Problem
i. Why should you use a multistart optimization search when estimating a "willingness to pay" space model?
ii. Give one reason why we typically compute the log-likelihood during maximum likelihood estimation rather than simply the likelihood.