According to Electronic Structure of NO2 Studied by Photoelectron and Vacuum-uv Spectroscopy and
Gaussian Orbital Calculations J. Chem. Phys. 53, 705 (1970) :
The highest molecular orbital $4a_1$ is occupied by the 1 unpaired electron.
The electron population of this orbital is (see table VI):
0.53 on the N atom (0.16 2s, 0.37 2pz)
0.24 on each O atom (0.24 pz)
Experimentally, Oxides and Oxyions of the Non-metals. Part II.
C02- and NO2 of the Chemical Society (1962): 2873-2880 collects values for the partition of unpaired electron density among N and O atomic orbitals.
The most purely experimental values are:
N 2s 0.094
N 2pz 0.364
N 2px 0.054
O 2pz 0.33
N 2s 0.103
N 2pz 0.471
O 2pz 0.33
(where the values for O are both O atoms added together)
There is extensive discussion of the various Lewis structures of NO2 from a dimerization point of view (that the opposite spin electrons should be on the atoms that form the dimer bond) starting on page 90 if the 2015 book Bonding in Electron-Rich Molecules: Qualitative Valence-Bond Approach via Increase-valence Structures , and also elsewhere in the book from a monomer point of view.
Overall, it is clear theoretically and experimentally that the unpaired electron is delocalized, having significant density on both N and the two O atoms. It is wrong to say that a particular one of the Lewis structures is "correct".