H 0 and the full Planck power spectra, within the EDE model one obtains a non-zero fraction of the total energy density in EDE at the critical redshift, f EDE (z c ) = 0.108 +0.035 −0.028 , with a corresponding Hubble parameter H 0 = 71.5 ± 1.2 km/s/Mpc [30] (adding supernovae (SNe) and baryon acoustic oscillation 'standard ruler' (BAO) data leads to insignificant shifts). Without the SH 0 ES prior, one has instead an upper bound of the form f EDE (z c ) < 0.088 at 95% confidence level (CL) and H 0 = 68.29 +0.75 −1.3 km/s/Mpc [30,31]. 2 Recent analyses of EDE using data from the Atacama Cosmology Telescope's fourth data release (ACT DR4) [34] alone have shown a slight (∼ 2.2σ) preference for the presence of an EDE component with a fraction f EDE (z c ) ∼ 0.15 and H 0 ∼ 74 km/Mpc/s [35,36]. Interestingly, the inclusion of large-scale CMB temperature measurements by the Wilkinson Microwave Anisotropy Probe (WMAP) [37] or the Planck satellite [2] restricted to the WMAP multipole range increases the preference to ∼ 3σ. A similar analysis using the third generation South Pole Telescope 2018 (SPT-3G) data [38] was presented in Ref. [39] (see also Refs. [28,29] for previous studies using SPTpol). There is no evidence for EDE over ΛCDM using SPT-3G alone or when combined with the Planck temperature power spectrum restricted to < 650, giving the marginalized constraint f EDE (z c ) < 0.2 at 95% CL in the latter case. Combining ACT DR4 and/or SPT-3G with the full Planck CMB power spectra returns an upper limit on f EDE (z c ), albeit less restrictive than for Planck alone.In Refs. [35,36] it was argued that the ACT DR4 preference for EDE is mainly driven by a feature in the ACT DR4 EE power spectrum around ∼ 500 when ACT DR4 is considered alone, with an additional broadlydistributed contribution from the TE spectrum when in combination with restricted Planck TT data ( < 650 or < 1060). Ref.[36] also considered the role of Planck polarization data, finding that the evidence for a nonzero f EDE (z c ) and an increased H 0 persists, as long as the Planck TT spectrum is restricted to < 1060.Building on these previous studies, the work presented here explores in more detail how the evidence for EDE using data from ACT DR4, SPT-3G or both data sets is impacted by the inclusion of the more precise intermediatescale (O( ) = 100) polarization measurements by Planck. We test the robustness of the results to changes in the Planck TE polarization efficiency and the dust contamination amplitudes in Planck EE. We also further investigate the role of Planck high-TT data as well as that of several non-CMB probes.This paper is organized as follows. In Section II we briefly summarize the numerical setup and cosmological data sets used in our analysis. In Section III we present