Statistic is kinda weirdly chosen: 200 men : 100 women is the same as 100 men : 50 women, yet the graphic would treat one of these as twice as extreme as the other (+100 vs -50).
Granted, for values close to 100 this doesn’t distort much.
Clarifying example:
For example, suppose we had 100 men and 100 women distributed in 2 countries. If we put 60 men with 40 women in country A and 40 men with 60 women in country B, this statistic would report: A: 150 men : 100 women B: 67 men : 100 women
And it would color the excess of men a darker blue than the excess of women (because one is +50 from 100, while the other is -33; note the spacing on their gradient). This would make it seem like there’s net more men than women, which is not true.
That doesn’t matter in this graph because everything is using the same base value comparing to 100. It would only be a problem if different countries had different denominators
That assumes comparison across country is all that folks see in the graph; but one could also read this from a gender-wars kinda lens. From that perspective, this would inflate the apparent # of men per woman if one averages over countries.
For example, suppose we had 100 men and 100 women distributed in 2 countries. If we put 60 men with 40 women in country A and 40 men with 60 women in country B, this statistic would report:
A: 150 men : 100 women
B: 67 men : 100 women
And it would color the excess of men a darker blue than the excess of women (because one is +50 from 100, while the other is -33; note the spacing on their gradient). This would make it seem like there’s net more men than women, which is not true.
Oh yeah, I get what you mean now. Because of the way percentages work, it looks more impressive going up than going down; when comparing the percentages against each other. It would have been clearer to show % of population male vs female.
Using their numbers: Qatar being the most extreme in men-heavy, has 249.9 men per 100.0 women; which could instead be displayed as 71.4% men and 28.6% women.
Moldova being the most extreme women heavy, has 85.1 men per 100.0 women; which could instead be displayed as 46.0% men and 54.0% women.
If we take the median (most countries seem to have more women than men), Uganda has 98.5 men to 100.0 women; which instead could be written as 49.6% men and 50.4% women.
The site makes it look like men are going extinct in most places, and are extremely disproportionality clustered around certain locations; when reality is not that extreme. A 0.8% population gender swing in the median country isn’t an interesting article, though; so I get why they presented it in this way.
Statistic is kinda weirdly chosen: 200 men : 100 women is the same as 100 men : 50 women, yet the graphic would treat one of these as twice as extreme as the other (+100 vs -50).
Granted, for values close to 100 this doesn’t distort much.
Clarifying example:
That doesn’t matter in this graph because everything is using the same base value comparing to 100. It would only be a problem if different countries had different denominators
That assumes comparison across country is all that folks see in the graph; but one could also read this from a gender-wars kinda lens. From that perspective, this would inflate the apparent # of men per woman if one averages over countries.
For example, suppose we had 100 men and 100 women distributed in 2 countries. If we put 60 men with 40 women in country A and 40 men with 60 women in country B, this statistic would report: A: 150 men : 100 women B: 67 men : 100 women
And it would color the excess of men a darker blue than the excess of women (because one is +50 from 100, while the other is -33; note the spacing on their gradient). This would make it seem like there’s net more men than women, which is not true.
Oh yeah, I get what you mean now. Because of the way percentages work, it looks more impressive going up than going down; when comparing the percentages against each other. It would have been clearer to show % of population male vs female.
Using their numbers: Qatar being the most extreme in men-heavy, has 249.9 men per 100.0 women; which could instead be displayed as 71.4% men and 28.6% women.
Moldova being the most extreme women heavy, has 85.1 men per 100.0 women; which could instead be displayed as 46.0% men and 54.0% women.
If we take the median (most countries seem to have more women than men), Uganda has 98.5 men to 100.0 women; which instead could be written as 49.6% men and 50.4% women.
The site makes it look like men are going extinct in most places, and are extremely disproportionality clustered around certain locations; when reality is not that extreme. A 0.8% population gender swing in the median country isn’t an interesting article, though; so I get why they presented it in this way.