Blog/2026-01-17/Citogenesis
Randall Munroe coined Citogenesis to reference the common tactic of putting something in Wikipedia, having reporters use it as a source un-cited, and then use that reporting to cite the Wikipedia article thereby closing the loop and making something seem real.
This kind of circular referencing is not feasible when all references are to other papers because of the fact that papers move only forward in time. In fact, given scientific papers that only reference other scientific papers, one can hope to trace the source of one mistake or the other back to its origin.
But another novel method of citogenesis exists: to reference a newspaper or magazine that does not reference its sources. By doing this, one can take a specific number one wishes to seem scientific, see it published in some magazine, then use a network of papers to launder that into a scientific reference.
This has particular strength now that a lot of people have access to Google and can speak the language of scientists: "citation" / "reference" / "source" and so on.
Here is a sequence someone put together recently to claim that Norwegian recidivism is 20% by comparison with the US's 52%[1].
Let's start with the paper people like to claim for the 20% number:
Denny, Meagan (2016). "Norway's Prison System: Investigating Recidivism and Reintegration". Bridges: A Journal of Student Research. 10 (10). Retrieved 2026-01-16.
If you actually read that paper, it sources Norway's recidivism rate from
Deady, Carolyn W. "Incarceration and Recidivism: Lessons from Abroad" (PDF). The Pell Center for International Relations and Public Policy at Salve Regina University. Archived from the original (PDF) on 2014-03-01. Retrieved 2015-11-29.
This one does not precisely state where its numbers are from, though two of the 'citations' state the 20% number:
Neither of them point to their primary sources. Now, it is entirely conceivable that the comparison is valid. It could entirely be the case that Norway's recidivism rate on the same parameters as the US is much lower. In fact, I think that is likely to be the case.
However, this is not the science that shows that. Through a circuitous path, all that has happened is that opinion articles in magazines and newspapers have been laundered into 'papers' with citations formatted in the style of the field, thereby creating an impression of scientific rigor greater than simply Googling for one's pre-existing beliefs and selecting the respective opinion articles that support one's views to put into one's paper.
By comparison, SSB Norway's report contemporaneous with many of those references shows how the details matter. They illustrate that if you vary the parameters you can get anywhere between 10% and 55%. There is no comparison there, but it does illustrate how a blanket claim about 20% is not meaningful without more detail.
Notes
- ↑ Cross-country statistics are risky to compare blindly. A typical mistake made is to use times where big change has happened (say pre-COVID and post-COVID or pre-GFC and post-GFC). But bigger mistakes can be made: crimes aren't equivalent across nations, and recidivism isn't measured in a consistent manner. Number of years after release matters significantly, for instance.
