Like the study from the AAUW, "Graduating to a Pay Gap." It claims to compare men and women one year after college graduation (finds that women earn 82%), and that even when they control for occupation, hours worked, etc. they still find a wage gap of 7 percentage points. Two problems with this, they bury the 7% figure in the body of the paper while putting the 82% figure in the summary, so people ted to cite that (including politicians). The second problem is that they didn't control for occupation correctly. That would require a great deal of work as there are thousands of occupations they would need to categorize, but instead they used 12 broad occupational categories. One of these was "other white collar jobs." Another was just "Other." You can't compare librarians to attorneys and say that you compared "apples to apples." Shoddy research.

How about "The Gender Wage Gap: Extent, Trends, and Explanations," by Blau and Kahn. This one is somewhat better, it's definitely more detailed and longer, but it's not better by much. It only uses 21 broad occupational categories.

Also from the AAUW was "The Simple Truth About the Wage Gap." This one is mostly just waffle about how we also have to consider the racial wage gap, or the sexuality wage gap, and how mothers are specifically affected. It only briefly attempts to prove that the wage gap is a meaningful statistic that suggests discrimination, and it does so by citing only three other studies. The two I already mentioned, and another that I still have to work through, but only deals with a very narrow scope (it's about the wage gap among certain medical researchers, and I have to study how the medical research field works and what factors within it might affect pay to be able to make an informed judgement about this one.) This paper spends far more time assuming that of course the wage gap is proof of discrimination, and that women only choose different occupations because they are "steered" (gender equality paradox? never heard of her).

Then the was "Behind the Pay Gap," again from the AAUW. It mad the same claims and had the same flaws as "Graduating to a Pay Gap." Only 11 occupations. Utter trash.

How about "Demystifying the Gender Pay Gap: Evidence from Glassdoor Salary Data," from 2016 by Glassdoor. This one actually had solid controls for things like occupation and job position. It has over 2,000 categories! It finds that once you control for a whole host of variables, the wage gap is reduced to about 5.4%. When this one came out, I saw it everywhere. Glassdoor hosted an event where Hilary Clinton talked about the need to close the wage gap, I saw an article about the paper in my local newspaper, and so many feminists have cited this one as proof that the wage gap is real. It took me just a minute to find the problem, it doesn't control for hours worked. Now according to the BLS American Time Use Survey, among men and women working full time, women work about 5-7% fewer hours than men. Maybe women earn 5.4% less than men because they work about 5% fewer hours? Just a thought.

Then of course there was the 2019 update from Glassdoor, "Progress on the Gender Pay Gap." I seethed as I watched their YouTube video about it in which they claimed that the arguments that the wage gap can be explained by factors such as occupation, position, and experience were wrong, and their new paper proved it. Of course, I suspected that this update would suffer the same problem as the 2016 paper, so I checked and it did.

Then there was Payscale's "Does Transparency Close the Gender Wage Gap?" It controls for the relevant variables. Curiously, it says that participants who strongly agreed that their organizations were transparent reported "slightly less pay equity at $0.99," relative to those who just agreed that their organizations were transparent, where "pay equalizes completely between $1 and $1.02." So if women are paid $0.01 less than men, that isn't very equitable. If men are paid $0.02 less than women, that's totally equitable!

How about "Gender Pay Inequality: Consequences for Women, Families and the Economy," a 2016 report by Representative Carolyn B Maloney (D). It's been used by the likes of Audi following their wage gap commercial as evidence that women are paid 21% less than men, and that this is an outrage and women deserve equal pay for equal work. Of course, this only compares the median annual earnings of men and women working full time. It turns out, if you read to the appendix, they did find that if they controlled for hours worked, that reduced the wage gap by about 6 percentage points, but like in Graduating to a pau gap, they didn't put the smaller gap in the executive summary, or in the press release. The smaller gap wasn't the figure Audi ran with, because saying "women are paid 15% less" isn't quite as sensational as 21% less. Still overall a mostly useless paper, but it does further confirm that hours worked accounts for about 5-7 percentage points (more than that according to Goldin).

I'll also throw in "Equity by Design,' from the American Institute of Architects. I'm an architecture graduate student who took a required architectural theory class which focused on race for four weeks and then feminism for another. One of the required readings claimed that women in architecture faced pay discrimination, and cited "Equity by Design." I will say, it controls for an impressive number of variables. I can't think of any that they missed. The problem, is that they never, ever, EVER, controlled for all of them at once. What it would do is say "we controlled for experience, but that didn't close the wage gap." Then they would say "we controlled for licensure, but that didn't close the wage gap." That's great, but this isn't whack-a-mole. We aren't looking for a single explanation, you have to control for all the variables at once. I thought about taking the data they provided and attempting to control for them all at once on my own, but because they are architects they had to make the graphics aesthetically pleasing, which sometimes meant just having bar graphs with no numbers, no specific date, and without direct access to their data, and maybe even with it, the potential error would just cast doubt on any results, and I might accidentally control for the same effect twice through different variables. Basically, this could have been useful, but it stops too short of actually proving anything about the wage gap.

Another required reading from that class cried wage discrimination and cited a survey that found that 47% of women in architecture believed that they were paid less than men for the same work. I don't even know what to do with that information. That isn't even a majority of women, and 47% of women can be wrong. It actually just makes it sound partisan if it's sitting at around half.

I am aware of the report by CONSAD which examined about 50 papers (including Graduating to a Pay Gap), but I'm still debating whether I should read it thoroughly. I have a rule where I only read papers that claim to find proof of a wage gap or support the feminist position on the matter. I do this so no one can claim that I cherry-pick my sources. however, one time I got into an argument where I provided my sources, this idiot said I was just cherry-picking my sources, I pointed out to him that I had already said I only look at feminist sources, and then I guess he actually decided to read the abstracts of my sources, because then he said that my own sources proved me wrong and said they found evidence of a wage gap. He also called me armchair educated, as if that was worse than just being entirely ignorant. I've seen the CONSAD report cited by both feminists and anti-feminists. From what I can gather, it doesn't really claim to reach a conclusion. I'm having the same thoughts about Claudia Goldin's work, but I guess I should just read it since it's so widely used.

I've gone through plenty of other sources that didn't deal directly with the wage gap but provided useful data. Like Georgetown University's "What's it Worth." I crunched the numbers and found that based on choice in college major alone (not even occupation), women would be expected to earn 11% less than men. Also the previously mentioned American Time Use Survey, and other similar sources.