When the real world refuses to conform with your ideological assumptions

What a dog's breakfast of ideology, unstated assumptions and capacity to bury the lede. From Bursting people’s political bubbles could make them even more partisan by Carolyn Y. Johnson.

There has been a strain in academia and elsewhere on the left which has always assumed that anyone on the right is both ignorant and irrational. That whatever assessments which they might make are based on ideology and malice. The assumption has always been that you could overcome that ignorance by exposing them to more information. That usually doesn't work. Then the assumption was that if you force them to interact with the "other" that would work at reducing animosities or pre-existing biases. It is now becoming apparent that that also does not reliably work.

These are not innocent assumptions free of consequences. All sorts of public policy have been built on these untested assumptions.

Johnson is reporting on some of the research that is now emerging that indicates that sometimes forcing people to interact with one another is actually counterproductive.

There are many strands coming together here. Reading the article, you see time and again sociology/psychology research and experiments which have all the characteristics which have led to the replication crisis which has brought down most of the exciting findings of the past three decades. Experiments which are designed with unintended ideological biases in them, too small test populations, a preference for lab research over real world observational studies, manufactured scenarios over natural scenarios, non-randomization of populations, absence of methodological rigor such as double-blinds, weak measurement approaches and weak effect sizes. Basically, left leaning academics are proving to themselves that their unstated assumptions are probably true.

The more rigor you build into the research the more the findings prove to be untrue.

Forcing people to interact with their "other" does not inherently make them more tolerant or change their knowledge base. It is very context dependent. Stereotype accuracy is one of the confounding strands. Trying to strip people of a stereotype which is actually accurate is not going to occur by exposing people who hold that stereotype to more people who conform to that stereotype.

Johnson spends a lot of time veering about, covering various postmodernist researchers who are finding problems with their assumptions. There is a lot of topical bleating about polarization and social media and sociological research which does not have empirical support.

The lede which is buried is that academia has been willing to treat untested ideological positions as true without testing those beliefs. They are willing to do this for decades at a time. They are willing to advocate for policies based on those untested beliefs. They are happy to force onto their fellow citizens policies inimical with those citizen's own beliefs and then are confounded when the real world outcomes are not what was expected and may have even been counterproductive.