“…The very useful statistic, PSD, the predicted standard deviation is obtained from the leave-one-out Finally, we combine the two data sets, together with some additional compounds taken from the literature (Alexander et al, 1977;Gharavi et al, 1983;Martin, Wu and Beerbower, 1984;Niimi, 1991;Pinsuwan, Li and Yalkowsky, 1995;Bernazzani et al, 1995;Kristl and Vesnaver, 1995;Bustamante, Pena and Barra, 1998;Martinez and Gomez, 2001;Avila and Martínez, 2002;Perlovich et al, 2003Perlovich et al, , 2010Perlovich et al, , 2011Ran et al, 2002;Sepassi and Yalkowsky, 2006;Quingzhu et al, 2007Quingzhu et al, , 2008Thimmasetty et al, 2008) or from our own previous work (Fletcher et al, 1998;Hoover et al, 2004Hoover et al, , 2005aHoover et al, , 2005bHoover et al, , 2006Charlton et al, 2005;Stovall et al, 2005aStovall et al, , 2005bFlanagan et al, 2006aFlanagan et al, , 2006bHolley et al, 2011;Ye et al, 2011;Stephens et al, 2012;Wilson et al, 2012;Bowen et al, 2013). For most of the common compounds in the Admire and Raevsky data sets, the listed log Soct(mol/L) values were the same.…”