How technology has been handling with diversity

Thiago Lima
8 min readMay 20, 2021
Image from Netflix documentary “Coded Bias”

Taking in consideration the statement: “technology will shape our future”, it’s inevitable to think how diverse this shape will be. Algorithms are getting deeper in our routines, dictating what content we should see in social media, what priorities we should receive in a hospital queue and if our CV’s should be analysed for a specific job.

Netflix approaches themes like these more deeply in documentaries like “Coded Bias” and “The Social Dilemma”, and inspires me to search for answers about diversity in the world of technology.

This study is also based on surveys of the last three years, made by Stackoverflow.

This study focuses only on respondents of the US, which has five of ten of the greatest big techs of the world, known as Amazon, Google, Apple, Facebook and Microsoft. Besides, the US has a great diversity according to the 2019 U.S Census Bureau, which points out about 40% of the population been non white people, which is perfect for this work.

How diverse is the tech workforce in the US?

The Stackoverflow’s survey has samples of 14.252, 17.471 and 9.402 respondents that live in the US, respectively in the years of 2018, 2019 and 2020.

Below, we can understand better how ethnical diversity is applied to the tech workforce in the US.

The US Census detected 60.1% of Non-Hispanic white people, but they were overrepresented in the tech workforce in 2020 with 83.5%, losing just 2 percentual points since 2018.

As well as white developers, Asian ones are overrepresented too. The Stackoverflow’s survey divides Asians in three different ethnicities being East Asian, South Asian and Southeast Asian, together in the US they sum 10.7% workforce representation in 2020. It’s little less than double identified by the 2019 US Census.

Hispanics and Latinos have increased almost 1.5 percentual points since 2018, reaching 6.7% representation in 2020 US tech workforce, which represents little more than a third of the value identified by the 2019 US Census.

Middle eastern and Black or African descent had the lowest increase between 2018 to 2020, achieving 2% and 3% respectively of the workforce. Despite the US Census accounts 13.4% of the population being Black or African descent and Middle eastern was not covered by the 2019 US Census.

Multiracial and Native American seem to have the most proportional representation in the US tech workforce. Stackoverflow surveys started counting Biracial and Multiracial just in 2019 separately, and in 2020 together they achieved 3.8% of the developers representation, against 2.8% of Multiracial detected by the US Census.

Native American ethnicity is divided by two in the 2019 US Census as Native Hawaiian and American Indian, together they sum 1.5% of US population representation. Native American are 1.4% representation of 2020 US tech workforce, losing 0.26% of representation in the last two years.

Below, we can see the distribution of the tech workforce by gender.

Men are the majority in the tech workforce, and they have been decreasing very slowly since 2018, achieving 88% in 2020 against 49.2% of representation of the US population, according to the 2019 US Census.

With a narrow growth between 2018 to 2020, Women achieved 10.9% of the workforce. Although Women represent 50.8% of the population according to the 2019 US Census.

Non-binary or non-conforming were not covered by the 2019 US Census, but they represent 2% of the workforce in 2020. Transgender was covered just in 2018 Stackoverflow’s survey with 1.2% of representation.

How satisfied are minorities among their peers?

According to the data crossing above, it’s possible to assume that minorities in the US tech workforce are Hispanics or Latinos, Middle eastern and Black or African descent by ethnicity and Women and Non-binary or non-conforming by gender.

One of the questions in the Stackoverflow’s survey was: “How satisfied are you with your current job?”. But, to answer this question, it’s mandatory the respondent had been engaged in a job.

In the 2018 survey, Stackoverflow used 7 scales of “Job Satisfaction”, whereas in the following years they used 5 scales. For this reason, this study will not use the 2018 survey to answer this question.

Taking this in consideration, it’s important to say that 95.9% of the 2019 sample was currently working, whereas in 2020 it was 90.3%.

Below we can see the table distribution of ethnicity and gender by each of the five scales of satisfaction in 2019 and 2020. White ethnicity was included in the tables to comparative effects, so was Man.

Table 1: Job satisfaction by ethnicity and gender in 2019

In 2019, White or European descent were much more satisfied with their jobs compared to minority ethnicities, to be more precise, they are 11.1 percentual points more represented compared to the Black or African descent which were very satisfied in their jobs. This difference falls to 7.7 percentual points comparing very satisfied White or European descent with Middle Eastern, and 5.8 points compared to Hispanic or Latino/a/x, in the same scale of satisfaction.

Although Women have less representation in the tech workforce, in 2019 they were much more satisfied with their jobs compared to the other genres. Non-binary or non-conforming were 11.3 percentual points less represented than Women that felt very satisfied with their jobs. Compared to Women, Men have 3 percentual points less in the same scale above.

Black or African descent and Women, who did not answered this question represent 10.2% and 6% respectively.

Table 2: Job satisfaction by ethnicity and gender in 2020

In 2020, ethnicities and genres were more compatible between themselves on the job satisfaction perspective, except for the Middle Eastern having 5.9 percentual points less representation compared to White or European descent very satisfied, and almost 4 points more compared to White or European descent very dissatisfied.

Black or African descent and Non-binary or non-conforming grew more than 10 percentual points very satisfied scale in 2020.

Stackoverflow asked developers what their job seeking status was in 2019 and 2020. So, it’s possible to find some false positives, i.e, developers who said they were very satisfied with their jobs and were actively looking for a job.

Below, we have two primary columns, subdivided by years. The first primary column is related to the percentage of false positives applied just to very satisfied scale in some ethnicities and genders. The second one, shows us the correction of very satisfied developers by ethnicities and genders, removing the false positives.

Table 3: Developers who felt very satisfied and were actively looking for a new job.

The highest false positive identified goes to Hispanic or Latino/a/x in 2020, which increased the gap between them and White or European descent very satisfied to almost 4 percentual points in 2020. The increase/decrease was not substantial to all the others ethnicities and genders.

Do minorities have been compensated equally in the US tech industry?

According to Forbes, salary is correlated to job satisfaction. So, it’s important to understand the difference in salaries across the last few years. As in the last question, it’s mandatory the respondent had been engaged in a job.

The 2018, 2019 and 2020 samples had compensation question not answered about 34.1%, 19.0% and 22.1% respectively.

Below, we can see the mean annual salary by minority ethnicities.

The annual salary gap between the minority ethnicities and White or European Descent has been decreasing since 2018, for most of the minorities. The Middle Eastern reached the highest mean to the compared ethnicities in 2020, and in the same year Black or African descent reached the shorter gap since 2018, compared to White or European descent, with just $4,000.00 less.

Hispanics or Latino/a/x achieved the highest annual salary gap compared to majority ethnicity in 2020, with $12,000.00 of difference, which have been increasing since 2018.

Below, we can see the mean annual salary by gender.

The annual salary gap between Men and Non-binary or non-conforming decreased in $1,200.00 comparing 2018 to 2020, accounting $8,800.00 of difference between these two ethnicities in 2020.

Women have the lowest mean annual salary of the category, with $12,300.00 of difference compared to Men in 2020.

The US tech industry has been absorbing the minorities’s labor?

One of the questions in Stackoverflow’s survey was related to the current employment status of the respondent.

In order to understand better if the tech industry has been absorbing the minorities’s labor, we need to understand the unemployment rate, i.e, people that are not employed, but are looking for work.

Below, we can see the unemployment rate by minorities across the last 3 years.

Table 4: Unemployment rate across the last 3 years.

Black or African descent continue to have the most unemployment rate in 2020 compared to other ethnicities, but this rate falls by half, compared to 2018. The other minority ethnicities got a minor decrease in unemployment rate in the period.

Non-binary or non-conforming were the most promissing gender, losing 4.8 percentual points of unemployment rate between 2018 to 2020. Women remain with the highest unemployment rate.

Conclusion

The technologies that will shape our future have been done mostly by White or European descent men, which renders these technologies more susceptible to bias.

Minority ethnicities and genres have been increasing in very slow steps. What suggests a field that should be worked more deeply.

Almost all minorities increased their job satisfaction rate, with a great highlight to Black or African descent that increased more than 7 times the numbers of very satisfied collaborators compared to very dissatisfied ones in 2020. However, Black or African descent still have the highest unemployment rate.

The annual salary gap between White or European descent and Hispanic or Latino/a/x increased considerably between 2018 and 2020, achieving $1,000.00 per month of difference.

What is your perception? What do you think that can be done to ensure more diversity in the future of the tech industry?

The detailed study can be found on my github.

--

--