Your source ignores benefits.

The source you're referring to explicitly discusses benefits. It even includes a chart that depicts the relationship between "wages and salaries" and "total benefits" over the past twenty years.

While I think it's important to distinguish between wages and total earnings, it's also important to differentiate between earnings and purchasing power. The fact that benefits have increased over time won't mean much to the average American if that increase simply reflects a corresponding increase in the cost of healthcare, for example.

In my view, the important question to ask isn't whether median earnings have been completely stagnant over recent decades (i.e. 0.00% growth in real terms), but whether increasing inequality represents some sort of structural change. Thinking in those terms, the range of varying estimates regarding the change in median earnings over recent decades seems relatively small and insignificant when compared to the larger economic picture.