Basic Statistics: Understanding Conventional Methods and Modern Insights 1st Edition

Very solid introductory book on using statistics and the proper use of each method. Wilcox often provides the subtle background while explaining the concepts which often makes all the difference in grasping the concepts. PLEASE NOTE that a second book that he wrote in 2016 covers the exact same material as in this book and is just as solid but with applications using R (I have both books). I would purchase that book instead simply because he is very helpful in explaining using R with this data, and if you’re going to start manipulating data using statistics then you need to understand basic scripting in R. That book is Understanding and Applying Basic Statistical Methods Using R.

They need to make it easier to understand

Easy to understand. The author also doesn't gloss over the difficulties of some statistical methods and how modern developments have helped.

The book is smaller than other textbooks therefore there is no burden in carrying it. It simplifies everything and explains it simply.

Good background reference and refresher.

The title here doesn't convey well that this book is distinctive, orprecisely why. It is very much an individual view of what is needed inan introduction to statistics, where "individual" means anything fromidiosyncratic through eccentric to perverse, depending on whether youagree with the author. That is both its great strength and its greatweakness. Here I give 3 stars because the author is not following amultitude to do evil, but trying something different, and that makes hisbook very useful in parts. But I also detail what could be unattractive,or at least not what you might want or need.To simplify, Wilcox's main stance is pretty much that none of thestandard statistical machinery can be trusted without modification, asmeans, standard deviations, t-tests, regression, analysis of variance,and so forth are all based on assumptions unlikely to be exactlycorrect. His main concern is that distributions are likely to be worsebehaved than the common assumption of normal (Gaussian) distributionsimplies: there are likely, for example, to be outliers and heaviertails. That stance alone doesn't differentiate him from moststatistically-minded people, but his precise mix of solutions isunusual. What he suggests is a battery of modified versions of standardtechniques, many involving trimmed or Winsorized means and similartechnology. That amounts to an insertion of techniques often onlycovered in specialist texts on robust statistics into an introductorytreatment. The positive side of this is detailed coverage of methodssuch as trimming and Winsorizing, which are relatively easy tounderstand, and which you might want to consider for your analyses. Ris recommended vigorously as software environment, but typically in theform of the author's own R functions. It often seems that the authorwill recommend strongly only what he has himself invented!Statisticians and other well-informed people are, however, likely todisagree with Wilcox on many details, not they will all agree amongthemselves. Some would regard his approach as too pessimistic, or atleast too complicated at this pedagogic level; some would preferdifferent robust methods to be emphasised (the field is highly fluidwith often considerable disagreement on what to recommend); and somewould want to stress other solutions. For example, transforming data isa standard alternative that is covered well by many introductory texts,yet it is dismissed by Wilcox briefly and quite unconvincingly. Thosewhose analyses (in economics or biology, say) successfully depend onworking on logarithmic or other scales will be surprised to learn thatsuch an approach is recommended against.Furthermore, the author's focus is in some ways one-dimensional.Assuming linearity when relationships aren't linear, or assumingindependence when it does not hold, or assuming lack of trend when thereis one are just some of the other assumptions adopted uncritically inuses of statistics. Naturally, the author can't write about everythingthat might be wrong, but his scope is distinctly limited.In many other respects this text is strikingly old-fashioned, and not"modern" at all.1. There is essentially no sense imparted that many standard statisticaltechniques are all special cases of generalised linear models, nor isthere much inkling that the problems tackled here using robust methodswould often now be tackled in quite different ways by assumingnon-normal distributions. In that respect the treatment is a few decadesout-of-date (and was so in 2009, the original publication date).2. The examples are often feeble. Many datasets are just fake orimagined or far too small to be taken seriously. Frequently there islittle sense conveyed of the scientific rationale for looking atparticular examples. The author often forgets to mention units ofmeasurement, in my experience diagnostic of dissociation from scientificobjectives. There are some striking, even good, exceptions, but too few.3. Use of graphics is weak. In too many cases, even for extendedexamples, the data being analysed are not plotted. When they areplotted, the graphs are often poor. More generally, exploratoryapproaches to data analysis are not much in evidence.Would this be a good book for you? Calculus is not assumed, but thatapart, this isn't a book for those uncomfortable with mathematics: it isassumed throughout that you are relaxed about handling notation andreading formal statements. A test case is that absolute value notationsuch as |x| is used without explanation; if that's an old friend, youshould be fine.The text would have profited from more careful editing andproof-reading. If you follow Strunk and White and dislike "compared to","centered around", or "prior to", then you would think less of thisbook. Independent clauses are frequently strung together with commas,see if it annoys you. Either way, there are many trivial typos.My guess is that this could be useful to you if you have somestatistical knowledge and experience and are confident at striking outon your own and interested in learning about the robust methods itcovers. I would be surprised to see this adopted formally in any coursesbut the author's, unless as recommended wider reading.

Over the many years I've spent teaching or tutoring undergraduate students in statistics for various sciences (social, behavioral, managerial, nursing, etc.), I've had to work with many, many introductory statistics textbooks. Some I've found to be better than others, but I believe they are all unnecessary as Dr. Wilcox's textbook is quite simply the best statistics textbooks for a wide-range of students. It requires almost no mathematical literacy yet clearly elucidates the essentials of statistical reasoning beyond the level of many a denser, more computationally demanding text. Additionally, Dr. Wilcox manages to incorporate a central component found in his other statistics textbooks: modern developments. The vast majority of introductory statistics textbooks continue to teach methods that were introduced as compromises to ideals in the first half of the 20th century and to make assertions about the most common statistical tests that were long ago shown to be false. Dr. Wilcox ends every chapter with a section of modern advances and insights. As the use of statistical software packages such as SPSS has increased, statistics textbooks have increasingly become more akin to manuals for using such software to run tests rather than covering their underlying logic. Even those textbooks which do emphasize the logical basis for the methods taught frequently perpetuate fundamental inaccuracies that lie at the heart of virtually all the covered material. Dr. Wilcox's text not only covers superior & modern methods in a clear, easily grasped manner, but also introduces the student (or casual reader) to the problems inherent in many common techniques and why the alternative methods introduced are improvements.Whether one is an undergraduate instructor looking for a textbook, a graduate student taking multivariate statistics having forgotten much of their introductory statistics class, or the average layperson interested in learning statistics but daunted by the mathematics, this textbook is ideal.

This book is certainly one of the best intro stat books. There're lots of examples in the book and the content is unique. It covers many modern statistical techniques. Really worth reading.