Reader Comments

Post a new comment on this article

Author’s response to comments!

Posted by knauff on 11 Jan 2015 at 16:37 GMT

We seriously thank all commentators for their interest in our article and the many stimulating remarks. The metrics of the paper, the discussions on the internet, and the comments on Twitter and Facebook show us that our research is a hot topic for many of the world's scientists. Many of you also asked us for our opinion about the criticism on our research. Here are some of our thoughts:

Copying is not writing? Of course, scholarly writing is much more than copying a given text. However, we do not believe that is a valid criticism on our study. Composing a scholarly text is a “thinking+writing” task, but we wanted to use a method that allowed us to explore the pure “writing part” of the more general task of composing a scholarly text. We think that is necessary, as the intention of our study is to compare the efficiency of LaTeX and Word as text preparation tools, and not to cover or intermix this issue with other cognitive processes. Another rationale for using predefined texts is that this is the only way to realize a fair comparison between the two software alternatives. Asking people to generate own texts would be methodologically inadequate, because this would result in a high variability between texts and therefore would not allow reasonable comparisons between the two software tools.

We understand that people who are not familiar with the experimental methods of psychology (and usability testing) are surprised about our decision to use predefined texts. However, this is the normal and methodologically common way to minimize the effects of different types of texts which otherwise would cover the pure differences of software usability. In this context, it is important to see that we used texts with high variability in text structure, difficulty, and complexity. We think this makes our texts representative for many sorts of scholarly texts and certainly more representative than self-generated texts would have been.

Is a three pages sample representative? Scholarly texts are typically much longer than our probe texts. However, we simply picked out a sample of material from a larger universe. Situations in which you can investigate all possible variations and circumstances of a task are very seldom. In science, we usually have to limit our research to a reasonable number of observations in order to make justified generalizations. Longer texts might lead to additional requirements, but the question is whether the requirements that come on top of shorter texts would significantly change the results obtained with shorter texts.

Do we compare apples and oranges? We think that is just verbal hairsplitting and does not say “that the entire premise of the paper is flawed because Word is a text editor and LaTeX is a file format, making any compassion difficult” ( First, we deliberately used the very general term “document preparation systems” (not e.g. “text processing systems”). Can it be denied that Word and LaTeX are both document preparation systems? Both are used for producing text documents with a layout, even if the systems have a different underlying logic. Second, we ourselves mentioned the differences in the paper, by saying that “another characteristic of our study is that it is practically impossible to evaluate LaTeX without also evaluating the used editors. In fact, our research measured the efficiency of Word against LaTeX in combination with some editor interfaces” (p. 10-11). We do not think that it is fair to make this point although we already discussed this issue in the paper. Moreover, we think the fastidious terminology distinction makes LaTeX resistant against any kind of criticism and comparison. If nothing is comparable to LaTeX than in conclusion nothing can be better. We are not sure whether this is a productive way to respond to an empirical usability study.

Working in isolation and in teams. We have the feeling that many LaTeX users are not adequately aware of the problems that LaTeX produces when scholars collaborate in teams. Word offers the helpful “track changes“ tool, which makes collaboration very easy and efficient. LaTeX users, in contrast, often use the uncomfortable comment function in PDF documents to work on the same document with non-LaTeX users. Alternatively, we have seen many LaTeX users converting LaTeX files into Word, Word documents into LaTeX and back and forth repeatedly. This is time consuming, error-prone, and unproductive in particular if LaTeX users and non-LaTeX users work together in teams. Word is easy to learn and to use for almost everybody, whereas learning LaTeX requires a significant amount of time and effort. We wished LaTeX users would keep that in mind in particular when collaborating with colleagues that do not use LaTeX .

Are we surprised by the negative reactions? No! Everything else would have been surprising. We were aware that the LaTeX community is convinced that LaTeX is much better than other document preparation systems. Yet, the motivation for our study was to provide an empirical basis that helps us making rational decisions about using LaTeX or a different, possibly more efficient document preparation system. What is surprising, however, is how irrational and emotional some of the responses seem to be. A lot of the criticism relies on assumptions, which are based on subjective impressions or opinions. From our point of view empirical results make a stronger case than claims that are not based on such empirical findings. It is astonishing how some commentators ignore the basic principles of scientific decision-making that is, collecting facts, control over variables, using systematic methods, careful measurement, connecting causes and effects, and making rational evidence-based decisions, instead of generalizing personal impressions or opinions. We invite all people who are not happy with our findings to see our study as a starting point for further empirical investigations on the advantages and disadvantages of different document preparation systems .

Why do so many people disagree with our conclusions? First, we think that this impression might be based on the fact that disbelievers more often express their opinions on the internet, whereas believers remain silent. Second, LaTeX users are very active on the internet and many LaTeX-related news are distributed over mailing lists, news groups, etc.. Third, from the beginning on we were aware that the issue is a highly emotional issue for many LaTeX users. Imagine we would have published a paper in which we criticize other software systems. Do you expect that, for example, the Word or PowerPoint or PaintShop community would have a similarly extreme response? In fact, we think that the passion is a special habitus of the LaTeX community. We think that there are several causes for this. Cognitive dissonance is just one of the reasons. Another is the following, which was expressed by a commentator: “The author doesn't understand signaling. With latex your signaling that you're part of the boys club, and that may be the most important reason people are using latex” (ccab at : http://www.econjobrumors....). We actually agree with this commentator. One reason for the LaTeX-preference might be that it shows that you have a special competence that others do not have. As we already said in the paper, some people may think that mastering LaTeX is a “must” for any “true” expert in their discipline and that it is a unique feature of their discipline. Another reason might be what psychologists call polarization. Polarization is given if a group of people is split based on opposing opinions. Over time, the opinions get more extreme. Also the so-called confirmation bias might play a role: people tend to search for information in a way that confirms their beliefs or hypotheses but have difficulties to account for information that contradict existing beliefs.

There might be some real reasons why one could favor LaTeX over Word. In particular, we explicitly said in our paper that LaTeX has advantages if a text is full of mathematical formulae. We thought the community would be pleased to have empirical evidence for this fact. We do not argue against LaTeX for such highly mathematical texts. Another reason might be that LaTeX is a non-commercial product and we admittedly have sympathy for the uncommercial and non-profit oriented attitude of the LaTeX community. There might be good reasons not to use software from the few software companies with their almost complete monopolies.

However, overall we think that there aren't many rational reasons why people should use LaTeX if they produce scholarly text, which are not heavily loaded with mathematical representations. In our opinion, a better solution would be to use more efficient document preparation systems or to develop a new system that, ideally, combines the advantages of Word and LaTeX. Of course this is not a trivial task and we can imagine that again some people would not be satisfied with it (Markdown, we think, is not such an alternative as it probably has many problems that LaTeX also has).

Let us close with saying that the lead author used LaTeX for many years in an interdisciplinary project with computer scientists. Therefore, we are well aware of the pros and cons of LaTeX. However, in our group nobody uses LaTeX anymore since an efficient way of collaborating in the same document is more important to us than individual software preferences.

Markus Knauff & Jelica Nejasmic

No competing interests declared.

RE: Author’s response to comments!

djao replied to knauff on 11 Jan 2015 at 21:17 GMT

1. The authors state "Copying is not writing." Correct. The authors state "Composing a scholarly text is a "thinking+writing" task." Correct. Notice what's missing? There is no link given between "copying" (which is what the authors measure) and "writing" (which is what scholarly research involves). The authors measured "copying". The authors did not measure "thinking". The authors did not measure "writing". So the authors did not measure either of the two components of scholarly writing. The authors claim that "copying" is close enough to "writing" that measuring "copying" amounts to measuring "writing". I not only disagree, I am staggered that anyone could possibly make this inference. As the authors themselves point out, copying is not writing. Copying is not related in any way to any scholarly research task.

2. The authors assert that a three-page sample can justifiably be generalized to larger samples. I disagree. The authors give no evidence for this assertion.

3. The authors rightly point out that Word has many software features that aid collaboration (such as "track changes"), but then completely fail to mention that LaTeX also similarly supports many software features that aid collaboration (such as git, which is not mentioned at all in the authors' response). This kind of one-sided perspective is exactly what we, readers, are complaining about, and deservedly so.

4. The authors go to great lengths to accuse commenters of irrational reponses, but completely ignore the rational criticisms. In addition, the one-sided bias on the part of the authors noted above is irrational.

Maybe we are not "disbelievers." Maybe the study, methodlogy, outcome, and conclusions are really flawed. Maybe the authors are the ones guilty of confirmation bias. I see no willingness on the part of the authors to question their own very questionable assumptions.

No competing interests declared.