Reality monitoring lie‐detection studies, like others that use raw frequency counts as primary data, seem consistently to underestimate the influence of the length of (or number of words in) the account. The decisions as to whether to standardise or not, or what method of standardisation to use, are rarely empirically driven, so it is still unclear as to whether reality monitoring is more effective before or after standardisation for length. Another factor that also has received little attention in the reality monitoring literature is whether statements are produced orally or in written form. To investigate these issues, 42 autobiographical statements, 21 truthful, and 21 deceptive, including 22 oral and 20 written accounts, were analysed before and after word count standardisation. Results showed that reality monitoring criteria only discriminated significantly between truthful and deceptive accounts when no attempt to control for word count was made. Also, oral statements contained more evidence of reality monitoring criteria before standardisation for word count, whereas written statements were denser and contained more evidence of reality monitoring criteria after standardisation. Implications are discussed. Copyright © 2014 John Wiley & Sons, Ltd.