If a social constructionist is a realist who’s just had their narrative dethroned, then we are all social constructionists now.
Donald Trump made a big deal out of the gap between FBI crime statistics and the Bureau of Justice Statistics’ National Crime Victimization Survey. Of course, he thinks the latter is better, as the diagram below shows:
But the Marshall Project, who put together this graph and analyzed it, emphasizes that the anomaly merits attention. It’s not a new anomaly, but the 2022 gap is the largest recorded by far. There is no obvious explanation for the jump: post-COVID crime (and crime report) “recovery,” the slow and inept transition to new reporting systems, crime report revision, and other factors could all play a part. The Marshall Project concludes that there’s not much reason to conclude there has been a true spike in urban crime, but also that the information is too patchy and inconclusive to explain what is going on in that graph above.
The disturbing trend here is not crime per se (at least, not to our knowledge), but uncertainty. You may remember when Germany and Italy reported drastically different death rates from COVID, Italy’s fatality rate four times that of Germany’s. One factor turned out to be that Germany was much more stringent in assigning COVID as cause of death, while Italy was more likely to classify anyone who died who happened to have COVID as a COVID death. That distinction never showed up in the summary statistics, making Germany look mysteriously more robust than Italy and feeding dozens of speculative narratives about what Germany was doing right and Italy was doing wrong—or, contrariwise, virulent skepticism about why the numbers couldn’t be trusted. Neither option was correct, but the best path forward—sifting through the sourcing of the data to understand the disparity—was a comparatively tedious and delayed auditing job that could hardly hope to gain attention in an online media environment that needs statistics yesterday.
What results is the creation of multiple metarealities, projected views on to the too-complex leviathan of life and data before us. And where in the 20th century, the scarcity of media channels tended to prune these metarealities evolutionarily until we were stuck with one or two that came to seem more or less objective (even if we continually found out later on where they were wrong), there’s no such forcing function any longer. For every graph like the one above, you can happily stick to the subset of interpretations that feed your prejudices and even “Fact Checks” will be unlikely to help, since even a deep dive by the Marshall Project could only muster a “Reply hazy; try again later” rather than a definitive interpretation. That level of skepticism doesn’t satisfy.
Given the ever-increasing amount of data collection and data analysis, the results of which are then further promulgated through aggregation and AI-summary, these anomalies are going to multiply rather than resolve, because the generation of these ambiguous metarealities takes far less time than the auditing of them. When something as seemingly concrete as crime becomes up for grabs, people fall back on intuition, superstition, prejudice, and anecdote—hardly a new tendency, but one against which we are losing tools to fight.
I have described this as the loss of consensus , which I define as "What everyone thinks everyone else thinks". Cronkite expressed the consensus, and not everyone actually agreed, but everyone thought that was the general opinion.
And the consensus was really ragged and mushy, not something that could be clearly stated, because everyone had somewhat different ideas about what the consensus was, and committment to the consensus belief varied widely and was unstable.
Metareality: I say mediated reality, where the important things in our life that we know directly are few and the things that we know through a hierarchy of imperfect intermediary sources (or not at all) are many. For a fictitious example, with the global economy interdependent the way it is, a big spike in the price of gallium might throw someone out of work , whilet their possibility of understanding what happened would depend on mediated sources which might not exist at all and might well be dishonest.
I use "mediated"