Writing a book punk-style

 

Our book on statistics is coming out, after three years of work.

The number of questions on “how did this happen?” and “how is it to write a book” I’ve been getting recently warrants a blog post with a bit of an early post-mortem – here it is!

How did this happen is answered quite easily: I’ve often advised colleagues on statistics, and they frequently asked, “which book should I get”? Despite suggesting some textbooks, colleagues generally found them too long, and/or complicated and/or missing important content while containing plenty of lower-relevance information. Over time, I understood why existing books do not quite meet their needs. Examples of problems include:

•  Length over 500 pages. This is completely fine for people who want to learn statistics in a certain depth. But most researchers in life sciences are busy with their own research and do not have training and priority for an in-depth dive into statistics. Specifically, I do not think that any book still going over probability distributions or principles of statistical inference on page 100 will work.
• Too much math. Again, for people without formal training, this is a huge cognitive load they will be reluctant to undertake. In an ideal world, every life scientist would master statistical theory - but in reality, most just need to apply methods appropriately. Since statistical analyses are done by computers anyway, it’s far more valuable for readers to spend time on what each method tests, its assumptions, and its limitations, rather than its mathematical background.
• Too general. General principles can be important, but what I found life scientists really want is an overview of statistical tests and their advantages/disadvantages, and conditions of use. They want a book which they read and almost immediately see how the content is relevant to them and their research.
• Missing important content. In contrast to the previous point, which is along the lines of “too much”, many books offer too little at the same time, despite considerable length. How many books go into multiple testing problem and corrections, or pseudoreplication? How many books describe common issues of ANOVA use? How many touch on experimental design? Very few unfortunately.

     Once I had this realization, I got in touch with Professor David Eisner, a distinguished physiologist with a keen interest in statistics and reproducibility. When I shared my thoughts, he immediately agreed: there really wasn’t a book that met these needs. So I said, “This is important, how about we write one?” He said yes, and so the journey started.

     Cultural side-comment: I don’t really have idols in the sense of believing a person is 100% perfect and aspiration-worthy (my wife gets really close though!). But there are many people whose thoughts speak to me. One of them is Joe Strummer, the punk legend, frontman of the Clash and later the Mescaleros. This quote from an interview with him, when he was asked about the punk attitude: “Punk ain’t the boots or the hair dye. I’ve been asked to define it many times so I’ve actually thought about it for a couple of seconds. It must be the attitude that you have, that approach everything in life with that attitude. Say that you come in here and the music sucks. I don’t care if the guy is big and in a bad mood. The first thing I do is go up to him and say ‘change that music!’ I do it in a cool way though but I don’t sit here fuming, getting sick and having to leave in twenty minutes. I go straight in, see what’s wrong and I fix it. If we’re meeting some new couples, the second someone lights a cigarette, I grab an ashtray and it’ll be there while everything’s going on.

Everyone else there will be standing around while their ashes fall off. That is a punk attitude because I wanna be aware of what’s going on. My motto is ‘never take your eye off the ball,’ which is a soccer motto. I like to be completely aware of what’s going on at all times, even if it’s four in the morning. She needs a chair or he needs a beer. There’s no long wait ’cause I’ve already clocked it while everyone’s going (jabbering) meh-meh-meh. I’m going meh-meh-meh too but I know what’s going on around me. This is punk rock.”

That is actually quite a lot how I approach life. It’s also why surviving in academia can be tricky for me. One piece of advice that I’ve been getting consistently is “focus on your one main thing and be very, very strategic”. OK, let’s write a book on statistics unrelated to my other activities, probably with zero CV-boosting value, what could go wrong? The problem with the advice on being strategic is not that it’s incorrect (it’s unfortunately correct, given the brutal competition), but that I find an excess of strategizing quite boring and unfulfilling. I do see people in academia who invest so much of themselves into surviving in the system that they seem to lose themselves a bit (well, a gigabyte in some cases), and I don’t want to end up the same way.

Making the book

Now, to the writing itself – there are several aspects that I would love to write shocking backstories about. Let’s split it into problems with the writing itself, problems working with a writing partner, and problems with a publisher.

First, the process of designing and writing the book.
I think in most cases, it’s worth conducting an analysis of existing books, their weaknesses, and sketching out how you’ll approach this differently. Somehow, after all the conversations with colleagues in physiology research I had, and going over existing books, I quickly had a fairly specific idea of what I wanted to deliver. I sketched out an expected table of contents and started writing sections… And to my surprise, it just went rather smoothly. Most of the early draft book was written in a jacuzzi in Davis, CA, although unlike Agatha Christie, I did not consume apples there.  OK, I thought it would be just one year of work, but there was a lot going on in my life, so I find that acceptable. Of course, there were areas where I had to do more-than-expected deep dive into statistical literature and that took longer than expected. There were a few instances where I got stuck on a block of text for weeks, knowing it is not quite clear enough, without knowing how to fix this right away. Making figures and sample source codes did get very tedious in the end, but nothing worse. And then countless iterations for clarity and making sure the book remains concise. Writing a short book rich in content is harder than writing a long one – already Blaise Pascal wrote “I would have written a shorter letter, but I did not have the time” after all… However, despite various very minor hurdles, the process remained quite light, relaxing, and fun throughout. In short, it was nothing like ordinary academic work, where progress means battling obstacles, going through pain and failure, and surviving rejections, with the best possible outcome being that you earn the privilege of doing it all over again.

Second, co-authoring a book with another person.
I did not write the book alone, but together with David. Over three years, surely there had to be disagreements and arguments, right? Except not at all! Somehow, we always agreed on things very quickly and productively. We had lots of great discussions, and I feel that collectively, we made faster and better decisions. I really appreciated how there was never even a whiff of “I’m more senior, so listen and obey”. I cannot recommend enough writing a book with someone else (who is kind and committed to the long-term task). I’m certain that if I was writing alone, the final product would be nowhere near as good. This holds even though I had, as I wrote, a clear vision and the writing flowed quite naturally. Having more perspectives on a topic is invaluable, and without David’s input, the book would not be as comprehensively useful and would miss certain points that I now see as very important. And finally – discussing things together and inspiring each other simply made the process much more fun. In retrospect, I think this is one major reason why I kept high morale and commitment to finish throughout the project (the other major reason, less positive, was observing the statistics use in academic papers and hearing “there was no difference, p~0.1, n=3” at least three times a day at conferences).

Third, working with a publisher.
Working with our publisher was also reasonably smooth for the most part, and I feel that their procedures helped us make the book better than if we were self-publishing. One memorable bump came when we discovered that the printed edition would be in black and white - after I had created all the figures in colour to make the book more inviting and modern. I then had to redesign and recreate every figure, right when I least needed extra work. Still, we managed. The online version includes the colour figures, and I think even the black-and-white ones turned out clear and effective in the end.

We were pondering whether to go for an open access online resource or a traditional book with a traditional publisher. We decided for the latter in this particular case, given that:

• Many in our target audience take a traditional book more seriously than a free online resource, however good it might be,
• It helps delivering the book to those who need it, e.g. via libraries,
• It can help us produce an ultimately better and better-looking book.

     As a researcher, I prefer to publish outputs and codes in an open way, free to the consumer. However, here, it felt like the point of view of target audience was more important than my own preference for doing things. Such decisions are highly case-specific though, writing a book for a different audience, a free online resource may be absolutely the way to go.

Concluding remarks

The complete lack of any substantial issues has made this blog post probably rather boring, I apologize for the lack of drama. It is like a strange version of a Jane Austen story, where the protagonists would say they fancy each other on page two and get together and married on page three, because nobody is opposing their marriage. They have children and live happily ever after. While their dog dies at some point, it’s the worst thing that happens in the book and they recover quickly. It would be a highly unusual novel, and likewise I feel slightly odd about this being such a purely positive experience so far (that’s what academia does to you). There will be more accounts of pain and suffering in the soon-coming blog series about the T-World model of human ventricular myocyte though, don’t worry!

To wrap up, if you are considering writing a book, I’d say it’s worth considering the following:

• Is there enough of a gap on the market? This may be both with regards to the topic, or how it is presented. If not, why to make another?
• Do you see how to address the gap reasonably clearly? If not, this will be a painful process…
• Who is your target audience, and what format of learning resource is the most effective for them, while deliverable by you? How much money do you want/need to make? How well can you typeset a book on your own? If we just stick to book-like resources, three common options are:
• Self-published book that is sold: most profitable per sale, but you have to do all the production and marketing work.
• Traditional book. Not hugely profitable, but I think it helps to make the end product a bit better for most authors, and can help spreading the book. (I’d say the improvement in final product quality that you get from working with a publisher, as opposed to self-publishing, is bought at the cost of substantial time investment, slower overall process, and much lesser profits. Completely fine for me with my motivations, but may not be fine for others with theirs.)
• Free online site. You do not profit financially, but you can reach out to more people if this is the format that works well for the target audience.
• A hybrid. E.g. Hadley Wickham’s R for Data Science is both a printed book and a free online resource. This would be probably my preference, but most publishers will refuse to do this.
• Do you want to write alone or with a writing partner? I recommend the latter, but it may not always work for everyone. If you work with someone with very different expectations regarding the pace of work, or someone who disagrees with you most of the time in an unproductive way, it will be demoralizing and unhelpful.
• Are you happy working for years on a writing project? Only you can answer this…

     I hope you’ve enjoyed reading, see you next time!