Xi'an's Og

One Response to “the polls weren’t wrong [alt book review]”

1. 

   Carl Allen Says:
   January 27, 2026 at 5:59 pm

   Thanks for sharing the alt-review. I agree the book was longer than it needed to be (for especially educated readers) but the book was intended to be for all audiences – not only experts. I am also mindful of fair criticism (like overt use of terms like lying). While it may come across as biased or conspiratorial, it’s language I chose intentionally, for the reason I stated in the book, paraphrasing:

   “If not for established traditions that says undecided voters can be allocated, with the allocation reported as the observation, this practice would be considered scientific misconduct” – and I’m right.

   Can you think of ANY other scientific field which collects data, observes a confounding variable, reports the data as if the confounders don’t exist (by projecting it) AND proceeds to calculate the accuracy of the measurement as if no projection had occurred?

   In my research, from epidemiology to sports sciences to economics, this practice is always flagged and rejected.

   Polling is the only domain I’m aware of where:A known confounder (undecided voters), is

   • Resolved via convention
   • Silently absorbed into the reported estimate
   • And that adjusted estimate is later scored as if it were a direct measurement

   You can argue it’s not “lying” in the moral sense – but it would meet the textbook definition of methodological misrepresentation in any other field. This is worse – the term was chosen for its appeal to the average reader. But I do believe characterizing it directly would fare better in my future work.

   In any case, I hope you’ll read the book (Christian) before writing reviews in the future.

   Reply