Among the horsemen, Apple is the most diverse

Here are the stats.

A shareholder pressure group called Open MIC issued a diversity report for the tech industry Wednesday that left me scratching my head. They had data I’d never seen before.

I’ve put their findings for Amazon, Apple, Facebook, Google and Microsoft—along with U.S. census data—in a bar chart to make it easier to spot the anomalies. Note, for example, the over-representation of Blacks and Latinos at Amazon. Or the outsized percentage of Asians at Facebook. Of the five horsemen, Apple and Facebook’s demographics most closely match the country’s. (For why this might be, see update below.)

Not seeing the graphic? Try the website. 

From the report:

  • Black people, Latinos and Native Americans are underrepresented in tech by 16-to-18 percentage points compared with their presence in the U.S. labor force overall.10 Black people and Latinos each comprise just 5.3 percent of the Professionals category in U.S. tech industry labor data.
  • While Asians are represented at a higher rate in the tech workforce than the private sector overall, white people are 1½ times more likely than Asians to rise to an executive rank.
  • Among people of color who do enter the industry, many report isolation, discrimination and toxic work environments that prompt them to take their talent else-where. People of color leave tech at more than 3.5 times the rate of white men.

Link: Breaking the mold.

UPDATE: As reader Fred Stein points out, the report’s footnotes shed light on Amazon and Apple’s stats:

** A note on the relatively high percentage of employees of color at Amazon: According to a report on Amazon by the Institute for Local Self Reliance, “Amazon [warehouse worker] wages were an average of 15 percent below the wages for comparable positions. … These low wages disproportionately affect African-American and Latino workers, who comprise 45 percent of Amazon’s warehouse workforce, but only 8 percent of the company’s management.” 29

*** Apple’s retail employee base is included in its overall numbers. In 2014, Apple had about 66,000 employees in the U.S., including 30,000 U.S. retail employees.30 According to the company’s latest disclosure, the retail staff is more racially diverse than the staff overall. Meanwhile, the tech staff is less racially diverse than the staff overall.

CORRECTION: In an earlier version of this item, the Asian and Latino U.S. population numbers were swapped. Thanks, George Row, for catching that.

11 Comments

  1. Fred Stein said:

    Their footnotes shed light. For Amazon and Apple respectively:
    ** A note on the relatively high percentage of employees of color at Amazon: According to a report on Amazon by the Institute for Local Self Reliance, “Amazon [warehouse worker] wages were an average of 15 percent below the wages for comparable positions. … These low wages disproportionately affect African-American and Latino workers, who comprise 45 percent of Amazon’s warehouse workforce, but only 8 percent of the company’s management.” 29

    *** Apple’s retail employee base is included in its overall numbers. In 2014, Apple had about 66,000 employees in the U.S., including 30,000 U.S. retail employees.30 According to the company’s latest disclosure, the retail staff is more racially diverse than the staff overall. Meanwhile, the tech staff is less racially diverse than the staff overall.

    2
    February 8, 2017
  2. George Row said:

    At first glance the figures for “Asian” and “Latino” in the US population as a whole, seemed to me to be the wrong way around. So I looked up some alternative figures. The US Census make the point that these are overlapping categories. (For example people have multiple heritages and some Spanish speakers would class themselves as both “Black” or “White” as well as “Latino”.)

    Nevertheless, if you look at Table 1 in this document:
    http://www.census.gov/prod/cen2010/briefs/c2010br-02.pdf
    from the US Census of 2010 the overall percentages – adjusted to fit the categories used in the Open MIC study – are, something like:
    – White: 63.7% (61%)
    – Hispanic or Latino: 16.3% (15.6%)
    – Black: 12.4% (11.9%)
    – Asian: 4.8% (4.6%)
    – Native American,Alaskan, Hawaiian: 1.1% (1%)
    – Other: 6.2% (5.9%)
    Yes this has a total of 104.5% – so the numbers in brackets are the normalised against that figure and total to 100%. Interestingly the in the normalised figures, 61% White and 12% Black match across these figures and the Open MIC figures. If you swap the Latino and Asian there is a match ±1%. So I think the “Asian” and “Latino” categories have been swapped. The question is was the swap done by the OpenMIC report or by PED?

    A quick glance at this chart:
    http://breakingthemold.openmic.org/OPE001_Workforce_Diversity_by_Race.pdf
    from the report, makes me think that it is you PED who has mixed up the columns or perhaps you took them from somewhere else in the report.

    2
    February 8, 2017
    • Doublechecking… You are right, George. It was I who swapped the columns. Should be correct now. Thanks for catching that.

      1
      February 8, 2017
    • Fred Stein said:

      Good points George. And in Silicon Valley, today, new workers, especially in the tech companies have parents to two distinct ethnic groups.

      0
      February 9, 2017
  3. Gregg Thurman said:

    Skills gravity to proportionately to pay scales and one’s desired geographic area. Programmers make more in Boston, Austin, Seattle and the Silicon Valley than they do in Denver, Miami, Atlanta or Chicago. Skills within ethnic groups follows the level of importance that ethnic group, as a whole, places on education.

    Reports like this are near worthless, because they report end result, and not the many factors that cause them, instead subtilely implying that fo-profit businesses will make personnel decisions not in the best interest of the business. While it may be anecdotally true, it is not the norm.

    In highly competitive industries requiring high levels of education and/or skills you recruit talent wherever you can find.it.

    0
    February 8, 2017
  4. Tom Wyrick said:

    Either it is an optical illusion, or the bars don’t add up to 100%. For example, the USA labels add up to 99% plus a grey segment without a % label that appears about 5%. Total = 104%. (The extra 4% are unregistered voters!)

    Amazon and Apple seem tied according to the labels on bar segments, but Amazon has an “other” segment that appears to put it over 100% while the “other” segment for Apple is too short to reach 100%.

    There are statistical measures of diversity (or its absence) that would shed far more light on the subject than bar charts, even if the bar charts were accurate. As one example, square the % shares for each company and sum them. A perfectly non-diverse company (100% of one race) would have a score of 10,000, while a company with 100 different cohorts each with a 1% share would receive a score of 100. Naturally, all companies will fall between 100 (most diverse) and 10,000 (least) on this scale, but it would provide a quantitative basis for comparison.

    0
    February 9, 2017

Leave a Reply