The effect of primer selection on real-time polymerase chain reaction (RT-PCR) performance was tested. Primer sets of varying length of product were used to amplify the sequence of beta-actin. Variability in length caused variability in RT-PCR performance. Kinetic parameters of PCR were studied by mathematical approximation of real-time data by means of a four-parametric sigmoid model. This model describes the full kinetics of the amplification trajectory. Statistical exploration of parameters yielded by this model revealed that reactions with higher amplification efficiency - primed by well-performing primers - proceed with lower variability and are therefore better suited for measurement purposes.