Infectious bronchitis virus (IBV) causes infectious bronchitis in poultry, a respiratory disease that is a source of major economic loss to the poultry industry. Detection and the study of the molecular pathogenesis of the virus often involves the use of real-time quantitative PCR assays (qPCR). To account for error within the experiments, the levels of target gene transcription are normalised to that of suitable reference genes. Despite publication of the MIQE guidelines in 2009, single un-tested reference genes are often used for normalization of qPCR assays in avian research studies. Here we use the geNorm algorithm to identify suitable reference genes in different avian cell types during infection with apathogenic and pathogenic strains of IBV. We discuss the importance of selecting an appropriate experimental sample subset for geNorm analysis, and show the effect that this selection can have on resultant reference gene selection. The effects of inappropriate normalization on the transcription pattern of a cellular signalling gene, AKT1, and the interferon-inducible, MX1, were studied. We identify the possibility of the misinterpretation of qPCR data when an inappropriate normalisation strategy is employed. This is most notable when measuring the transcription of AKT1, where changes are minimal during infection.