Data of the three replications (i.e., nine measurements per formulation) were pooled and analyzed using analysis of variance of the General Linear Models procedure of SASâ
software, and the LSMEANS procedure for generating SEM, where SE and probabilities were calculated using Type III mean squares for “chub(formulation storage time)” as an error term (SAS Institute, 1990; Littell et al.,1991). A 4 ´ 6 factorial (i.e., four formulations and six
storage intervals) and a 5 ´ 3 factorial (i.e., five formulations and three storage intervals) were employed for the refrigerated and frozen studies, respectively. Differences among formulation means were determined using Duncan’s multiple range test (SAS Institute, 1990).
Interaction between formulations and storage time; and replications was tested.