TBH, no, I was just nit-picking the headline - as my OP was a one-liner that re-interpreted the headline for context. Y’know, like how sometimes we do a “fixed that for you” snark in this forum…
I am not pretending to have any of the answers to the bigger issues you are talking about - I’m just looking at honesty in statistical reporting for its own sake.
IMO, misleading reporting can weaken an otherwise good argument, and I critique arguments for causes I personally believe in more harshly than others because I care about intellectual honesty when I’m expecting things to be taken seriously. Because I see racism is a pervasive problem then we should not make up claims to get the point across.
With scientific rigour in mind - It’s true a null result is still a result, and it’s fair to report on it (we found no abnormal trend at all) more. In fact, I think there was an article to that effect on BB earlier this year, but I can’t find the linky.
That Fusion site in your first link? They only let me read the article if I Like them on Facebook. A dubious practice, and since I refuse to give Zuckerberg my personal data I can’t read it.
In this study, black hosts were shown to discriminate against black guests to the same extent as anyone (table 5) - the fact was even called out in the article.
Yeah–if it’s racism at work it means blacks are just as racist against blacks as whites are. This doesn’t make much sense, it’s much more reasonable to conclude there’s some more substantial factor going on here. A likely candidate is looking at what a “black” name means. People are usually named after a family member or after someone the family looks up to. What does that say about a “black” name? If it’s from family, nothing. If it’s from looking up to–it says the parents look up to blackness–racist and also associated with people who don’t fit into the normal working world well.
Combine that with no photo to show respectability and it’s quite understandable people would be reluctant to rent to them. When you do such testing with a position out of the ordinary (in this case, no profile photo) don’t expect good results!
Did you know that most cameras don’t photograph black skin as well as white skin? It is very hard to get a decent photo with your average everyday camera if you have dark skin. Why would anyone attach a photo that makes them look like a dark blob?
“Maybe there’s something about black people, Scalia responded, that makes them gravitate towards lower level accommodations. Perhaps there is something about these airBnB accommodations that is unsuitable for them.”
And most don’t give it a second thought. So, white person hearing another white person say something discriminatory does not set off a mental alarm. It passes by unnoticed most of the time.
Unless somebody has cracked the magic of dynamic range on a budget, no cameras are going to work as well.
Human teeth and sclera are all pretty much the same color(barring not-always-serious dental matters and usually-dire liver malfunction); and under reasonably bright light surfaces reflective enough to show ‘highlights’ also tend to be pretty similar.
With light colored objects in the field, you need considerably less dynamic range to get good results on also-fairly-light-colored tissue(though, in the cheap seats, it’s still pretty common for the detail of dark hair to just dissolve into a blob with a couple of shiny parts) than you do to get good results on darker tissue.
If you don’t have enough dynamic range to solve the problem, you can still err in different ways to suit different uses(if you concentrating on resolving detail in dark regions of the image you’ll ‘blow out’ light regions into flat blobs of pure white; if you focus on resolving detail in light regions, you’ll lose darker areas to flat black blobs); but you can’t get good results simultaneously without improving the dynamic range of your sensor, which is costly, where the technology allows for it at all.(One of the reasons why the ‘High Dynamic Range’ images you see are frequently reconstructed in postprocessing from multiple shots taken at different exposures; because even quality DSLR sensors aren’t nearly enough for the dynamic range of the final processed image(and, for truly HDR shots, your monitor, video card, most of your software, and even the interconnect protocol between your video card and monitor also aren’t good enough).
Given that most film and silicon image sensors have struggled to reach adequacy; I’d very much suspect that there are a lot of creative hacks in use, and that most of them are chosen for best results on the people the vendors care more about; but dynamic range is a virtue that nobody can get as much as they want of at reasonable cost; so ‘neutral’ cameras simply aren’t on the table(at any mass-market price point, I’m not sure how the high end is to parity or superiority with human vision); and I’d imagine that most of the compromises have not been in favor of getting the best results from dark skin.
