This first variationist study of ancient Egyptian examines two possessive constructions in Middle and Late Egyptian. Egyptologists maintain that the innovative form does not gain popularity until Late Egyptian, and that in the Late Egyptian stage, the language had two different registers – each favouring a different possessive variant (Allen 2013, Junge 2005). I confirm that the innovative form rises in use from Middle to Late Egyptian, and that both variants appear throughout official and vernacular texts during the Late Egyptian period. I then use logistic regression to show that the new variant is disfavoured in official texts, supporting the Egyptological hypothesis. An extension of the Egyptological position is the implication that two registers may be an example of diglossia, and that because of this the rate of change (the rate of the spread of the new variant) in each register should be different (Junge 2005). The data were modeled using a logistic function (Kroch 1989), and it was found instead that language change occurs at the same rate in both contexts, contradicting the disglossia hypothesis and supporting the style shift account as well as the Constant Rate Hypothesis.