Good methods of estimation will have the properties of accuracy (low bias), efficiency (low variability) and consistency. Comparison of maximized log likelihoods under different models may be used as an indication of model suitability. For cases where maximum likelihood is difficult computationally and where only uncensored data are available, a simple, fairly efficient method of estimation is provided by equating population and sample moments – the method of moments. Most of the general goodness-of-fit tests apply to grouped data or where data are compared to a completely specified distribution, that is, the parameters are not estimated from the data but given hypothesized values. There are a number of goodness-of-fit tests on the general lines of that due to Kolmogorov and Smirnov. It should be emphasized that maximum likelihood estimation identifies the member of a particular model family most appropriate to the data, but does not on its own demonstrate the suitability of the model family.