Stop trusting the literature review: trace the citation telephone
MethodologyComments
It is usually the latter. We saw this with the early misuse of p-hacking in psychology, where "significant" results were cited as absolute truths for decades before replication crises hit.
But what about when you're working across five different disciplines... isn't it almost impossible to be the primary expert on every single root source? how do we decide which citations are "critical" enough to trace back?
Maybe we can look at the rise of open-source annotation tools. They might allow researchers to flag these "telephone" errors in real-time for others to see.
This is exactly why local health guidelines often contradict the newest papers. By the time a finding hits a government white paper, it has been filtered through three different committees and lost all its nuance.
Given how those white papers are written, do you think the simplification is intentional for the sake of "clarity," or just lazy?
I disagree that the committees are the main cause of the nuance loss. Usually, the distortion happens in the initial abstract or a high-impact review paper long before it reaches the policy level.
The upside here is that this realization is driving the "Open Science" movement. We are seeing a shift toward preregistration and raw data sharing, which makes auditing these citation chains much more transparent.
the 2016 study on citation bias showed a significant percentage of papers cite others without ever reading the original.