Pitfalls in limiting studies of reactivity to a family of compounds

Case studies of the oxidation potentials for MeCpMn(CO)(2)B where B represents nitrogen and phosphorus donors, as well as the CO stretching frequencies of eta-CpFe(CO)(COCH3)PX(3) adducts are used to illustrate potential errors in data interpretation that can arise when studies are restricted to one family of compounds. These systems show that the uncertainty in the parameters from the data fit exceeds that indicated by R(2) or the standard deviations in the fit when the C-B/E(B) ratios of the donors employed do not vary significantly. When an improper scale is used in data analyses, the R(2) for data fits is improved by limiting the correlation to a single family of compounds at the expense of meaningful interpretations. By definition, substituent constant correlations are limited to a single family, so care must be exercised in interpreting the parameters from these data fits.