Let’s assume you own a modeling agency. You’ve just discovered a beautiful new model, who seems to have everything. She’ll be a star, the toast of the fashion show, modeling all the latest gowns and swimsuits. And you will be a star too, because you discovered her. Yet, when the fashions start coming in, lo and behold, they don’t fit. Nothing fits the beautiful new model. What went wrong?

Well, we can’t speak for the fashion industry, but in statistics when the real-world data doesn’t fit the model, you may have what’s called overdispersion. It occurs when the actual results vary more than those of the model, and it’s said that “overdispersion is a rule rather than an exception.”  Now there is a guide to overdispersion specifically for the SAS world. Jorge Morel and Nagaraj Neerchal, both long-time SAS users from the fields of industry and academia respectively, have just published Overdispersion Models in SAS.

With a combination of theory and methodology, real world examples and working SAS code, the authors lead you through the thorny jungle of overdispersion, covering generalized linear models, likelihood models, binomial models, multinomial models, and much, much more.  Throughout, they offer explanations and, most importantly, guidance. In their nearly 400 pages, you will find the tools for both understanding and dealing with the mysterious but ubiquitous phenomenon of overdispersion.

Want to learn more? Read a free chapter or buy a copy today.