Education is the first step to hiring diversely
“Fixing the underrepresentation of women and people of color in the [tech] industry is not just about making sure we have the necessary people. It is also a matter of social justice, and indeed creating better products and services” — Dr. Nicole Forsgren and Jez Humble “The Core Belief Keeping Marginalized Groups Out of Tech,” Model View Culture 2015 Quarterly No. 2
Job descriptions are one of the first contact points between people and a company. Even when combined with other hiring efforts, they act as a gatekeeper and provide the first opportunity for a candidate to imagine themselves at a company with a specific role. From the perspective of diversity, they can help with hiring from marginalized groups.
There is more to the issue of hiring and diversity than representation. As Dr. Nicole Forsgren and Jez Humble state in the quote above, this is an issue of social justice and one that is heavily embedded in the larger culture.
So where does one start with trying to fix this issue of underrepresentation? Knowledge and recognition of the problem is key.
Knowledge of Unconscious Bias
Recently, I went through the process of rewriting our job description for “Online Course Facilitator” as a way of testing out Textio, a tool to help improve job descriptions. One key feature is the ability to see gender bias in words and phrases.
The gender recognition algorithm of Textio is built on data mining and analysis of ‘historical’ job descriptions. So it highlights words or phrases that have been identified to attract more male or female applicants based on the corpus of job descriptions and survey results conducted by Textio. English, unlike numerous other languages, does not have many words that are inherently gendered. Textio’s recognition of a word or phrase as gendered is based on a correlation in the data that is available.
As someone who has expressed skepticism due to unrecognized bias in datasets, I think Textio provides a fantastic entry point to discuss unconscious bias in our culture as well.
Let’s take two examples of gendered terms in Textio. The word ‘expert’ is identified as a masculine term. In other words, it “draws more male job-seekers.” I would guess that this has to do with the commonly discussed issue of imposter syndrome and how it more commonly impedes marginalized groups entry into technical fields, such as the technology sector. Recognition of the unconscious bias the term ‘expert’ holds can be a starting point for a company to think about imposter syndrome and how it affects hiring and internal culture.
The word ‘teaching,’ on the other hand, is identified as feminine. Education, as a field in the US, has a gender-gap that adversely affects men. The gender-gap is particularly pronounced with the earlier grades — the 2014 US Bureau of Labor population survey only had 2.8% of pre-K and kindergarten faculty as male.
What could be causing some of these biases? In an analysis of the sex panic created by homophobia in America, Jim D’Entremont observes “There was a widespread assumption that any man who sought work with very young children must be gay.” This is during a time when being labelled as gay was a mark of abnormality in the US.
Is the current low representation of men in education a remnant of some of these biases? I honestly don’t know. But they both indicate a culturally embedded male role that strongly disassociates with young children.
Questions about language and gender roles arise from uncovering of these hidden biases. Even though there may not be a readily available answer, knowledge allows room to explore and discover potential solutions or steps towards a more lasting solution.
Deeper Understanding of Identities
Having data on the representation of marginalized groups is also a key to improving the situation. The new US government initiative on diversity in tech, TechHire, includes a large component around data and using data.
For marginalized communities, that means understanding the complexity of identities.
While I previously discussed the lack of gender identities represented in data, I didn’t talk about the concept of intersectionality: how marginalization is not isolated and being part of multiple marginalized communities is a different experience from either marginalized identity on its own. This term was created by African American women, specifically Kimberlé Crenshaw and popularized by Patricia Hill Collins, and their recognition that they had different concerns from ‘women’ that included stigmas associated with being African American.
Discussion about marginalized and intersectional identities is much needed within the technology sector, where diversity is code for white women. Without recognition, there isn’t the ability to collect data and improve the situation. How can TechHire improve diversity through data if there isn’t diversity in the data?
Textio does not address intersectional identities: there is simply a scale between men and women. While I don’t think this is a fault of Textio, it does highlight a hidden bias in the dataset they are working off of.
It highlights a marginalization and erasure of these intersectional identities in conversations of diversity.
For the LGBTQIA+ community in the US, erasure of identity in tech has been a consistent problem. While some gains have been made, such as Facebook’s expanded gender identity options, there is a significant lack of understanding and knowledge about these identities. Despite allowing additional options, Facebook put in place the ‘real name’ policy that marginalizes, endangers, and erases LGBTQIA+ populations. These two different policies create a dissonance and demonstrate a strong lack of understanding around the issues of gender identity.
If hiring inclusively is truly a social justice issue, knowledge is a key factor to deconstructing the current situation and moving beyond it. We need tools like Textio to provide concrete examples of hidden biases. We also need better understanding of identities and marginalization to ensure our well-intentioned efforts aren’t leading to greater marginalization.
Perhaps most importantly, we need to allow the marginalized communities to have the primary voice in this process to ensure we are inclusive with our hiring and the services we provide.
Author’s Closing Note: Even though I focused on gender and race, these are not the only aspects of diversity. Neurodiverse individuals, people with disabilities, and other marginalized groups bring unique and very important perspectives to the technology sector and should also be included in conversations around diversity.
Here is an uncomprehensive list of awesome people or organizations working to bring marginalized voices in tech to the forefront of conversations.