Statistical analysis should be performed on log2-transformed fold change (i.e., -ΔΔCt) or on ΔCt. The advantage of using ΔCt is that, after running the test on your chosen softwater, the descriptive stats (mean difference, SD, SEM, 95%CI) will be expressed in -ΔΔCt, which you can easily transform to fold change (2^-ΔΔCt) if you wish.
Although the original Livak paper performed stats on fold changes, this parameter tends to deviate from a Gaussian distribution and is only useful for graphic presentation or summary.
Ref.
Article The Ultimate qPCR Experiment: Producing Publication Quality,...
Article Statistical analysis of real-time RT-PCR data