The Strange Situation Procedure (SSP) is increasingly being used to study attachment between dogs and humans. It has been developed from the Ainsworth Strange Situation Procedure, which is used extensively to investigate attachment between children and their parents. In this experiment, 12 female beagle dogs were tested in two treatments to identify possible order effects in the test, a potential weakness in the SSP. In one treatment (FS), dogs participated together with a 'familiar person' and a 'stranger'. In a control treatment (SS), the same dogs participated together with two unfamiliar people, 'stranger A' and 'stranger B'. Comparisons were made between episodes within as well as between treatments. As predicted in FS, dogs explored more in the presence of the familiar person than the stranger. Importantly, they also explored more in the presence of stranger A (who appeared in the same order as the familiar person and followed the same procedure) than stranger B in SS. Furthermore, comparisons between treatments, where a familiar person was present in FS and stranger A was present in SS, showed no differences in exploration. In combination, these results indicate that the effect of a familiar person on dogs' exploratory behaviour, a key feature when assessing secure attachment styles, could not be tested reliably due to the order in which the familiar person and the stranger appear. It is proposed that in the future only counterbalanced versions of the SSP are used. Alternatively, since dogs reliably initiated more contact with the familiar person compared to the strangers, it is suggested that future studies on attachment in dogs towards humans should focus either on the behaviour of the dog in those episodes of the SSP when the person returns, or on reunion behaviour in other studies, specially designed to address dog-human interactions at this time.