13 Statistical Techniques for the Interpretation of Analytical Data 687
Whithout With
Skins
30
35
40
45
50
55
60
65
1-propanol
Whithout SO 2
With SO 2
Fig. 13.2Means plot of 1-propanol content for each levels of factors
13.1.4.6 Two-way ANOVA with no Replication
When there is only one observation foreach combination of the levels of the
two factors (m = 1), the model isxi,j = μ+αi+βj+εi,j,without the
interaction term. The purpose of this factorial design is only to test the main
effects because interaction and error are confounded. This model is also valid
for the case of one factorrandomized block designor repeated measures design.
Two-way ANOVA with no replication hasbeen used: to test the influence of the
yeast strain and aging time factors on different compounds in sparkling wines
(Mart ́ınez-Rodr ́ıguez et al. 2002; Hidalgo et al. 2004), to test differences between
the quality scores of three wines from ten tasters (O’Mahony 1986), to compare
the concentrations reported by five analytical methods for six samples (Sharaf
et al. 1986), to test the effect of vineyard and aging time factors on the phenolic,
volatile and nitrogen compounds of the wines (Pozo-Bay ́on et al. 2004), to test the
effect of time and blend factors in wines (Monagas et al. 2007), to test the effect
of technological and time factors on the nitrogen compounds in wines (Alcaide-
Hidalgo et al. 2007) and on the phenolic compounds in wines (Hern ́andez et al.
2006).
Multifactorial ANOVA can be used to test the effect of more than two factors
(Pozo-Bay ́on 2003a; Mart ́ın-Alvarez et al. 2006). ́