I suppose this is a perennial concern of some folks, and thus I suppose it is to be perennially misunderstood.
The current story goes: income distribution has gotten much worse since the 1980s, in the sense that any real growth in income has gone to the top 1%, leaving the middle class the same (in real terms) as they were in the 1980s. I can (sort of) understand that non-economists buy into this story, but I'm perplexed that economists still do. But, I guess it's good for some people, since debunking this story gives them something to do. In all seriousness, Terry Fitzgerald is doing really important work on this issue: http://www.minneapolisfed.org/research/economists/staff_display.cfm?id=365
Terry's main point is that the story is, well, distorted to say the least. It's not out-and-out wrong, but the case is exaggerated to a significant degree. First, it is true that more of the income growth has gone to the top income earnings. Thus there has been some widening of the gap between middle-income and top-income groups. However, what's wrong is that the middle-income earnings have not seen real growth.
The case is built on incorrect estimates of inflation, and changing definitions of "households." The CPI is the commonly used price deflator to go from nominal to real. It is well known, at least among economists, that the CPI overstates actual inflation that would affect households. This overstatement is between 1 and 1.5%/year. Google "Boskin Commission" for more on the CPI story. That overstatement makes a significant difference and explains a lack of real growth. So when you use the PCE (implicit price deflator on personal expenditure), which is preferred by macroeconomists, you get some growth in real personal income.
Next, the definition of household. In the 1980s, a household, on average had two income earners. Now more households have one income earner (single person household). So you have to compare individuals to individuals, not households to households. This explains more of the difference and so doing individual income to individual income you get more growth.
Finally, there is non-money benefits. Obviously benefits are not as good as money income, since they don't have the option value of cash (i.e. I can use cash to buy whatever I want; maybe I don't want health insurance). But these are a form of income, since benefits reduce expenditure on certain things and thus free up income for other uses. Including benefits explains more growth in the middle class income.
That's the background I want to get down. I will take up the topic again later (tomorrow maybe).