Gender and Race Preferences in Hiring at Silicon Valley Tech Firms [pdf]

I have heard story after story from peers (around my age), about passing numerous rounds of phone and test interviews, only to see the interviewer's face fall, as soon as they walk in the door. It's happened to me, but, thankfully, before being flown out (if it's a contract application, then we're on the hook for the cost), and revealing the age.

It happens with highly qualified older white guys, so I'm sure that it will happen with other distinguishing characteristics.

The simple fact of the matter is, is that people have certain traits they are looking for, and the final call is always a "gut feeling."

I've learned to just avoid all the agita by making it clear that I'm "of a certain age," right up front. I don't get many offers, but at least I don't have those blasted recruiters, gushing about my résumé, only to shred my contact card, as soon as they find out I'm old (I guess the industry is crawling with 32-year-olds with 30 years' experience).

I don't think it's a good idea to have AIs do the selection (as has been suggested numerous times -usually by folks in the AI business). Humans need to work with humans. If an AI puts someone into a situation where they are treated like garbage by their managers and/or co-workers, then they will be miserable (and unproductive); especially if they have done something like uproot their lives to move somewhere (I have also heard many stories from people that have uprooted their lives to go to "the perfect" job, only to come back, a few months later, broken and cynical).

If the people involved don't want to work with someone like me, then I don't want to work with them. That's easy for me to say, though. It can really stink, when you're hungry and need the work.

(and vice versa for ethnicity vs gender).

The fact that they don't show these breakdowns is a major weakness of this study.

Or strength. Depending on whether the desired metric is scientific correctness, or political marketability.

IMHO, if the tech industry wants to live up to its narrative of meritocracy, this is one obvious improvement over existing processes. No, it won't take care of pipeline problems; no, it won't solve the "tipping point" problem (i.e. where candidates of underrepresented groups are dissuaded by a lack of pre-existing representation, making it very hard to go from zero to one, so to speak). That said, we're uniquely positioned as an industry to do this - technical interviews are similar to auditions in that they hinge on skill-based performance - so why not? It can't be any less arbitrary than asking random questions about manhole covers and light bulbs.

I'd posit that even the "show your thinking process" parts could be done in this way - e.g. via text chat, or inline comments, or maybe even using voice obfuscation and/or neutral avatars.

    Women:                       +9 to +10% chance of callback relative to men
    Black, Hispanic, and Asian:  -8 to -13% chance of callback relative to White people.

I'm making my way through the main paper but there's a lot of tables in there that would take me a lot of reading to decipher and it's really not clear which are best to use for any purpose. The tables do break out the separate ethnicities separately though (along with callback/interview/offer rates, not just callback) and it does appear that Black, Hispanic, and Asian very much do not have the same experience. In particular, the graph on page 22 seems to show that, when it comes to receiving offers, female candidates have a 29.5% advantage in receiving offers, Black 4%, -21% Asian, and -24% Hispanic. The previous column (received interview) is much worse for Asian, Black, and Hispanic applicants though; all were significantly less likely to receive interviews than whites.

Anyway, there's lots of potentially good data in here, assuming you trust their methodology, but I personally feel like I need someone who knows this stuff better to write up a longer abstract based on all this data to really explain it to me properly.

Candidate A: Had worked in the industry, had all of the qualifications, already chock-full of some interesting ideas I wanted to hear more of from the interview alone. Excited at the prospect.

Candidate B: Had never worked in the industry, had only a handful of qualifications, barely responsive. Seemed indifferent to getting the job. Additionally, not too fluent in English, to the point where it was more than a little difficult to communicate.

Candidate A was a white man, Candidate B was a recent immigrant and a woman. The immediate supervisor for the position -- a woman -- wanted Candidate A, as did most others. However, the person running the show said, out loud I might add, that our group already had "too many pale males." I would like to repeat that: too many pale males. A significant glance was then cast at me and the guy in the wheelchair on the hiring committee, both being not-particularly-dark men. Presumably by "virtue" of our disabilities we would automatically be down for the Diversity Squad.

Candidate B was hired and turned out exactly as she was in the interview: disinterested in doing the job, lacking even some bare understanding of how to accomplish many things, always trying to find ways to do her grad school homework while on the job and pushing off her duties on someone else, rather than trying to learn her tasks. Her poor English was a significant barrier. She remained a leaden weight until she went off to be someone else's problem. She wasn't a drag due to her skin color or sex, but she was hired because of those things.

My guess is that this kind of discrimination (I use that word specifically) goes on all the time in at least some segments of the industry.

The challenge is, I suspect they ask for race & gender is intentionally requested to add bias. I’ve worked with recruiters and part of the job is indeed targeting “under represented” groups to improve the figures.

Got hired into a company that did everything possible to avoid hiring straight white men.

Everything was completely dominated by not offending anyone over anything. Unless you were the aforementioned group.

Turns out there simply isn’t a lot of the right race/gender/sex. I’m sure that’ll be different as time progresses, but for now there is clearly some catching up to do.

So it was all jr developers. A couple of which truly cared about technology, and I see good things in their future.

The rest were the most arrogant sexist bigots imaginable.

Most days were filled with hatred. Full rant fest about how evil “my kind” were. But it wasn’t personal as I was clearly one of the good ones...

So yeah, n=1 personal experience. But crude that echo chamber was scary.

I knew someone who worked as a recruiter at Boeing a few years ago, and she told me the pressure from higher ups to give preferential treatment to female applicants was crazy. She said they couldn't explicitly discriminate, but would use things like "Women in Tech" events to try to increase their female applicant pool.

We need more studies in this area. I'd also like to see studies that include other possible sources of discrimination (gender identity, age) that are known to exist in some areas. Without those other sources included in the data it's possible those sources could be skewing the data.

One scenario might be that most of the women were young (20s-early 30s), while many of the men were older (late 30s to early 50s). Just one possible scenario, I'm not saying that was the case. But in that scenario the bias could be against older candidates rather than for female ones.

It seems financial is a similar boat: Wells Fargo recently claimed a “limited talent pool" but got criticized by AOC for lacking "talent to recruit Black workers".

Obviously society, including industry, could do more to get kids into school. But right now, in the resume pile, what should employers be doing, and why is "not enough X talent" not a good defense?

https://www.washingtonpost.com/business/2020/09/23/wells-far...

I guess in the case of SV hiring outcomes, this is literally true!

I fully imagine there is bias towards Asians within SV but it'd be heavily weighted towards recent arrivals who either have uncertain visa status or aren't yet at native speaking English proficiency surely? Whereas you'd have to imagine the vast majority of black applicants meet both of those requirements with ease.

How accurate this is, however, depends a lot on how well controlling for applicant attributes worked.

> An obvious set of confounders is the applicant’s objective qualifications such as years of experience, educational attainment, and field of study, all of which affect the outcome of an application. It is widely known that many of these attributes differ across demographic groups – for example, women are less likely to major in STEM subjects, Asian Americans are more likely to have graduate degrees (Camilie Ryan and Kurt Buaman 2015). To account for these confounds, we control for total years of experience, average tenure, educational attainment (associate or less, bachelors, masters, and doctorate), field of study, and rank of the university attended (Top 10, 21-50, 51-10, ). For experience controls, we use the total number of years of experience at the time of application parsed from resume text. For average tenure, we divide the total years of experience by the number of jobs held. For university rank, we parse the Education section of the applicant’s LinkedIn profile and join this against the U.S News Global University Rankings list. If an applicant attended multiple universities, we take the lowest rank. For the field of study, we parse the Education section of the applicant’s LinkedIn profile and bucket them into one of the following categories: Technical – mathematics, computer science, engineering, economics, etc. Business – business administration, finance, accounting, marketing, etc. Law – law and legal studies. Science – natural sciences such as biology, chemistry, etc. Other – all other majors.

> An applicant’s professional and social network is another important signal that employers use to screen applicants (Fernandez and Weinberg 1997; Sterling 2014). Since one’s network tends to be demographically homogeneous, the effect of gender and race could be confounded by these affiliations. We control for this in two ways. First, we use a Referral indicator from the ATS, which indicates whether an applicant has a referral from an existing employee of the firm. Second, we identify whether an applicant has worked at the company’s talent competitor. We identify a company’s talent competitors by taking the top 10 companies from which its current employee pool comes from based on all of LinkedIn data. For example, to identify Company A’s talent competitors, we first search for all the employees of Company A using all of LinkedIn data. Once these employees are identified, we look at the previous company these employees worked at before joining Company A. We then aggregate these previous companies by count, and take the top 10 companies from which Company A’s current employee pool comes from.

> Finally, an applicant’s skills, previous job responsibilities, and fit for the job to which they applied are perhaps the most important factors in determining the success of an application. We operationalize this using a text-analytics method called Word2Vec to measure the similarity between skills and competencies listed in the applicant’s resume and the job description (Mikolov et al. 2013). To do so, we first train a Word2vec model on a corpus of resumes. Using this model, we transform each document (i.e resumes and job descriptions) into a vector representation based on skills listed in each document, and measure the cosine similarity between the resume vector vR and job description vector vJ . The higher the cosine similarity between the job description and resume vector, the better the fit. This type of approach is often used in automatic application screening tools

I'm skeptical that this actually captures what hiring managers or recruiters care about when looking at resumes, which means I'm not sure I trust the callback numbers. Submitting identical resumes with different demographic characteristics seems like a much more appropriate experimental approach here?

As for whether candidates receive an offer, they don't have anything here where they have actually evaluated the candidates' skills. I've given over 200 technical interviews, and resumes are just not that good a predictor of technical competence. It really doesn't seem to me like they have good enough controls to run this as a correlational experiment.

   Women:                       +9 to +10% chance of callback relative to men
   Black, Hispanic, and Asian:  -8 to -13% chance of callback relative to White people.