ABSTRACT
In this chapter we look at a class of estimators known as maximum likelihood estimators. There are two main reasons that this is important. First, maximum likelihood estimators are very common in empirical work. If you have done any empirical research before, it is likely that you have used a maximum likelihood estimator – even if you didn’t realise you were doing so! Understanding the concept of a maximum likelihood estimator – and the relative strengths and weaknesses of those estimators – is therefore very important for doing empirical work, and for interpreting others’ empirical results. Second, maximum likelihood estimation is very flexible; it can be adapted to many different kinds of model, in many different empirical settings. It is therefore very important if we want to tie a theoretical model directly to an empirical estimation. This is known as structural modelling, and we discuss this in Chapters 19 and 20.
