The standard definition of poor is those below the poverty line. There are various definitions of "near poor" which usually starts at the poverty line and goes up to 125-150% of the poverty line.
It's reasonable to assume that a significant majority of people with income greater than 100% of the poverty level have jobs. The only other possibility is that they are retired and are drawing money from SS, other savings, or disability.
As to how much it skews the numbers - I can't access the full paper so I'll make an assumption to illustrate my point. The authors claim "most" people with less than 125% of the FPL income have jobs. I'm going to interpret most to mean 55%.
Now I know from the census bureau that about 14% of the population is below the FPL and another 5% is between 100% and 125% of the FPL.
http://www.census.gov/hhes/www/poverty/data/historical/hstpov6.xls
If 80% of the people in the 100-125% range have jobs and we back that out of the 55% number, we find that only 46% of the true poor have jobs. I'd guess that more than 80% of the near poor have jobs so the effect may be even greater.