A very simple, straight-forward piece at Carpe Diem, along with a table of data, helps make things much clearer [h/t Cafe Hayek]. Here is the conclusion:
The highest-income quintile has four times more people working per household than the lowest quintile (2.08 earners vs. 0.48), and individuals in those households are far more likely to be well-educated, married and working full-time in their prime earning years. In contrast, those individuals with low incomes are far more likely to be less-educated and working part-time, and either very young or very old living in single-parent households. Given these significant differences in household characteristics, it's not too surprising that there are huge differences in incomes among American households. It's also very likely that those individuals in the highest quintile were once in the lower quintiles before they acquired job experience and education, and they'll likely be in a lower quintile again when they retire.
Understanding the factors explaining income inequality would also help explain why income inequality changes over time. For example, compared to previous years, in 2009 there were both: a) more single-parent households, and b) more married, dual-earner households, following trends going back to the 1960s, and both of those trends would explain rising income inequality over time.