Interesting to note that male 'nonfathers' fall below female 'nonmothers' on all the measured axes, and female 'nonmothers' were rated equal to or higher than confirmed fathers in all categories except 'proportion recommended for management'.

The methodology appears sound and the results are internally consistent: Mothers and singly males suffer discrimination of about 1 and 1/2 standard deviation, respectively. Single women and fathers are generally rated equally, although they were never directly compared.

I'm not so convinced by her research. Firstly, her undergraduate experiment is close to worthless in my view. I have zero reason to believe an experiment on a bunch of undergraduates has any external validity. If you read enough experimental economics papers, you'll find out pretty fast that what a bunch of undergrads do in a laboratory experiment often diverges greatly from how people operate in the real world. 

However, I'll take the audit study a bit more seriously. The problem with audit studies, is that they are very difficult to interpret - especially when doing labor market studies. Search and training costs are extremely high for employers. The marginal product of labor of any given employee is unobservable prior to incurring the search and training costs. This is why employers act on expected costs - not actual; hence, signals matter and signals are hard to measure and so many empirical / statistical models end up misspecified. For example, she cites Bertrand and Mullainathan's study on "black-name" resume applications, which supposedly demonstrated discrimination against blacks because they were much less likely to get called in for an interview if the resume had a black-sounding name (thanks mom for naming me Theodore!)  However, Fryer and Levitt came along and demonstrated that "black sounding names" were correlated with all sorts of other socioeconomic variables, including, having a single black mother; have young and less-educated parents; lower birth weight; less income; even if your parents had insurance coverage. Growing up in Detroit, my own anecdotal experience confirms Fryer-Levitt's result that more outlandish ethnically black named children were most likely raised in these backgrounds. Once they controlled for socioeconomic background in life outcomes, they found almost no relationship between life outcomes and having a "black name." Now, it's trivially obvious that this social background hinders human capital accumulation (to which I speak of broadly to include social and cultural capital). If employer's are aware of this (and why wouldn't they be?) then having an ethnically black name is just a signal to the employer of your type and there is no nefarious discrimination going on here (except discrimination against those with expected lower MP of labor), nor is there any rational public policy that could resolve this signal problem. 

I know I got sidetracked a bit, but I thought it was important to lay out where audit studies can yield seemingly obvious interpretations, yet be quite misleading. Motherhood is definitely correlated with all sorts of social variables, and almost certainly has a causal effect on female labor supply and human capital accumulation. One imagines that the social characteristics of mothers vs. non-mothers also varies in perhaps relevant ways. Those are main reason I find the regression she runs on the applications difficult to interpret - there is just too much potential for omitted variable bias. She also seemingly ignores all the economic literature on this topic. The economic literature has offered a variety of interpretations that are more plausible that either employer's just hate women or they irrational think women, especially mothers, suck at work. For example, female labor supply, especially that of mothers, on both the extensive and intensive margins, is significantly less than that of men over the lifecycle (around 40% difference in hours worked between 20 and 40 iirc). This leads to significant differences in human capital accumulation. Hence, women, and especially mothers, all things equal, will have lower wages and lower prospects in employment because of the expected future path of their labor supply and / or their expected human capital accumulation. Economists have found that when calibrated in structural models, you can explain a lot of the gender gap. What Correll is seeing as an incompetency stereotype may just be a reflection of the empirical fact women usually have less human capital accumulation than men do. Discrimination, of the type favored by Correll that people just believe women and mothers are incompetent, appear to be second-order. I suppose you can call the signal-based hiring behavior "discrimination," but it's not nefarious (or at least any more nefarious than only hiring people with college degrees, for example) and there is little that public policy can do to correct these problems because informational frictions are hard to correct. If labor market signaling-based discrimination is the primary cause, there isn't a lot we can do outside of finding ways to increase female labor supply. Of course, these economics papers aren't perfect either (what paper is in social science?), but they have a strong theoretical explanation and appear to match the quantitative data well. Correll should have engaged this literature if she is serious about this topic. Also, all Correll did was look at application behaviors. I'm more interested in actual outcomes. Her research doesn't allow us to figure out how much gender discrimination matters, perhaps the effects are relatively small in the real world despite stereotypes. An interesting research topic would be to compare the outcomes of single father's versus single mother's, it's hardily perfect, but at least this would better isolate any gender discrimination effect. 

My more general point is Correll's research is hardily definitive. I would call it generous to say it is suggestive, there is just too many other potential, plausible explanations. I don't think it's clear she is wrong, but not accounting for labor supply data is just mind-boggling. My personal belief is that actual and expected labor supply, and how it affects human capital accumulation, are the primary reasons for the gender gap, with discrimination playing second-order effects. The social issue isn't employer's beliefs, but how our culture promotes women as the primary child-rearers - and well past the infant age - and burdens them with most of the household production. It's hard to imagine, given the labor supply data, how this social norm couldn't have a large influence on female wages. I think if men played a larger role in child rearing, without this behavior being stigmatized or viewed as abnormal, we would see more equitable economic outcomes among men and women. I don't think inequities in economic outcomes for genders will be entirely resolved until either this social norm is changed, or women start accumulating a lot more human capital than men. How you go about changing social norms is a whole different topic. Although this is interesting from an economic point of view. Here's a fun little side-consequence of this. In a model of familial bargaining over home production, where labor supply elasticities between men and women endogenously emerge because men have superior bargaining power (a proxy for cultural expectations on women) and men have a preference for paid labor (because of the usual cultural reasons that we devalue unpaid / household production), it implies that women should receive a lump-sum tax transfer and men should face higher marginal tax rates, as this is Ramsey optimal due to the difference in elasticities. It will also lead to a smaller difference in labor supply elasticities, so, perhaps, gender-based taxation is the way to correct the problem of lower wages for women and initiative significant changes in social norms, not to mention it would make the tax code more efficient on top of it. 

This is clearly visible just at the level of census and income statistics. Who is poor, and who is not? People have known this for years- women with children have less hours in a day to both work and come home and take care of things. There has been much discussion and around this subject in and out of academia. These women will not be as mobile in the workforce because they have the second shift, and that is not 100% of the pay gap, though perhaps a lot of it. This is still true when you control for race and other factors.

However, the fact that there is (throughout most of society) really no such thing as co-parenting is NOT the best reason for women to stay out of the workforce. Things may seem less "fair" between working women and their husbands in one way, but women who never work are putting themselves (and their children) in a precarious position. If they ever divorce- or perhaps their husband even passes away, they may find themselves middle-aged with no experience, teenage children, entering the workforce for the first time. Additionally, even married women are not gauranteed that their spouses will help economically to the degree that they expect, or are able. A family is not a single economic unit in this way- the various members' interests actually compete. Women spend a higher percentage of their incomes on their families when they do work.

Taxing men and women differently may make some difference. If married men and single women are comparable, it is still important to remember that most people do not remain single. Engineering of this kind will only do so much. Ultimately, women who do not want to be poor will have fewer and fewer children. This is the trend that is seen in all developed nations. Where women are able to have less children, they choose not to have them because, whatever individual faults they may possess, the majority of them actually understand that they are going to be doing the care, and maybe even paying for everything. Trillions of hours of care will go into raising the next generation; women will provide the vast majority- 90+ percent of this. Men will not be providing the same proportion of financial support even if you consider only legitimate children.