46 Ordered Regression Models
The MLE method requires the construction of the “likelihood” or “likelihood func-
tion.” Many researchers confuse likelihood with probability. Probability refers to the
chance of observing an event or series of events given the population parameters, and
the likelihood is proportional to the probability of the population parameters given
the sample data. Because MLE typically involves maximizing a function based on
population parameters, the parts of the likelihood function that do not contain these
parameters (e.g., constants) are usually dropped. It is in this sense that the likelihood is
proportional to the probability. MLE involves maximizing the natural log of the likeli-
hood function because it is a mon