I have done my job for 16 years and there is one thing I think we should eliminate: Microsoft Word. I must say that Microsoft Word is the root of all evils. Coupling with the poor quality publishers, it makes the problem even bigger. But Microsoft Word is like our financial system, it is too big to eliminate. What I would like to scold at this time, is it’s monstrous accomplice: scientific publishers.

I have a lot of my papers screwed by the publishers. I have a good paper published in a very prestigious journal called CHEST during my very early career. For the social scientists reading this, CHEST is the second or third most important journal in the pulmonary / respiratory medicine. I was responsible for submitting the manuscript to the journal and I have followed the entire R&R (Rewrite and Resubmit) cycle and the proofing thing. The manuscript I’ve sent out was a Microsoft Word file produced in a Windows machine using Chinese locale. In Microsoft Word, the punctuations are locale dependent. For some unknown reasons, when one types punctuations in early version of Microsoft Word in Chinese locale, they are translated into Chinese version of punctuations (full-width form, 全型標點, can you notice the difference between ! and !), including a lot of mathematical symbols. In that paper, I did the statistics and I have written a lot of “P < 0.05”. I didn’t have a lot of experience by that time and I didn’t realise “smaller signs” have been translated by Microsoft Word into Chinese punctuations. I sent out the paper, did the R&R, and got it accepted. Up to this point, everything was fine because the reviewers were dealing with a Word-converted version of PDF. Once again, everything was fine.

So now, the paper had been accepted. For those who don’t know, the paper will be typesetted and a “galley proof” (or simply “proof”) version of the paper will be sent back to the authors for proofreading and copy-editing. In most of the time, publishers ask the authors questions about the problems they have during typesetting.

When I received the proof, I quickly discovered a lot of problems. One of the biggest was, some “P < 0.05” in tables were randomly mistyped into “P > 0.05”. I thought the problem was originated from those Chinese punctuations and their DTP machine couldn’t handle them. Then the typesetter guessed what those missing symbols should be, but failed. In the proof, I have pointed out all the incorrectly typesetted symbols. Also, in the query, the typesetter would like to know the country where Hong Kong is located. I have answered this question, unwillingly, as China. Only CHINA. These information I gave to the publisher is called “corrections”. The publishers should handle all these corrections and produce a corrected version for final publication. After the authors submitted the corrections, in most of the situation publishers will not contact the authors again and put the corrected version directly online and in print. When I later saw the final version online, I was totally shocked. Firstly, all mistyped “P > 0.05” statements were not corrected. Secondly, the publisher has written a complete country name rather than China but they have written my address as “Kwong Wah Hospital, Hong Kong SAR, ROC (Republic of China)”. If you know the geopolitical situation in Asia, you should know how embarrassing it is.


Of course, I have quickly noticed the problem and reported that immediately to my ex-boss. My ex-boss asked me what should we do. I suggested to write to the publisher to complain for their sloppy job and demand a correction. My ex-boss said the readers should be able to discern those statements are mistakes and the “Republic of China” is so minor. And therefore, up to this day, that paper remains screwed and makes me as a coauthor look like a fool. It exists also as a reminder to myself, these corrections are not handled seriously. This observation is right because all major scientific publishers have outsourced their typesetting and proofing operations to some third world countries such as India or People’s Republic of China.

Since then, I always called the proof / correction procedure “Schrödinger‘s corrections”. You can never sure if the publishers incorporate your corrections. You can only know the death or alive of your paper after your paper is officially published. Later in my career, many of my papers were also screwed. But the damage was not as big as in the CHEST case. Until last week.

I have a paper accepted by the Journal of Information Technology and Politics (JITP), a good journal published by the Taylor & Francis group. Actually, the R&R cycle was very unpleasant for me because I have written tons of emails to ask for the progress of the review. Turned out what I was waiting is just a 3-sentence unhelpful review that is almost impossible to address (“the paper is badly written”). After 2 and a half years, my paper is finally accepted and I was very happy. But I didn’t realise that was the beginning of another agony. My old nemesis.

After I had received the proof of my JITP article, I was again shocked. All captions of figures were missing. A formula was missing. The funding information was mistyped. The publisher did query about some of the problems and I have turned in the corrections accordingly. (I have retained a copy of my correction as a black-n-white proof.) Few days have passed and I received the notification about my OCRID update. When I checked my article on the JITP website, all of my corrections are not corrected. Therefore, I have an article without captions of figures and a formula. I look like a fool again and another paper is screwed by the publisher.


This time I am my own boss and I have immediately contacted Taylor & Francis for that. So far, no action has been taken. I have also contacted the chief editor of JITP. From my experience of sending him tons of e-mail, he doesn’t reply even one.

I left with no choice. I never put anything on preprint server before. I think I need to do my paper justice. So, if you want to read my paper, please read it on SocArXiv.

I think the origin of the problem is purely a technical one. Why is Microsoft Word an established standard in social science? If the social science community is like the hardcore science community (e.g. Physics, Mathematics) to use LaTeX, then both the authors and the DTP are talking in the same language. There are other good plain text formats such as RMarkdown than can be transpiled to LaTeX. Arguably, RMarkdown is easier to use than Microsoft Word.

Publishers should also take our corrections seriously. We crafted our papers professionally, maybe for years. But why do you handle the final step of the pipeline like an amateur?