In this work we consider the QCD predictions for spin correlations in \( t\overline{t} \) production in hadronic collisions. In view of recent tensions between experimental data and theoretical calculations, it has been argued that one should include in the predictions also the effects of the production of the ηt, i.e. the pseudoscalar \( t\overline{t} \) bound state, or alternatively the full effects of the non-relativistic dynamics of the \( t\overline{t} \) pair near threshold. This implies the resummation of all corrections that scale like powers of αs/v (where v is the velocity of the top quark in the \( t\overline{t} \) rest frame) which are dominated by values of v of order αs. In this work, we show that, since the observables that are usually considered for these studies are integrated cross sections up to a \( t\overline{t} \) mass cut that is not small, it is possible to perform the calculation using perturbation theory, considering only the contributions that scale as the first few powers of αs/v. We examine the implications of our approach by computing corrections to nominal Monte Carlo results for correlation-sensitive observables, and compare them with available data, showing that the tension with data is no longer present.
