3 ms·

> The reality is that wealth is inversely correlated with fertility. Yes and no, it's a lot more nuanced. There can be higher fertility and the very low high a

> The reality is that wealth is inversely correlated with fertility. Yes and no, it's a lot more nuanced. There can be higher fertility and the very low high and very high end of wealth, but it's also not consistent nor a solid link. This article looked at as an well illustrated one to me: https://ifstudies.org/blog/more-money-more-babies-whats-the-relationship-between-income-fertility https://ifstudies.org/blog/more-money-more-babies-whats-the-...

somenameforme 1y ago
Their graph is setting off red flags for me. A decent sized sample of data over any sort of remotely normal distribution doesn't look like that where you're seeing a 3% change on the x axis drive a 30% shift on the y, and then back again. Here [1] are some of the data that they probably used. Table 7 in particular. Here is the percent of women in each income bracket who gave birth in the past year by ascending income per household member (their choice of data, not mine - I would prefer completed fertility): 7.96 7.51 6.39 5.14 4.47 3.78 3.18 Here is the percent by total household income in ascending order of income: 6.27 5.23 5.64 5.88 5.26 5.30 5.26 4.98 4.64 4.75 In both cases there is a practically linear, and sharp, inverse correlation between income and fertility. I have no idea how they derived their graph as that data does not seem to be directly provided, but there's no combination of the lines in their graphs that would yield these data as an average, so I suspect they made a mistake. I would not dispute that there is a U curved shape to fertility, but it's misleading as the tail end is in extremely high incomes. And in any case, their graphs look more like some sort of messed up sine waves, which is obviously just wrong! [1] - https://www.census.gov/data/tables/2022/demo/fertility/women-fertility.html#par_list_63 https://www.census.gov/data/tables/2022/demo/fertility/women... ---- EDIT: Actually I have a hypothesis. The census tables give an extremely useless datum that's very easy to misinterpret. The column is simply labeled "percent". It waxes and wanes all over the place, very much like their graph does. But it's the percent of all births that came from a given income group. But that is completely meaningless, because what matters is the data I gave (and had to manually calculate - by adding a new column) which is the percent of each group that is having children. Otherwise you're graphing some bastardization of population size at each percentile (a bell curve) multiplied by a pseudo-randomizing linear decreasing factor (fertility). So you get a graph that looks weird and makes no sense.

somenameforme 1y ago

Their graph is setting off red flags for me. A decent sized sample of data over any sort of remotely normal distribution doesn't look like that where you're seeing a 3% change on the x axis drive a 30% shift on the y, and then back again. Here [1] are some of the data that they probably used. Table 7 in particular. Here is the percent of women in each income bracket who gave birth in the past year by ascending income per household member (their choice of data, not mine - I would prefer completed fertility): 7.96 7.51 6.39 5.14 4.47 3.78 3.18 Here is the percent by total household income in ascending order of income: 6.27 5.23 5.64 5.88 5.26 5.30 5.26 4.98 4.64 4.75 In both cases there is a practically linear, and sharp, inverse correlation between income and fertility. I have no idea how they derived their graph as that data does not seem to be directly provided, but there's no combination of the lines in their graphs that would yield these data as an average, so I suspect they made a mistake. I would not dispute that there is a U curved shape to fertility, but it's misleading as the tail end is in extremely high incomes. And in any case, their graphs look more like some sort of messed up sine waves, which is obviously just wrong! [1] - https://www.census.gov/data/tables/2022/demo/fertility/women-fertility.html#par_list_63 https://www.census.gov/data/tables/2022/demo/fertility/women... ---- EDIT: Actually I have a hypothesis. The census tables give an extremely useless datum that's very easy to misinterpret. The column is simply labeled "percent". It waxes and wanes all over the place, very much like their graph does. But it's the percent of all births that came from a given income group. But that is completely meaningless, because what matters is the data I gave (and had to manually calculate - by adding a new column) which is the percent of each group that is having children. Otherwise you're graphing some bastardization of population size at each percentile (a bell curve) multiplied by a pseudo-randomizing linear decreasing factor (fertility). So you get a graph that looks weird and makes no sense.