It's possible the negative income tax experiments results were more complicated than thought -- especially with how the results were portrayed:

http://www.widerquist.com/karl/Articles--scholarly/Failure2c...