Thursday, February 16, 2017

Comments on Diversity report

A news article on a study about diversity efforts got me into a discussion with a proponent. He argued I needed to read the report, which I did. I am offering the bullet-point commentary here in order to facilitate the discussion.


1. Correlations - not causation
2. Does an all minority firm out perform?
a. None available to determine
b. yes: then the marketplace will deal - more investment in out-performing companies
c. no: is there a tipping point? vs the current level?
3. interior Culture and performance demands different than existing exterior culture norms
4. Affirmative action impact on retention - ie no AA standard in performance kills AA standard in initial hire
5. 170 firms offering info are self-selected. Those with positive results tout, those with negative don't
6. I ABSOLUTELY reject % of population to % within company. failure by absence is not evidence of failure
7. Tech sector is more oriented towards technical competence than personal interaction (personal experience too)
8. Google is going after more STEM support in schools - I absolutely support - but not via affirmative action
9. Knowledge base failure within study: limited data and evidence. as Econ grad, I understand the limits of such a data pool
10. Reports rely upon 'feeling of what seems to be working'
11. A significant study of 700 companies = no positive effect on diversity efforts and may hurt Black women
12. "...efforts to isolate impact of female leadership was inconclusive" BUT LOOK!!! 9% HIGHER VALUATION!!
13. Due to minority population size, areas/fields that aggregate significant minority representation removes diversity from elsewhere
14. Women in STEM has decreased as the # of 'programs' in sociological studies increased...related?
15. What the study calls bias I call animosity - not of hatred but of 'not what I have seen work' bias via work experience (arguably false sometimes.
16. Diversity training = no positive effects so....
17. ...switch to 'unconscious bias training' - but may have same longterm results - NONE. Recall Hawthorne Study.
18. pointed out WIPRO as example - also in my business plan as a foreign competitor
19. regression analysis bar for significance = .10 this is a low bar though not 'significantly' outside the norm. But the study never indicates how FAR over that bar the results were. .11, .15. These are 'noticeable' however, given the sample size, NOT USEFUL.

All in all, the study takes a minor blip and runs like the wind with it.

Other issues. 70% of Black children are born out of wedlock, 60% of Hispanic. And these are long term trends. These populations are NOT going to create a pool of STEM candidates for the industry as a whole. Only the top 1% or less will reach excellence levels necessary for prominent positions in corporations. By the measure of calculating populatio = 30% likely to succeed, you get the 3.5-4% representation of Blacks in the tech workforce.

Diversity in corporations doesn't change culture and diversity in culture is NOT beneficial. Understand I used the term CULTURE.

No comments: