updates to Occam's exercises

[hawkrobe]

• add a 'warm-up' question asking for discrete Big/Small model in chapter to be tested with continuous observation from differently-sized uniform distributions
• causal support question part C needs to explicitly say the dataset to run it on (i.e. the 8/8 + 8/8 $$Delta$$P dataset) and have them compare it to the 2/2 + 2/2 dataset to get the point that causal support changes inference as more data comes but causal power doesn't. Also need to explicitly ask to interpret the results of both the causal support model and the 'original' causal power model.