The aim of the study was to test the hypothesis that polymerase slippage correlates to the length of repeat stretches consisting of uniform repeats against the alternative hypothesis that the total number of repeats is most relevant. Two short tetrameric short tandem repeats (STRs) with different repeat structures were investigated: D3S1545 containing only homogeneous (GATA) n repeat stretches and D7S1517 with compound repeat arrays of GAAA and CAAA repeats. Additionally two different polymerases (Herculase and AmpliTaq Gold) were used which gave comparable results. No correlation was found for the hypothesis ''total repeat number against percent of stutter''; in contrast, the other hypothesis that the number of uniform repeats is relevant for the degree of stutter gave a strong positive correlation (0.82 for selected D7S1517 alleles) which confirmed the hypothesis that polymerase slippage correlates to the length of repeat stretches consisting of uniform repeats.