[R] t.test or wilcox.test?

[Serebrenik, A.]

Dear all,

I have two samples that I need to compare. Running Shapiro Wilk normality test does not allow me to reject the normality assumption neither for the first sample (p-value = 0.8938) not for the second one (p-value = 0.07905). So the t.test seems to be applicable. However, the p-value for the second sample seems to be quite close to 0.05 threshold, so wouldn't it be better to use a non-parametric test such as wilcox.test?

Best regards and thanks in advance,
Alexander