“…For example, cross-validation is required with a sample of smaller, privately held firms because, as Table 1 shows, the average organizational size in our SiRi was largeapproximately 92,000 employees per firm. More important, though, may be the possibility that the type of CSP measure used in this study-even when decomposed into its stakeholder components-is affected by significant biases, conflicts of interest, or validity concerns (e.g., Carroll, 2000;Chelli & Gendron, 2013;Graafland, Eijffinger, & SmidJohan, 2004;Igalens & Gond, 2005;Liston-Heyes & Ceton, 2009;Orlitzky, 2013), so that alternatives (see, e.g., Chen & Delmas, 2011;Orlitzky & Swanson, 2012;Turker, 2009) (Hough, 2006), ANOVA and VCA are purely descriptive (McGahan & Porter, 2005;Rumelt, 1991). In terms of causality, many higher-level effects are very likely to be driven by managerial actions (McGahan & Porter, 2005, pp.…”