias requires understanding how and where it manifests. Combating bias requires understanding how and where it manifests. In light of the substantial gender imbalance in real-world investments, our results suggest a larger-than-expected bias against female entrepreneurs in other settings. They seem to be trying to combat bias without having even established that bias is the cause. Indeed, they take the statistically significant discrimination in favor of women in the initial stage as evidence that there must be even more bias against women in later stages. Given that they haven’t cited any research that would explain why it must be bias and not some other factor, I wonder why the paper is written with this assumption? Perhaps the authors felt it would be poorly received if it so much as kept an open mind about other possible causes? Or maybe this is just common knowledge in the field?
    - inglor_cz 6y ago
      "It’s interesting that they don’t entertain the possibility that the gap might be caused by something other than bias" Imagine a medicinal study that would concentrate on the hypothesis that sinners get cholera as a divine retribution for their sinful lives, and would not even entertain other possible causes, such as ... dirty drinking water. We would rightly locate such a study into, say, 1840, but definitely not 2021. That is where authors of this study belong mentally. Into the science of the era when Lincoln was young.
      - throwaway894345 6y ago
        > We would rightly locate such a study into, say, 1840, but definitely not 2021. You give far too little credit to 1840 and far too much to 2021.
        inglor_cz 6y ago
        Look up 1854 Broad Street cholera outbreak. That was when the scientific breakthrough regarding bad water took place - and many actually refused to accept the results. Before that, the most "scientific" theory about cholera outbreaks was the miasma theory, but religious and pseudoreligious beliefs about divine punishment or influence of the stars were not fringe yet.
        throwaway894345 6y ago
        To be clear, my earlier comment was made in jest. I’m familiar with the Broad Street outbreak, but I’m not aware that religious theories were commonplace among scientists of the time. Notably that science wasn’t taken seriously by society more broadly doesn’t mean that the science of the time was “tainted by religion” (I say this as a religious person). Perhaps we’re making different points: mine is that science today is contaminated by a fervor resembling religion (though I don’t want to overstate the problem—overwhelmingly I think the effect is very low for the time being and I suspect it is highest in areas that are most pertinent to the ideology in question, which thankfully tends not yet to be the hard sciences or engineering disciplines lest our bridges fail, buildings collapse, and rockets explode on the launchpad).
- sltkr 6y ago
  The whole thing is written from the assumption that investors discriminate against women and nonwhites, and that the purpose of research is to confirm this pre-existing belief, rather than find out the objective truth. > [..] future research on investor biases should focus on networks and in-person interactions. Read: the results contradicted our hypothesis, so rather than update our hypothesis, let's shift the focus of research in hopes that we find evidence to support it.
  - maccard 6y ago
    This is exactly how science should work. You can't perform a study, look at the data and pick a result that suits your bias. You're supposed to suggest a hypothesis, and conduct your experiment. If the results contradict your hypothesis, it doesn't mean the opposite of your hypothesis is true, simply that your hypothrsis was false, and that's ok. It's also more impressive that they have simply published their results, rather than burying them and running another experiment.
    - mc32 6y ago
      So what if they had found bias, do they say “look we found bias!” Or do they do additional studies to make sure it’s not an aberration or improper methods?
      - sokoloff 6y ago
        It seems ideal if they did both of those things. Publish and make recommendations for future research, exactly as this paper does.
      - maccard 6y ago
        Both. They publish and say "yes, we found bias on X, and our next study will verify via Y"
    - bluecalm 6y ago
      So you think the data is only relevant to the hypothesis you decided it's relevant for before you collect it? "Have you seen the fire sir?" "I have only seen smoke" "But have you seen the fire?" "It's irrelevant because I was only looking for smoke!". I mean I know they teach it in naive introductory experiment design in some soft science departments but c'mon it's neither smart nor correct. If you need a separate experiments for everything because otherwise you will run into "one of the 100 hypothesis is bound to be correct" problem then your priors are too weak, you're relying on getting lucky anyway and just end up discarding useful datapoints. If it's 5 out of 100 hypothesis fitting your data or 5 hypothesis out of 100 experiments passing "statistical significance" check doesn't matter at the end of the day. Both lead to shaky conclusions. The data being relevant depends on the experiment design (how it's collected). It doesn't depend on what you thought it's relevant for beforehand.
- mc32 6y ago
  From omission it seems they reply less to black and white men. But these groups are large, however, sadly I don’t think it’d be easy to test on feigned education/class, so we won’t know what causes the bias.
  - dash2 6y ago
    They only did Asian versus white.
  - throwaway894345 6y ago
    If you’re curious about why they didn’t include black names: > After long and thoughtful consideration, we discarded the idea of assigning African-American names to our fictitious entrepreneurs. First, there has been a concern about the ability to disentangle race and socioeconomic perception by using distinctly African-American names (Fryer and Levitt (2004)). Second, and most importantly, African-American entrepreneurs are woefully underrepresented in high-impact entrepreneurship (less than 1%, see Gompers and Wang (2017)). This poses a substantial challenge for our setup, because even names that are disproportionately likely to be African-American in the general population (such as “Martin Jackson”) may not be perceived as African-American by investors in this context. Studying discrimination against African-Americans and other underrepresented minorities in high-impact entrepreneurship remains an important problem.
mc32 6y ago
My question would be why don’t they do a cursory ‘bg’ check on who these people are before replying? With so much scamming and the different phishing attacks, do you respond to random inquiries? That would seem rather naive, if true.

nairboon 6y ago

>Our results suggest that investors do not discriminate against female or Asian entrepreneurs when evaluating unsolicited pitch emails... Are they implying that investors do discriminate against non-Asian males? Why write it the other way around.

jhgb 6y ago
> Why write it the other way around. Maybe that's how the hypothesis under test was phrased?
- throwaway894345 6y ago
  Indeed, reading further into the article they seem to be trying to answer, “At what stage in the funding pipeline do VCs become biased against women?”. It’s interesting that they don’t entertain the possibility that the gap might be caused by something other than bias (at least so far... I’m only 6 pages in). For example: > Combating bias requires understanding how and where it manifests. Combating bias requires understanding how and where it manifests. In light of the substantial gender imbalance in real-world investments, our results suggest a larger-than-expected bias against female entrepreneurs in other settings. They seem to be trying to combat bias without having even established that bias is the cause. Indeed, they take the statistically significant discrimination in favor of women in the initial stage as evidence that there must be even more bias against women in later stages. Given that they haven’t cited any research that would explain why it must be bias and not some other factor, I wonder why the paper is written with this assumption? Perhaps the authors felt it would be poorly received if it so much as kept an open mind about other possible causes? Or maybe this is just common knowledge in the field?
  - inglor_cz 6y ago
    "It’s interesting that they don’t entertain the possibility that the gap might be caused by something other than bias" Imagine a medicinal study that would concentrate on the hypothesis that sinners get cholera as a divine retribution for their sinful lives, and would not even entertain other possible causes, such as ... dirty drinking water. We would rightly locate such a study into, say, 1840, but definitely not 2021. That is where authors of this study belong mentally. Into the science of the era when Lincoln was young.
    - throwaway894345 6y ago
      > We would rightly locate such a study into, say, 1840, but definitely not 2021. You give far too little credit to 1840 and far too much to 2021.
      - inglor_cz 6y ago
        Look up 1854 Broad Street cholera outbreak. That was when the scientific breakthrough regarding bad water took place - and many actually refused to accept the results. Before that, the most "scientific" theory about cholera outbreaks was the miasma theory, but religious and pseudoreligious beliefs about divine punishment or influence of the stars were not fringe yet.
        throwaway894345 6y ago
        To be clear, my earlier comment was made in jest. I’m familiar with the Broad Street outbreak, but I’m not aware that religious theories were commonplace among scientists of the time. Notably that science wasn’t taken seriously by society more broadly doesn’t mean that the science of the time was “tainted by religion” (I say this as a religious person). Perhaps we’re making different points: mine is that science today is contaminated by a fervor resembling religion (though I don’t want to overstate the problem—overwhelmingly I think the effect is very low for the time being and I suspect it is highest in areas that are most pertinent to the ideology in question, which thankfully tends not yet to be the hard sciences or engineering disciplines lest our bridges fail, buildings collapse, and rockets explode on the launchpad).
sltkr 6y ago
The whole thing is written from the assumption that investors discriminate against women and nonwhites, and that the purpose of research is to confirm this pre-existing belief, rather than find out the objective truth. > [..] future research on investor biases should focus on networks and in-person interactions. Read: the results contradicted our hypothesis, so rather than update our hypothesis, let's shift the focus of research in hopes that we find evidence to support it.
- maccard 6y ago
  This is exactly how science should work. You can't perform a study, look at the data and pick a result that suits your bias. You're supposed to suggest a hypothesis, and conduct your experiment. If the results contradict your hypothesis, it doesn't mean the opposite of your hypothesis is true, simply that your hypothrsis was false, and that's ok. It's also more impressive that they have simply published their results, rather than burying them and running another experiment.
  - mc32 6y ago
    So what if they had found bias, do they say “look we found bias!” Or do they do additional studies to make sure it’s not an aberration or improper methods?
    - sokoloff 6y ago
      It seems ideal if they did both of those things. Publish and make recommendations for future research, exactly as this paper does.
    - maccard 6y ago
      Both. They publish and say "yes, we found bias on X, and our next study will verify via Y"
  - bluecalm 6y ago
    So you think the data is only relevant to the hypothesis you decided it's relevant for before you collect it? "Have you seen the fire sir?" "I have only seen smoke" "But have you seen the fire?" "It's irrelevant because I was only looking for smoke!". I mean I know they teach it in naive introductory experiment design in some soft science departments but c'mon it's neither smart nor correct. If you need a separate experiments for everything because otherwise you will run into "one of the 100 hypothesis is bound to be correct" problem then your priors are too weak, you're relying on getting lucky anyway and just end up discarding useful datapoints. If it's 5 out of 100 hypothesis fitting your data or 5 hypothesis out of 100 experiments passing "statistical significance" check doesn't matter at the end of the day. Both lead to shaky conclusions. The data being relevant depends on the experiment design (how it's collected). It doesn't depend on what you thought it's relevant for beforehand.
mc32 6y ago
From omission it seems they reply less to black and white men. But these groups are large, however, sadly I don’t think it’d be easy to test on feigned education/class, so we won’t know what causes the bias.
- dash2 6y ago
  They only did Asian versus white.
- throwaway894345 6y ago
  If you’re curious about why they didn’t include black names: > After long and thoughtful consideration, we discarded the idea of assigning African-American names to our fictitious entrepreneurs. First, there has been a concern about the ability to disentangle race and socioeconomic perception by using distinctly African-American names (Fryer and Levitt (2004)). Second, and most importantly, African-American entrepreneurs are woefully underrepresented in high-impact entrepreneurship (less than 1%, see Gompers and Wang (2017)). This poses a substantial challenge for our setup, because even names that are disproportionately likely to be African-American in the general population (such as “Martin Jackson”) may not be perceived as African-American by investors in this context. Studying discrimination against African-Americans and other underrepresented minorities in high-impact entrepreneurship remains an important problem.

mc32 6y ago

My question would be why don’t they do a cursory ‘bg’ check on who these people are before replying? With so much scamming and the different phishing attacks, do you respond to random inquiries? That would seem rather naive, if true.