Since you cant really see if the cemented went okay, its many thousands of feet underground, its hard to tell if this is happening. When the high pressure drilling fluids are injected, they would easily flow right up that channel into the groundwater supply.
Actually, there is technology for this: http://en.wikipedia.org/wiki/Cement_bond_log
Since there is an economic incentive for the petroleum company to apply all that pressure and frac fluids ($$$ to the tune of 1/4 to 2/3 of the total cost of drilling the well) only to the hopefully-producing zone and nowhere else, you can rely on the cementing operations to be ordinarily done right, and then verified prior to frac-ing. Which isn't to say that things don't sometimes go wrong with cementing: http://en.wikipedia.org/wiki/Deepwater_Horizon_oil_spill
Note that in that well, I doubt that fracturing was even contemplated, since the reservoir was clearly capable of significant production without aid.
They say in the propoganda that there is many thousands of feet of impermeable rock between the pay dirt layer and the groundwater, but this doesnt mean much as you just drilled a hole through it all.
"Dirt" isn't the best choice of words to describe a rock formation which requires fracturing to liberate economic amounts of hydrocarbons. But, that said, we clearly know that ground water eventually seeps down to a certain depth when it encounters something that stops it and allows it to accumulate. The reverse is true of the natural gas trapped down below. It wants to escape upward. Something stops it. So the rock itself clearly is impermeable. During drilling, the fluid (drilling mud) is designed to be viscous and heavy enough to move the rock cuttings back to the surface and protect the rock formation itself. Again, there is an economic incentive to get this right, as you don't want to either have a blowout during drilling, if the mud is too light, or for the mud to be so heavy that it forces itself out into the formations you're drilling through, causing loss of circulation of the cuttings. Best practices require drilling an oversize hole with just water as the fluid down below the bottom of potential fresh water aquifer, then setting an initial "surface casing" and cementing that alone in order to protect the fresh water, before proceeding drilling within that casing down to the (much) deeper productive zones. Incidentally, that surface casing is also what gives you the opportunity to place a "blowout preventer (BOP)" in case something goes badly wrong during drilling. If you didn't have the surface casing, you'd have nothing to attach the BOP to--nothing to allow you to seal off the well in case of major problems.