Quantcast
subscriber help

artisanal film reviews | by maryann johanson

Where Are the Women?: critics are slightly more likely to rate a film highly if it represents women well

watwcritics

My Where Are the Women? project, analyzing the films of 2015 for their representation of women, is now finished, and I’ve crunched some numbers.

Get an introduction to this analysis here. You can examine a comprehensive spreadsheet of the details about these 153 films here.

see also:

only 22% of 2015’s movies had female protagonists
best and worst representations of women on film in 2015 (and the average WATW score for the year)
• mainstream moviegoers are not turned off by films with female protagonists
movies that represent women well are just as likely to be profitable as movies that don’t, and are less risky as business propositions


Here’s what it looks like when we compare a film’s WATW score to its Rotten Tomatoes score:

(Hover over a dot for the film’s WATW score, title, and RT score. Click here to see the chart in a separate zoomable window.)

There is a 0.22 correlation between a film’s Rotten Tomatoes score and its WATW score, meaning that there is a very slight tendency for a film that represents women well to rate more highly with RT critics.tweet

And here’s what happens when we compare a film’s WATW score to its Metacritic score:

(Hover over a dot for the film’s WATW score, title, and MC score. Click here to see the chart in a separate zoomable window.)

The correlation between a film’s Metacritic score and its WATW score is a little bit higher — 0.28 — so these elite critics are slightly more likely still to rate a film better if it represents women well.tweet

What this means, basically, is that Rotten Tomatoes critics are 22 percent more likely to give a good score to films that represent women well, and Metacritic critics are 28 percent more likely to do so. What that means is that movies about women are more likely to be better movies than movies about men, as determined by criticstweet. This may be because movies about men are so predominant and tell so many similar stories over and over again, so movies about women are simply going to feel fresher.

(Big thanks to reader RogerBW for pointing out the numerical correlations, and for sending me in the direction to learn more about statistical correlations. As I said to him, I never took a statistics course!)

What about comparing Rotten Tomatoes scores with the gender of a film’s protagonist?

(Hover over a dot for the film’s title and RT score. Click here to see the chart in a separate zoomable window.)

There is a positive correlation here, but it’s very tiny: only 0.05.tweet

And Metacritic scores versus the gender of a film’s protagonist?

(Hover over a dot for the film’s title and MC score. Click here to see the chart in a separate zoomable window.)

Again, there is a very small yet still positive correlation here: 0.08.tweet

What about my own personal critical reactions to movies? I’ve made a point of keeping the WATW ratings separate from my reviews because — as I’ve said many times — a movie that represents women well can still be a crappy movie, and a movie that doesn’t represent women well can still be a great movie. (The problem isn’t that any given individual movie doesn’t represent women well but that so damn many of them don’t.) That said, how does that hold up across my reviews of these films?

(Hover over a dot for the film’s title and WATW score. Click here to see the chart in a separate zoomable window.)

The correlation here is pretty strong — 0.49 — but not absolute. I have rated highly films that do not score well on the WATW test, and have rated poorly films that do well on WATW.

And how do my grades compare with the gender of a film’s protagonist?

(Hover over a dot for the film’s title. Click here to see the chart in a separate zoomable window.)

The correlation here is weaker but still positive: 0.23. But here is scientific proof! I do not automatically like movies better if they’re about women (which I’ve often been accused of). If I did, the correlation would be 1.

More analysis!


Where Are the Women? was partially supported by a Kickstarter campaign — a HUGE thank-you to my generous Kickstarter supporters — but those funds did not begin to cover all the time, effort, and expense that went into this project.

If you find Where Are the Women? useful, interesting, or important, please support it now by:

buying some Where Are the Women? merch
becoming a monthly or yearly subscriber of FlickFilospher.com
making a one-time donation via Paypal


posted in:
where are the women
  • Jan_Willem

    Facinating stuff! A worthy conclusion to a worthwhile project. (Donation on its way.)

  • RogerBW

    It’s hard to say what’s a significant correlation, but a rule of thumb is to ignore everything below about r=0.8.

    Still, this is a small sample. It would be great if we had more, but…

  • Danielm80

    I’m curious how well Best Picture nominees represent women. There appeared to be a fair amount of green on the lists of Oscar nominees, but that may have been wishful thinking on my part.

  • I rated and ranking all the Oscar nominees from the past two years:

    http://www.flickfilosopher.com/2015/07/women-2014-oscar-nominated-films-ranked.html
    http://www.flickfilosopher.com/2016/02/where-are-the-women-the-2015-oscar-nominated-films-ranked.html

    If you’re talking about all Best Pic nominees across Oscar history, I don’t think it would be reasonable to apply the WATW criteria I developed to any films from decades ago. The movie environment has changed too much… and, if we’re lucky, things will improve in coming years so that the WATW criteria no longer accurately reflect the situation.

  • 0.8 seems like a really high bar, at least in this particular situation.

  • RogerBW

    Yes, it is. Social sciences tend to allow lower values than, say, pharmacology, but it’s all a bit of a black art anyway.

  • Danielm80

    Wow. That’s a fast response. I asked a question, and you provided exactly the information I was looking for, retroactively, two months ago. I really appreciate the in-depth data analysis, even if I didn’t back in February.

  • Jurgan

    The way I’ve generally heard it explained is that r=.22 means that 22% of the y-variable is determined by the x-variable. So 22% of the critics’ scores can be predicted by the WATW score, but the other 78% comes from elsewhere. I think I’m interpreting that right- I’m a mathematician, but statistics isn’t my area. .8 is a high bar, but that’s intentional. You’ve never claimed that 80% of a movie’s quality is based on how it represents women.

  • I’m sorry you missed these links. Maybe the site needs another redesign…

  • Danielm80

    I saw the first link, when it first appeared, but I overlooked the second one. But you have more than enough work to do without taking on another redesign, or blaming yourself for my spaciness. If you enjoy doing redesigns, knock yourself out, but the WATW project is already a big enough workload for any sane human being—and very much appreciated.

  • I worry when people say they can find stuff here. It means there’s a problem. :-(

Pin It on Pinterest