If you deprotonate nitromethane, the negative charge will without doubt mostly reside on the oxygens rather than on the carbon atom as
1) oxygen is more electronegative than carbon and
2) there are two of them. Thus O-alkylation will predominate.
This is also the reason we cannot alkylate nitroethane with benzyl chloride to form phenyl-2-nitropropanes directly.
I think I have seen only one exception to this rule and that was the alkylation of 2-nitropropane with p-nitrobenzyl chloride, forming p-nitrophenyl-2-nitro-2-methylpropane, and the authors of the article somehow concluded that no other combinations worked. para-aminophentermine, anyone?