Actually, I don't think it needs to be done at all. It doesn't really serve a purpose, in fact, it's misleading. It references Mulan and mentions that the dragon has more words than her, but he's a blithering, cowardly idiot, and Mulan is a stoic and reserved hero. Yet, the article focuses on the fact that the dragon has more words said than Mulan. Mulan is the strong, silent type archetype while the dragon is the mouthy, cowardly sidekick archetype that Eddie Murphy would do in the Shrek movies as well. I haven't seen Mulan in ages, so perhaps I could be less than 100% accurate here, but Mulan is a film about female empowerment and trying to make the claim that a strong woman can do anything a strong man can do. I don't think it's inclusion in a study trying to prove male dominance is a good idea unless your idea is just to provide a raw set of data with no context.
Perhaps it would be best to consider that this study is analyzing larger trends in industry representation, and is not an act of criticism toward any specific film on an artistic level. Naturally there will be films, like Mulan, where the word disparity has a reason that relates to the film itself, but this is not what the study is trying to argue. It's getting at the fact that, on a systemic level, we tend to give larger roles to men more often, as evidenced by the fact that male characters speak more than female characters across the board. For that to be the case, it wouldn't matter that some of the individual films have a reason for the disparity, but that all of the films, or at least a supermajority of them, do, and calls attention to the types of films we tend to see made, with the types of roles and the size of roles given, on a larger scale. The fact that Mulan is a strong silent type and the Dragon is a blabbermouth is irrelevant to the larger industry trends.
Now, on a rhetorical level, choosing that film as an example of the problem they are examining would be a mistake, if that's what they had done. However, it isn't. They even go out of their way to explicit make the point I have, and use Mulan as an example of how the study's methodology is not a critique of the films studied.
This dataset isn’t perfect. As with Mulan, a plot can center around a character, even though the dialogue doesn’t reflect it. And all of our data is based on screenplays, not a perfect transcription of a film.
They realize that Mulan is an example of positive representation for women, they used that example to show that the data is only useful as a way of discussing the issue, not as an arbitrary barometer for good representation. In the methodology section, they even go on to talk in more detail about their methodology and its limitations. The researchers fully recognize that the study isn't above reproach, but also clearly state that their intent isn't to comment on the quality of each individual film, but to provide context through data that explains trends about representation of women in film. If you continue to scroll down they also have data sorted by genre, data about the age of women (how women over 40 face unique challenges when it comes to getting roles while roles become steadily more available with age for men) etc. The study is actually surprisingly comprehensive and aware of it's limitations.
As with all studies of this nature, the data collected only can point us in the direction of a larger discussion of an issue, it's up to the researchers, and ourselves, the readers, to determine exactly what that data means and what we should do with the information. When I read a study like this, it illustrates using data the larger problems mainstream cinema has with gender representation, and suggests a real concrete effect of those problems, an effect the industry and the audience can look to correct, by actively seeking to create more opportunity for women-led projects, both in front of and behind camera, and supporting those projects. Now, you may not consider this to be an issue at all, and that would be unfortunate but I suppose right now I'll have to just agree to disagree, but it would be more productive to be upfront about that disagreement and talk about why you feel it isn't an issue, than to misinterpret a study by focusing on one data point in a much larger trend as a way of deflecting from the real issue.