If you were only stuck with either statistical significance tests or algebraic models the removal of which one would harm science progress more?
It is possible to do statistical significance tests like t tests to see if two groups are statistically different without algebraic models that predict a output for a given input
Historically people used algebraic models with no statisical significance tests until they were recently invented
The two do not necessarily go together
newton worked for royal mint and made sure the foundry workers werent doing the "sawdust in bread" fuckery when making currency?
i would say the scientific method is more important just validate making new theories another way instead of statistics i dunno lol