In an initial investigation of differential expression of genes caused by cis-acting regulatory elements in rice, the lack of reproducibility led us to question the basic premise of allelic expression imbalance determination: namely that departures of cDNA expression ratios from those observed in genomic DNA provide unequivocal evidence of cis-acting polymorphisms. This paper describes experiments designed to demonstrate that stochastic variation in low copy number of targets in PCR reactions give variable allelic ratios even when starting with the same copy numbers of the two alleles. These significant departures from an expected 1:1 ratio provide an explanation to the lack of reproducibility observed for our cDNA measurements.