velocity histograms can give you an impression of the amount of "peak-locking", but I don't know of a way of quantifying it through histogram data...
2. How to find amount of error related to correlation peak through the results obtained?
I think that there are many sources of error that relate to the correlation peak: The position of the peak itself may be incorrect. Or the subpixel estimation might be incorrect. Raffels book has a nice chapter on this. PIVlab is also just PIV, so the facts that are mentioned in Raffels book can be applied to PIVlab too.