#ShutDownStem

[fan27]

The Cultural Revolution is gaining momentum. Now the problem is systemic racism in math and science. Soon there will not be any part of American society that is not irrevocably infected with systemic racism.

https://www.cnbc.com/2020/06/10/tho...n-strike-to-protest-systemic-racism-stem.html

 

[gaussian]

Black and Hispanic workers are significantly underrepresented in the STEM workforce. In the U.S., only 9% of STEM workers are black and 7% are Hispanic, according to the Pew Research Center.

Lying through statistics. Almost like they deliberately avoid the question of population size in order to spark unnecessary outrage.

https://www.census.gov/quickfacts/fact/table/US/PST045219

So black people comprise approximately 13% of the population according to the census but have 9% representation in STEM. Hispanic people comprise 18% of the population and have a 7% representation.

Do we really have a problem? I'm confused by CNBCs mental gymnastics here. Can someone who believes this explain to me how you arrive at the conclusion of under representation based on population size?

 

[Cuddles]

Lying through statistics. Almost like they deliberately avoid the question of population size in order to spark unnecessary outrage.

https://www.census.gov/quickfacts/fact/table/US/PST045219

So black people comprise approximately 13% of the population according to the census but have 9% representation in STEM. Hispanic people comprise 18% of the population and have a 7% representation.

Do we really have a problem? I'm confused by CNBCs mental gymnastics here. Can someone who believes this explain to me how you arrive at the conclusion of under representation based on population size?

Methinks the argument is that STEM distribution should be commensurate to the population make up. So when using your pop. stats, they're still underrepresented.

 

[gaussian]

Methinks the argument is that STEM distribution should be commensurate to the population make up. So when using your pop. stats, they're still underrepresented.

Using the numbers from Pew 17.3 million people are in STEM. This would put Black participation at ~1.6M and Hispanic participation at ~1.2M.

According to the census 328,239,523 are in the US.

This gives us approximately 42.7M Black people and ~59.0M Hispanic people.

Let's devise a new measure. Population Adjusted Representation. This number is defined to be [Population in STEM] / [Population of US].

So their population adjusted representation is 4% for Black people and 2% for Hispanic people.

Comparing this to White participation:

There 60.4% of the US is White. Around 198M people. According to the article 69% of the field is White. Though this is slightly disingenuous because "White" is a far less clear definition of race than "Hispanic" or "Black". "Asian" is also ambiguous - did they include Indians?

Anyway, 69% of 17.3M gives us around 11.9M people.

So for Whites, the population adjusted representation is 6%.

Summarizing: Whites have a 6% PAR, Blacks have a 4% PAR, and Hispanics have a 2% PAR.

We could make an argument that Hispanics are marginally under-represented according to PAR but Blacks and Whites have a nearly fair PAR.

If you really wanted to make a fuss that 2% gap could be fixed but 2% is not nearly as bad as this article playing chicken little makes it out to be.

 

[fan27]

The flawed logic is that under representation equates to systemic racism. Perhaps we should have a quota system where any company, university, sports team, governmental agency, etc. has the racial makeup reflective of society.

 

[Cuddles]

Using the numbers from Pew 17.3 million people are in STEM. This would put Black participation at ~1.6M and Hispanic participation at ~1.2M.

According to the census 328,239,523 are in the US.

This gives us approximately 42.7M Black people and ~59.0M Hispanic people.

Let's devise a new measure. Population Adjusted Representation. This number is defined to be [Population in STEM] / [Population of US].

So their population adjusted representation is 4% for Black people and 2% for Hispanic people.

Comparing this to White participation:

There 60.4% of the US is White. Around 198M people. According to the article 69% of the field is White. Though this is slightly disingenuous because "White" is a far less clear definition of race than "Hispanic" or "Black". "Asian" is also ambiguous - did they include Indians?

Anyway, 69% of 17.3M gives us around 11.9M people.

So for Whites, the population adjusted representation is 6%.

Summarizing: Whites have a 6% PAR, Blacks have a 4% PAR, and Hispanics have a 2% PAR.

We could make an argument that Hispanics are marginally under-represented according to PAR but Blacks and Whites have a nearly fair PAR.

If you really wanted to make a fuss that 2% gap could be fixed but 2% is not nearly as bad as this article playing chicken little makes it out to be.

Why is a "new measure" required? You're assigning bias by tying STEM choice to make up of the population with that metric.

 

[gaussian]

Why is a "new measure" required? You're assigning bias by tying STEM choice to make up of the population with that metric.

Because if you don't compare representation of a sub-population to representation of the total population your numbers are necessarily skewed the wrong direction. Imagine a population of X and Y. Xs are 50% and Ys are 50%. If 10% of the total population is in this field you may expect something around 5% of each X and Y to comprise that total under the assumption of equal participation.

Looking at under-representation in the most generous light you can interpret it as "A population at large should be represented at least as much in the sub-population as another population". Moreover, not only is this interpretation generous it's also the only one we can assign a measure to. Every other one is subjective. This measure ignores cultural factors and looks strictly at population-to-population comparison in a sub-population without considering socioeconomic or other factors which are non-quantitative.

The only way to do this is to ascribe another measure. A percentage of the sub-population to total population in order to calculate this interpretation of under-representation. It's not perfect and there are many socioeconomic factors not considered (for example we could create a more complicated model using correlates like STEM education graduation rates) but this napkin math at least gives you an idea of the degree of under-representation which isn't nearly as bad as one would think by just reading this article.

 

[userque]

The flawed logic is that under representation equates to systemic racism. ...

If not some form of discrimination/racism, somewhere in the chain, what are the other possibilities?

 

[userque]

Because if you don't compare representation of a sub-population to representation of the total population your numbers are necessarily skewed the wrong direction. Imagine a population of X and Y. Xs are 50% and Ys are 50%. If 10% of the total population is in this field you may expect something around 5% of each X and Y to comprise that total under the assumption of equal participation.

Exactly. if Y's make up 50% of the overall population, then you'd expect them to make up 50% of the STEM population, all other thing being equal...especially since we are dealing with very large numbers.

 

[fan27]

If not some form of discrimination/racism, somewhere in the chain, what are the other possibilities?

Who do you think is going to have a higher interest in studying math and science?

1. An Asian American who comes from a middle class family whose peer group is interested in math and science and whose parents encourages pursuits in the math and sciences.

2. An African American who comes from a lower income family who has very few if any of his peers interested in math and science.

3. A White American who comes from a lower income family who has very few if any of his peers interested in math and science.

Of course the Asian American in this example is more likely to pursue match and science and I would guess that the white and black american both have the same likely hood of pursuing math and science. Show me any university that is discriminating against minority students...it is usually the opposite. Universities and large companies are thrilled to admit/hire minorities...I have seen this first hand in tech.

EDIT: I bet that if you figured out the racial break down of middle and upper class America, you would find STEM fields represented proportionally based on the racial make up of that population.