You've probably heard the widely quoted figure that women earn only 77 cents for every dollar a man earns. Suzy Khim drops some statistical controls on that analysis and comes away quoting researchers who find that "Women are more likely to work in lower-paying occupations and leave the workforce when they have children, for example" and that when you control for those factors the pay gap drops to "only" 91 cents or so.
And yet I don't think that simply saying we have 9 cents of discrimination and then 14 cents of life choices is very satisfying.
After all, assume you have a two-earner household and now due to childbirth one of the workers is going to shift away from market labor and toward household production. All else being equal, it would make sense for the shift to be done by the person who has the lower wages. So if a discrimination-driven wage differential exists, then it can produce a "choices"-driven differential. And this goes on down the line for a whole range of life choices. People make lots of money-vs.-something tradeoffs in life, and if women are going to be discriminated against in efforts to earn money then you'll see men and women weighting life choices differently.
And all that's before you consider the really deep questions. Existing workplace norms were set up with the understanding that there's a woman behind every successful man. Then we changed the rules to say that women are allowed to try to compete with men for professional success. But we didn't equip women with the convenient fulltime homemaker partners who can help them succeed. Even at the highest levels, someone like Angela Merkel's husband has a full-time career as a research scientist while Michelle Obama has a full-time job as First Lady of the United States. It's not "discrimination" that denies Obama a salary, but it would be naive in the extreme to just note that she left the labor force voluntarily and chalk it up to life choices.