Few other comments (edited and just leaving the technical aspects):
VexIQ is a great value, and I think it has great
potential. The build quality on the plastic parts (tolerances and ease of assembly) isn't quite to the level of Lego. Removing pins can be difficult, especially for elementary aged students, and the pins do wear a bit over time. Fortunately, bags of connectors are reasonably cheap. Removing the corner braces and hard rubber spaces (especially against a gear) is especially tough for the young ones.
ModKit was basically unusable when I tried it in December 2013, with very limited functionality. Robot C was quite a bit better, although it is expensive per license at ~$80. Larger programs can get extremely reasonable annual licenses at ~$250/30 seats/year, but it is still an added cost. I actually took a few minutes to revisit the forum and see a couple more options coming along with the ability to program in Python that is intriguing. I didn't have any trouble teaching students as young as 3rd grade to program in Robot C though. I would show them working programs, let them tweak parameters to understand what the commands mean, and then turn the commands they know into a script.
I'm REALLY excited about some
comments Art Dutra (IFI) made on the VexIQ I2C sensors, essentially about making the protocol open to the community so we can integrate our own own sensors back to the controllers (like the Lego community). If this comes to fruition, I'd love to play with it myself more and with our high school students as a quick and dirty platform to teach programmers. Hopefully, they take the same approach to the mechanical side and encourage custom parts development to extend the base build.
Long story short, I'd say that understand VexIQ a developing community and product, but with a large upside. I think the hardware itself is superior at this point, but the scale of the program is obviously behind FLL, and regional event availability varies. A big upside is the controller integration as noted for tele-op control, but a note of caution is that I found it difficult to focus students on the autonomous aspects after the first clawbot build. There are some students that would much prefer to just drive it around for an hour as a toy, versus iterate on the design and learn a bit more.