The effects of uncertainties and errors in various constraints used in the analysis of multi-component life-time spectra of positrons annihilating in metals containing defects have been investigated in detail using computer simulated decay spectra and subsequent analysis. It is found that the errors in the fitted values of the main components lifetimes and intensities introduced from incorrect values of the instrumental resolution function and of the source-surface components can easily exceed the statistical uncertainties. The effect of an incorrect resolution function may be reduced by excluding the peak regions of the spectra from the analysis. The influence of using incorrect source-surface components in the analysis may on the other hand be reduced by including the peak regions of the spectra. A main conclusion of the work is that extreme caution should be exercised to avoid introducing large errors through the constraints used in the analysis of experimental lifetime data.