How should I interpret this statement from this article:
> In an academic study from MIT and Melbourne Business School, researchers decided to test the accuracy of third-party marketing data. So, how accurate is gender targeting? It’s accurate 42.3% of the time.
If this statement is correct, does it mean that gender targeting is worse-than-chance? I am willing to believe that the data is poor, but find it had to believe that it is THAT poor. Am I misunderstanding something?
No it's not correct to say it's worse than chance. That only works for binary classification problems. Without knowing how many categories people were sorted into, we can't quite say how bad chance (or random label assignment) would do. But anything less than 1/ (number of categories) is worse than random.
You're right. There are more than two categories of gender. 1/(number of categories) works if the categories are the same size. They are not the same size. Male and female cover a vast majority of cases.
What percentage would all the categories that are not male and not female need to be in order for 42.3% to be better-than-chance?
Regardless of how you see it, being correct 42.3% seems worse-than-chance, which is the worse kind of incorrect.
> The other genders are faddish nonsense.
Gender is a touchy subject. While there is a lot of room for disagreement, it is hard to deny that some people are born in a way that does not fall neatly into the traditional male and female categories. It's dismissive to talk about it as faddish nonsense. It does not help people to understand the world around us better. Instead, it reinforces simplistic thinking about a complicated world.
Yes but intersex has nothing to do with unambiguously male or female people deciding to adopt an alternative gender that doesn't match their sex.
One is a biological condition of sex development, the other is a social issue, largely caused by the idea that there are certain thoughts and behaviours that are inherently masculine or feminine.
So for example, if you're a woman who doesn't "feel like a woman" (actually: at odds with cultural stereotypes of women) you might choose to identify as non-binary. Which is happening a lot more these days, amongst some populations - so I think this could reasonably be called a fad.
> In an academic study from MIT and Melbourne Business School, researchers decided to test the accuracy of third-party marketing data. So, how accurate is gender targeting? It’s accurate 42.3% of the time.
If this statement is correct, does it mean that gender targeting is worse-than-chance? I am willing to believe that the data is poor, but find it had to believe that it is THAT poor. Am I misunderstanding something?