Hi!
Right now cpt_pop_function() assumes uniform probabilities for age (1–80) and splits sex proportion using a single slider. In practice, age and sex distributions are not uniform in most populations and this assumption can bias likelihood ratios. A possible improvement would be to load a small demographic table (e.g., census-like data) so the age prior reflects real frequencies and sex ratio can vary by age group. The app could default to the current uniform setup but include an option like “Use demographic priors”. This would make the model statistically more grounded without changing the UI flow or CPT structure.
Hi!
Right now cpt_pop_function() assumes uniform probabilities for age (1–80) and splits sex proportion using a single slider. In practice, age and sex distributions are not uniform in most populations and this assumption can bias likelihood ratios. A possible improvement would be to load a small demographic table (e.g., census-like data) so the age prior reflects real frequencies and sex ratio can vary by age group. The app could default to the current uniform setup but include an option like “Use demographic priors”. This would make the model statistically more grounded without changing the UI flow or CPT structure.