In the first-ever empirical study of the question, I show evidence that larger DEI bureaucracies often hurt and almost never help the speech climate on college campuses.
Good to see some quantitative analysis of the effects of DEI on campus climate. As far per the stated goals of DEI -- this is an Orwellian lie. If we analyze what they actually do in terms of concrete policies and activities, it becomes obvious that a more accurate operational definition of 'DEI' is Discrimination, Entitlement, and Intimidation. This should not come as a surprise of one considers the theoretical foundations on which this apparat operates -- postmodernism, critical theories, and ultimately plain old Marxism.
DEI bureaucracies are categorical enemies of humanism and liberalism. We must purge our universities from them:
"To state this in more concrete terms, DEI bureaucracies appear to shape the campus speech climate not by stigmatizing ideas like “Black Lives Matter is a hate group” and “Transgender people have a mental disorder” but, instead, by normalizing ideas such as “Religious liberty is used as an excuse to discriminate against gays and lesbians” and “White people are collectively responsible for structural racism and use it to protect their privilege.”"
Note that the ideas normalized both stigmatize conservative perspectives and are directly related to conservative expression!
Excellent point! I'm going to use this insight in my updated version of this analysis (which uses the 2023 FIRE survey, a larger sample of universities, and a more sophisticated modeling approach to show the same pattern of results).
KW, I think FIRE should normalize their prompts. For example, they could make sure that all prompts have roughly the same percentage of support from the general population. I have a similar comment about their "attacks from the right" and "attacks from the left" classification scheme. Often (but not always) an "attack from the right" looks like the public getting angry because a professor says something very extreme like "the US deserved 9/11" or "the Hamas attack was legitimate resistance." In contrast an "attack from the left" looks like students/administrators attacking a professor for saying something most people support like "admissions should be done based on merit." I've suggested to them that they normalize based on percentage public support for a position, but it hasn't happened. -Dorian
Thanks! I'm flattered. Your writing (both here and elsewhere) was one of the reasons I started doing this.
I have an updated version of this analysis that expands the number of universities considerably and uses a more sophisticated statistical analysis (the aggregated approach presented here works well for telling the story but academic publication will likely require a multilevel model that accounts for the clustering). I will be polishing that up and submitting it somewhere soon.
r2 just equal r x r, so who cares? FWIW r is more easily translatable into an intuitive effect size, see the Binomial Effect Size Display. And r=.50 most likely has a p-value below .0001, which is pretty obvious to anyone familiar with data.
Good to see some quantitative analysis of the effects of DEI on campus climate. As far per the stated goals of DEI -- this is an Orwellian lie. If we analyze what they actually do in terms of concrete policies and activities, it becomes obvious that a more accurate operational definition of 'DEI' is Discrimination, Entitlement, and Intimidation. This should not come as a surprise of one considers the theoretical foundations on which this apparat operates -- postmodernism, critical theories, and ultimately plain old Marxism.
DEI bureaucracies are categorical enemies of humanism and liberalism. We must purge our universities from them:
https://hxstem.substack.com/p/fighting-the-good-fight-in-an-age
https://freeblackthought.substack.com/p/dei-colleagues-your-anti-semitism
Excellent article, thank you
"To state this in more concrete terms, DEI bureaucracies appear to shape the campus speech climate not by stigmatizing ideas like “Black Lives Matter is a hate group” and “Transgender people have a mental disorder” but, instead, by normalizing ideas such as “Religious liberty is used as an excuse to discriminate against gays and lesbians” and “White people are collectively responsible for structural racism and use it to protect their privilege.”"
Note that the ideas normalized both stigmatize conservative perspectives and are directly related to conservative expression!
Excellent point! I'm going to use this insight in my updated version of this analysis (which uses the 2023 FIRE survey, a larger sample of universities, and a more sophisticated modeling approach to show the same pattern of results).
KW, I think FIRE should normalize their prompts. For example, they could make sure that all prompts have roughly the same percentage of support from the general population. I have a similar comment about their "attacks from the right" and "attacks from the left" classification scheme. Often (but not always) an "attack from the right" looks like the public getting angry because a professor says something very extreme like "the US deserved 9/11" or "the Hamas attack was legitimate resistance." In contrast an "attack from the left" looks like students/administrators attacking a professor for saying something most people support like "admissions should be done based on merit." I've suggested to them that they normalize based on percentage public support for a position, but it hasn't happened. -Dorian
Maybe 20 years ago, a friend gave me a big button to wear at meetings. It read “We’ve got charts and graphs to back us up so fuck off!”
Not all of the visuals in this article are particularly clear. But admittedly a few are thought-provoking.
Hey, I just saw this. Great post! You planning to publish this in an academic journal?
Thanks! I'm flattered. Your writing (both here and elsewhere) was one of the reasons I started doing this.
I have an updated version of this analysis that expands the number of universities considerably and uses a more sophisticated statistical analysis (the aggregated approach presented here works well for telling the story but academic publication will likely require a multilevel model that accounts for the clustering). I will be polishing that up and submitting it somewhere soon.
Not sure about you stats
r=-.50, p=.00 ?
Why not r2?
And your p needs more sig digits
r2 just equal r x r, so who cares? FWIW r is more easily translatable into an intuitive effect size, see the Binomial Effect Size Display. And r=.50 most likely has a p-value below .0001, which is pretty obvious to anyone familiar with data.