The overall (dotted) line shows a rise from ~27,000 to ~31,000, which is consistent with the data I linked.
Perhaps "consistent" but still not "a complete picture".
Your table lists gains for median incomes gains for sectors of the population over a 20 year span (ending a decade ago BTW) with increases of 31335/28939=1.08, 17648/10741=1.64, 22740/17390=1.31, 18379/9944=1.85 , while the doted overall line, as you state, is 31/27=1.15
None of the educational lines are any where near 60-80% gains, or even half that of 30-40% gains - are women (at least the black and white ones) really such a small part of the economy that when you nearly double their income it has no visible effect on the overall population numbers?
If we go with "all males" and "all females" from the wikipedia table we get 30513/27206=1.12 and 17629/10683=1.65, are there really so few women workers that their substantive gains are virtually washed out by all those guys.
As I said, I don't think the linked table gives a complete picture. I don't have a great source that does give such a picture.