As a goat's separation from or reintegration into its group is likely to have an adverse effect on the welfare of both the separated goat and the remaining goats in the group, management procedures need to be carried out in a way that minimises their negative impact. In the present study, we tested the effects of two treatments of separation and reintegration by individually separating 12 goats from four experimental groups, each composed of seven horned, non-lactating female goats. In the ‘no-contact’ treatment the separation allowed for acoustic contact with the group only, whereas in the ‘contact’ treatment tactile and visual contact was also possible. The separation lasted two days, with each separated goat experiencing both treatments (i.e. there were 24 separations in total). Per group, one separated goat was of high, medium and low rank, respectively. The effects of separation and reintegration were assessed by evaluating social interactions, lying and feeding behaviour, and the concentration of cortisol metabolites in faecal samples of separated and resident goats. Data were collected during three phases: a reference period (days −7 to −1), a separation period (days 0 and 1) and a reintegration period (days 2, 3 and 4). Separated goats fed less and had higher concentrations of faecal cortisol metabolites during the separation than during the reference period. In addition, goats undergoing the ‘no-contact’ treatment spent less time lying during separation. On the first day of the reintegration period, newly reintroduced goats were more likely to display agonistic and sniffing behaviour towards resident goats and had higher concentrations of faecal cortisol metabolites than during the reference period. The faecal cortisol metabolite levels of the recently separated goats tended to be higher and the probability of recently separated goats displaying sniffing behaviour towards resident goats was higher in the ‘no-contact’ than in the ‘contact’ treatment. By contrast, resident goats were scarcely affected by the separation and reintegration of a group member. The rank of both separated and resident goats influenced behaviour per se, but did not interact with treatment or day. In conclusion, our results indicate that separation had a greater impact on the welfare of the individual goats than did reintegration. To mitigate negative effects on the goats’ welfare, it may therefore be advantageous to allow tactile, visual and acoustic contact during separation.