We discuss the possibility to explain the anomalies in short-baselineneutrino oscillation experiments in terms of sterile neutrinos. We work in a3+1 framework and pay special attention to recent new data from reactorexperiments, IceCube and MINOS+. We find that results from the DANSS and NEOSreactor experiments support the sterile neutrino explanation of the reactoranomaly, based on an analysis that relies solely on the relative comparison ofmeasured reactor spectra. Global data from the $\nu_e$ disappearance channelfavour sterile neutrino oscillations at the $3\sigma$ level with $\Deltam^2_{41} \approx 1.3$ eV$^2$ and $|U_{e4}| \approx 0.1$, even without anyassumptions on predicted reactor fluxes. In contrast, the anomalies in the$\nu_e$ appearance channel (dominated by LSND) are in strong tension withimproved bounds on $\nu_\mu$ disappearance, mostly driven by MINOS+ andIceCube. Under the sterile neutrino oscillation hypothesis, the p-value forthose data sets being consistent is less than $2.6\times 10^{-6}$. Therefore,an explanation of the LSND anomaly in terms of sterile neutrino oscillations inthe 3+1 scenario is excluded at the $4.7\sigma$ level. This result is robustwith respect to variations in the analysis and used data, in particular itdepends neither on the theoretically predicted reactor neutrino fluxes, nor onconstraints from any single experiment. Irrespective of the anomalies, weprovide updated constraints on the allowed mixing strengths $|U_{\alpha 4}|$($\alpha = e,\mu,\tau$) of active neutrinos with a fourth neutrino mass statein the eV range.