[1]. I'd transform degree to create our treatment variable d. What would you do and why? [2]. Build a model to predict d from x, our controls. Comment on how tight the fit is, and what that implies for estimation of a treatment effect. [3]. Use predictions from [2] in an estimator for effect of d on loan. [4]. Compare the results from [3] to those from a straight (naive) lasso for loan on d and x- Explain why they are similar or different. [5]. Bootstrap your estimator from [3] and describe the uncertainty. [+]. Can you think of how you'd design an experiment to estimate the treatment effect of network degree