“…With such a design, and absent any other methodological guidance, researchers resort to outdated methods or otherwise finesse their analytical problem so that it conforms to this requirement of a dichotomous or continuous X. For instance, some researchers use a variant of the causal steps method popularized by Baron and Kenny (1986) by assessing whether group differences on Y revealed in an analysis of variance (ANOVA) disappear after controlling for a proposed mediator (see, for example, Pandelaere, Briers, Dewitte, & Warlop, 2010;Wirtz & McColl-Kennedy, 2010). Others have treated a discrete, ordinal independent variable as interval level and used standard regression-based techniques (Chandler & Pronin, 2012;Legault, Gutsell, & Inzlicht, 2011).…”