17 Conclusion

In this book we have covered many topics that comprise an introductory and first course in statistics. These topics and methods form the basics of the discipline, and all higher level courses depend on this material. We have covered the modelling chapters in a generic form that allows the material to be easily extended to more advanced models. While the examples and problem sets are more science-based, the material is applicable to any discipline.

We hope that this book has given the student a good basis on which to build their further knowledge of statistics. We further hope that the student has realised that there is a lot more to statistics than what this book covers. But the knowledge gained from this book should allow the student to talk more confidently to a statistician, ask more informed questions and also make more sense of the answers and ask further questions.

We would advise the student who will need statistics in their discipline to take further courses in statistics. In particular, this book covers models where the response variable is continuous. These are called normal-based models. If the response variable is binary or count data, as is often the case, then the models we have covered do not hold. These require what are called Generalised Linear Models, or Log-linear models. Further, if the assumptions of the linear model are not satisfied, then again the models we have covered do not hold. In particular, if the residuals are not normal, or the homogeneous variance assumption is not satisfied, or the responses are not independent, then we need more advanced models.

The UWA team wishes you success in your pursuit of statistics.

R. Nazim Khan

May 2024.