The sign of the theoretically predicted steric asymmetry S in rotational inelastic state resolved molecule–atom collisions is questioned. It is shown that the sign of the T-matrix obtained on a basis of non-oriented rotational states of the molecule depends on the choice of the Jacobi coordinates in which the collision problem is solved. Explicit expressions for the state-to-state dependence of the integral and differential cross-sections for oriented and non-oriented molecules are presented. The effect of the choice of Jacobi angles and the inertial frame of reference on the sign of S are discussed in detail. The sign of the earliest obtained expression for the steric asymmetry of the integral cross-section (as reported by van Leuken et al.) is found to be correct, whereas that of the orientational- dependent contribution of the differential cross-section (as reported by Alexander and Stolte) requires a negative multiplication factor. Quantum mechanical calculations on collisions of OH with Ar are performed. These calculations do not agree with the experimental sign of S and cannot be interpreted in terms of a simple ball and stick model. Inspection of the HIBRIDON source code shows that the prepared wavefunction carries an orientation that is opposite to the one assumed. This could offer an explanation for the disagreement for S. For collisions of NO with Ar, the signs of the quantum mechanically calculated and the experimental values of S appear to disagree with the simple ball and stick model. The experimental sign of S has been reinvestigated. Previous experimental results are confirmed.