I haven't been able to find good data on individual income, sadly. At least in part because if one uses AGI then all you have to go with are tax returns and if a married couple files jointly the AGI will just be the joint number and that's that. But people also keep claiming this is "household" income, which doesn't quite match AGI in many cases (e.g. married-filing-separately situations, households which consist of taxpayers who are not married to each other, households with more than 2 taxpayers, etc).
I really wish the data-gathering methodology were actually spelled out somewhere. :(
I don't think the graph excludes single-person households. It's certainly easier to be in the 1% with two incomes, but it's not a prerequisite.
This whole 99/1% label thing is silly.
I suspect that roughly 30% of the engineering population at Google is in the 1% (given what I know of salary scales and what fraction of engineers are at each eng level). That works out to maybe 4-5k engineers (assuming half the company is engineering), out of the 6,746 in "Computer & data processing services". You really think that Google accounts for 80% of the 1% in software developers? Where's Facebook and Twitter in there - them alone should push it over the count given?
No - what's happening is that most of those well-paid software engineers are single. Of the ones that aren't, many of them married college sweethearts who work as teachers or yoga instructors or urban planners or other professions that don't make a whole lot of money. One Google senior-SWE total comp will put you in the 1% as a single taxpayer, two of them will put you in the 1% as a household, but one of them plus a teacher's salary will not get you anywhere close to the 1%.