Guess we don't have to worry about gentrification (or do we?)

The Pew Research Center just came out with a much-touted new study showing that American neighborhoods are becoming more economically segregated (or at least purporting to show this) 

I was a little curious, so I actually clicked on the link to the Pew report's maps for individual metro areas.  The Philadelphia map  showed that most urban neighborhoods are either mixed-income or lower-class; in fact, it showed only one homogenously high-income area within the city limits of Philadelphia (the easternmost part of downtown near the Delaware River)  At first glance, I thought: well, at least gentrification isn't driving out poor people.

But then I zoomed in on the map and found some even more interesting results.  The area where I used to live, just north of Rittenhouse Square between Broad Street and the Schuylkill river, was labelled "low-income."  Rittenhouse Square?  Low-income?  What are these people smoking? 

Then I looked at Boston; at least Back Bay and Beacon Hill weren't labelled as poor- but they did count as mixed, the same category as my truly middle-class neighborhood in Queens.  That can't be right! 

What went wrong here?  The study defines lower-income as having a household income below $34,000 and upper-income as a household income above $104,000.  Back Bay and Beacon Hill are just below the $104,000 threshold.

But this methodology makes no sense in the context of urban areas.  Urban areas (especially close to downtown) tend to have smaller households, and thus smaller household incomes even if they are well off.  For example, even if every single person in Beacon Hill made $100,000, it would be poorer (according to Pew's methodology) than a drive-to-qualify suburb where every household contained a husband and wife earning $51,000 apiece.  And in some cities, such as Philadelphia (where two major universities are within commuting distance of downtown) urban areas may be more student-heavy than suburbs, thus artificially depressing household incomes in cities (since students are often not earning money at all).   So relying solely on household income makes city neighborhoods look artifically poor.

But of course, the purpose of the study was not to compare city and suburb but to see whether rich areas are becoming more homogenously rich.  I'm not sure I can answer that question, but I think to do so I'd want to see if other income measurements yield the same result as Pew's methodology.