# Review of the book "Laborstatistik für technische Assistenten und Studierende " (Laboratory statistics for technical assistants and students)

In the meantime, I’ve read - still enthusiastically - another "essentials" booklet by Patric U.B. Vogel. In today's post, I’d like to introduce the book "Laboratory Statistics for Technical Assistants and Students" (2021 published by Springer Spektrum, Wiesbaden; ISBN: 978-3-658-33206-8) and provide my personal feedback.

## About the content

This book of 70 pages including references provides a good overview of the statistical methods most used in different biological / chemical laboratories. It is aimed at staff from laboratories in the academic environment as well as pharmaceutical quality control. The introduction explains the context, why and for what statistics is / can be necessary and introduces the subsequent chapters. In the next chapter, the reader learns about the different types of data and the parameters of descriptive statistics, as well as common sources of error. The explanations are starting completely from scratch, which is absolutely fine here. In the next chapter, rather "unusual" things, i.e. techniques not used in every laboratory, such as testing for outliers, normal distribution and sample calculation, are discussed. The presence of a normal distribution is a prerequisite for the statistical tests to compare data sets using the t-test or ANOVA explained in the following chapter. The book concludes with a chapter on linear regression and correlation to be able to assess potential connections. We also learn about the purpose of confidence intervals and how to calculate them. Furthermore, at the end, the content is summarized and rounded off by the bibliography. I guess, I don't need to mention that the proven concept (getting started and exit pages with bullet points, introduction and summary, super illustrations, and bold keywords) has also been applied to this book ????.

## Pro's and Con's

First of all, in this booklet it is extremely noticeable - compared to the previous books already read - that there are significantly fewer spelling mistakes. As always, it is written in a super-comprehensible writing style and throughout the whole book, **how** something is calculated is explained excellently (with the help of pictures, among other things). I also like the fact that an example of a weight determination of hormonally treated laboratory rats, chosen right at the beginning, is used throughout the entire book, and is taken up again and again each time with a new perspective for different aspects.

In addition, I like that the topics are written for students in laboratories of the academic environment as well as for technical assistants or laboratory technicians in pharmaceutical laboratories and that both points of view are always considered. In this context, the selection of topics is also very well done in my opinion.

I liked very much - as in the last review - the critical view of the author, when it is mentioned, for example, that in the academic field the determination of the sample size is often subjective, axes might be glossed over to dramatize, or data manipulation can be practiced when using outlier tests. Equally critical is the self-selected example, where it is clearly stated that the sample size chosen for the example was too low and only by luck a significant difference between the two groups could be demonstrated.

As mentioned at the beginning, the description of HOW the Dixon test for outliers is calculated is excellent, but I missed the explanation of WHY this particular test is used. A little more background on the advantages of this test over others would have been nice as well as an indication of up to what sample size it can be used. Speaking of backgrounds... These were also missing to understand why I should use a Shapiro-Wilk test to check for normal distribution of a data set with a sample size of 50 or why a value of 0.2 is often assumed as β-error or why I should only enter a positive but not a negative error value in the bar charts... Unfortunately, it is also somehow clumsy that although the backgrounds to the hormone therapy of the rat example are super explained at page 36, but that they are already anticipated on page 28 and at that time, the reader does not yet know anything about the hormone therapy on which the experiments are based... In my opinion, however, the explanation of the background for the "how & why" of the T-test and the mini-introduction to the F-test for the purpose of selecting the correct type for the T-test was very nice. Compliments!

I also liked, as already with the book about Trending, that non-parametric statistical methods for non-normally distributed data were not discussed here, in order not to "get out of hand", but the reference to a good source for own further education (as with Trending) would have been nice... And while we are already at the comparison with the Trending booklet, I also discovered in this book with the explanation of normal distribution based on the example of body size unfortunately again a passage with exactly the same wording, which is in my opinion still a taint...

Furthermore, there are some small things to criticize, such as errors at two places in the presented calculation examples, which makes it a bit more difficult to understand in case the text provides you with a value of 0.84**2**, but in the calculation 0.84**5** is given or in case the text tells about a standard deviation of 15 g and references to a table which only displays 10.1 g... or if "yes" and "no" are missing in a decision tree ???? Also, it’s not necessary to mention several times that squaring is multiplying by itself, this should (still) be known to everyone... In the case of confidence intervals, it would also have been nice to have a little more application reference with examples from the pharmaceutical perspective, in order to develop a better understanding of when and for what confidence intervals could be used in the regulated environment, as it was briefly explained in the Trending booklet.

Unfortunately, the lector lacked thoroughness in this book as well, since it should have been noticeable if the text refers to a figure with colored references ("blue dots", "red line"), but the figure is unfortunately only shown in black and white... Also, the use of different designations / symbols (MW versus x or SD versus s versus σ) for the same facts could have been noticed. From my point of view this is didactically not very clever, a uniform presentation would be more valuable ???? In addition, other small things could have been noticed, such as a reference not arranged in alphabetical order, a missing dash at x, the use of "ß" instead of the Greek beta-sign or an abbreviation not introduced...

Furthermore, the integration of the illustration / tabulation into the text is not always optimally managed, for example, on one page there is a reference to a table, which is shown 6 pages later, which makes the reading flow somewhat difficult.

## And last but not least

Even though I spend more time in the pharmaceutical area nowadays, I’d first like to take up the cudgels for academia. I still remember how, at the beginning of my doctorate, my supervisor and I sought the advice of a statistician after execution of the first trials in order to evaluate how many patients we would need to include into the groups of our study (keyword: sample size)...

And what else did I take away from this booklet? Among other things, the FDA's rounding rules were interesting to me, since they are relevant to the pharmaceutical field. In addition, every book is a source of inspiration for one or the other suggestion for a new blog post or for improvement or potential additions to already existing ones.

Furthermore, the reference of the frequently incorrectly used Excel formula for calculating the standard deviation has also caused a bit of discussion with one client...

Regarding the target group to be addressed (students and TAs), my personal impression is that this booklet is perfectly aimed at them, both in terms of content, examples, and the way of conveying knowledge.

Thus, as a conclusion, it can be stated that my own expectations regarding this book were fully met, and I can recommend it without reservation.