Maximum Likelihood Estimation

Do you teach maximum likelihood estimation for some or all of the following reasons?

• It�s the estimate that is most likely
• Other estimates would yield probabilities only 0.997 or even 0.977 the size of the probabilities stemming from use of maximum likelihood estimates.
• If the data were normally distributed�it would be a minimum variance estimate (but then after normality you�ll be telling me there are fairies at the bottom of your garden).
• It allows you to show off your knowledge of the calculus.
• It fits in with your plan to teach point and interval estimation on consecutive days.

����������� Jaques Cuze