Statistical Methods for the Information Professional: A Practical, Painless Approach to Understanding, Using and Interpreting Statistics

Citation

Ole Pors, N. (2003), "Statistical Methods for the Information Professional: A Practical, Painless Approach to Understanding, Using and Interpreting Statistics", New Library World, Vol. 104 No. 7/8, pp. 324-325. https://doi.org/10.1108/03074800310488121

It is not an exaggeration to state that only a minority of library and information science students really yearn for a module or course in statistical analysis. This makes it very important to produce good introductory textbooks with examples to which the students can relate.

The book has approximately the same size as the excellent Simple Statistics for Library and Information Professionals, 2nd ed., London, 1997, by Peter Stephens and Susan Hornby.

Without making this a comparative review, allow me to state some of the qualities and features of Stephens and Hornby’s book. It contains chapters on probability, index number, time series and inferences from proportions. It has a lot of questions and exercises and answers to the practice questions. The book has many and diversified short examples from information work and it presented formulas through the text. It is in many ways a stand‐alone workbook for the student.

The aims of Vaughan’s book are not that different but the approach taken is different. In the preface, the author emphasises that the book introduces the logical reasoning instead of mathematical deduction to explain statistical concepts and procedures. Vaughan also emphasises the approach to formulating hypotheses, the use of computer software to analyse and the pertinent interpretation of the output. Third, she aims to reduce the number of formulas and technicalities explaining the usefulness of spreadsheets and statistical packages.

Overall, I think the author is very successful in achieving the aims. There are relative few formulas, figures and tables in the text. Most of the pages consist of letters. This is unusual in an introductory text in statistics.

The book contains 13 chapters covering the traditional topics like types of data, descriptive statistics by numbers, charts and tables. It has also chapters on the traditional inferential procedures including a chapter on sampling problems. This chapter includes a precise and short discussion of Web‐based surveys. The book introduces the traditional t‐tests, chi‐square tests, different measures for correlation and different non‐parametric tests.

It is more unusual to see chapters on ANOVA – procedures, multiple regression and two‐way ANOVA in an introductory text. The author succeeds in explaining the logic and fascinating features of these topics in a way that is easy understandable.

One of the strengths in the book is the discussion about the assumptions embedded in the single test and test procedure.

The book starts off with a short chapter on different data types and levels on measurement and it continues with a chapter on spreadsheets and statistical software. The next two chapters are on descriptive statistics like graphical presentations, measures of central tendency and measures of variability. It is good to see that it is possible to present the fundamentals in a concise way on relatively few pages. The author emphasises the relationship between the graphical presentation and the numerical presentation of data sets. The one thing I miss in these chapters is discussions about how far, in practice, you are allowed to deviate from the ground rules, e.g. presenting ordinal data, as do most questionnaire surveys, as if they were ratio data. This is one question that many students ask in a classroom situation because they have seen the discrepancy between practice and theory in this respect.

The rest of the book is about inferential statistics and associations of different kinds. Vaughan does a very good job explaining the underlying logic of the single tests and measures of association. She gives in every chapter a good description of the assumptions of the test and the requirements for using the test. The text is readable and you can actually sit in an armchair and follow the argument. I like the inclusion of ANOVA tests as part of the basic statistics. She also introduces some of the better‐known non‐parametric tests like Mann‐Whitney and Wilcoxon Sign test. It is probably not a good idea to demonstrate the Mann‐Whitney test using ratio data such as salaries.

The book includes a very useful road‐map for using the most appropriate test. This road‐ map goes very well together with a lucid discussion about the advantages and disadvantages of the different measures of associations and tests.

The library and information science literature often includes the use of statistics. I think it is just to say that the library and information science literature to a certain degree contains too simplistic a statistical analysis. By simplistic analysis I simply mean that some authors do not use multivariate methods when appropriate. The book by Vaughan can help remedy this situation. She has a very good introduction to some of the more advanced statistical methods like two‐way ANOVA, multiple regression and the LISREL model. Her interpretation of the SPSS output is illuminating for students and her introduction to the concept of partial correlation coefficients is also very precise. One could have wished for a deeper analysis of these so‐called advanced methods. Anyway, the basic principles get a good presentation.

It is evident from this review that I like the book very much and recommend it as one of the best introductory texts around at the moment. I am not sure it can stand alone in teaching statistics. It presents the logic of statistical reasoning very well but students have to experience the many complexities of real life analysis using other means. It could be a guide to a statistical package or a book containing exercises.

Teaching my courses in research methodology, this is a book to which I will return.