Results were statistically analysed using the Matlab 7.6.0 (R2008a) software.
Analyses of variance were performed according to a two-factor full factorial design
of experiments. ‘GFRP waste type’ and ‘GFRP waste content’ were considered as factors,
with two (CW and FW) and four (0%, 4%, 8% and 12%) variation levels, respectively.
A 21 41 full factorial design leads to eight different formulations, however,
both formulations CW0 and FW0 were in fact the same composition: 20% of resin,
80% of foundry sand and 0% of CW (or FW) admixture. Hence, for data treatment
purposes, these mix design formulations, with equal composition, share the same
replicates.