2.4. Statistical analysis
Proximate composition data and lipid profiles, obtained from the 4
mortadella formulations (0%, 25%, 50%, and 100% soybean oil) at
7 days of storage, were analyzed as a completely randomized design.
One mortadella unit individually obtained from each batch was considered
an experimental unit for each treatment. The data were analyzed
with the GLM procedure (SAS Institute, 2002) using the
following model: Y = μ + Oi + e; where μ = the overall mean,
Oi = the effect of substituting pork fat with different concentrations
of soybean oil (i = 1 to 4), and e = the experimental error. The treatment
effects were then tested by using contrasts for linear, quadratic,
and cubic effects of soybean oil incorporation, assuming a level of significance
of P b 0.05.