Wednesday, May 27, 2015

Someone (Ahem) Did a Statistical Analysis of Vox.com Photos, And It's Actually Pretty Depressing

    The other day, Vox.com ran a piece entitled "Someone did a statistical analysis of New Yorker cartoons, and it's actually pretty depressing." The semi-serious article presented the findings in the "semi-satirical journal Proceedings of the Natural Institute of Science" regarding the race and gender of characters in the semi-iconic New Yorker cartoons. The study found that "[o]ut of 1,810 total characters, 1,277 (about 70.6 percent) were male, and 1,714 (94.7 percent) were white." Vox writer Joseph Stromberg argues that there is a "need for accurate representation of nonwhite guys", and not just in movies and on TV. Stromberg quotes Amy Rothschild of the FiveThirtyEight blog asserting that it's a problem "[w]hen whole groups of people appear as stereotypes or not at all[.]" Rothschild was referring to children's books, but Stromberg says "adults are impressionable, too."
     After reading the article, I became curious about what impressionable adults might find to skew their something-other-than-the-content-of-a-persons-character worldview on the site where the article appeared, Vox.com. Since Vox does not regularly produce cartoons, I decided to use as a proxy a Google search of the Vox.com site of the first 100 facial images since the beginning of 2015. (Small sample size, but I have staff limitations.) You won't, as they say, believe what happened next:

    Read 'em and weep:
  • 81% "white dudes", as Vox likes to say
  • 9% women, all white
  • 10% non-whites (and four of them are President Obama)
     Others may come up with slightly different figures. I am particularly bad at guessing a person's race (maybe that's a good thing?)
     Anyway, to try to eliminate the possibility that Google itself is the culprit in white-male-skewing the results, I also conducted a general (the whole web) image search for "faces" since the beginning of the year.

     At a glance, the results seem more eclectic (certainly more women than men), but I did not conduct a full statistical analysis (again, staffing issues...).
     I decided to do a few more for comparison. Here's the New York Times:

     Looks like quite a variety.

     How about MSNBC?

     Whoa. Elizabeth Warren is looking pretty lonely there, eh, MSNBC?
     And finally, just for fun, The New Yorker's website:

     It appears the "white dude" dominance at The New Yorker is largely confined to the cartoons.

     So what conclusions can we draw from these analyses? Got me. I think I got a C in Statistics in college. I told you we have staffing issues.

No comments:

Post a Comment