“…Other studies have included conditions where: ratings were made by participants, but choices were made by the computer (Salti et al, 2014); choices were implicit, rather than explicit (Alós-Ferrer et al, 2012); or participants were blind to the choices as they made them (Izuma et al, 2015;Johansson et al, 2014;Luo & Yu, 2017;Miyagi et al, 2017;Nakamura & Kawabata, 2013;Sharot et al, 2010;Taya et al, 2014). The SoA effect has thus been shown to be highly robust, supported by many recent studies published even after the revelation of the statistical artifact explanation (Chammat et al, 2017;Colosio et al, 2017;Coppin et al, 2012Coppin et al, , 2014Greenberg & Spiller, 2016;Hagège et al, 2018;Ito et al, 2019;Izuma et al, 2010;Koster et al, 2015;Lee & Coricelli, 2020;Lee & Daunizeau, 2020Lee & Holyoak, 2021a, 2021bSharot et al, 2012;Tandetnik et al, 2021;Voigt et al, 2017Voigt et al, , 2019; see (Enisman et al, 2021) for a meta-analysis. This suggests that while the statistical artifact aspect of SoA may be valid, it is not enough to rule out true choice-induced preference changes.…”