Ideas for (graded?) exercises

[briatte]

• There are sessions without exercises: Week 6 (association), and Week 11 (classification).
• The exercise for Week 10 (logit) is a follow-up to Week 8 (linear regression), but leave it that way.
• The exercise for Week 12 is a no-code exercise.

1. 'secret weapon' on births and female educ.

• have students start with average over years
• use QOG, or ask for students to assemble the data themselves
• use a t-test instead of linear regression, and make it the exercise for Week 6?

2. regression on survey + PCA on aggregated vars

• Use regional data?
• Make it exercise for Week 11?

3. complex data wrangling with EES 2019 / PartyFacts

• aggregate country/party-level answers
• complex merge of DTA/CSV (CSV has party names)
• add a merge to e.g. PartyFacts?
• compute PTV by party family

… when to have that?