Group housing of does is hypothesized to improve animal welfare through favouring positive social interactions. We built a part-time group housing system for does by connecting 4 individual housing units through hatches. The part-time group housing system included a platform, a gnawing block, and a burrow. Spatial position, health and injury frequency of does and kits were evaluated for 32 females and their litters housed in 8 part-time group housing systems. Reproductive performance were compared to a control group (n = 8 females housed individually). The grouping started at D12 (D1 = artificial insemination), it was interrupted after 10 days (at D22) due to increased injuries rates and severity caused by fighting: injured does increased from 25% (D13) to 63% (D22), and severely injured does increased from 0% (D13) to 28% (D22). The weight gain during the experiment was similar between groups. The litter size at weaning was lower in the part-time group housing system than in individual housing (9.2 vs. 10.0, P < 0.01). At D22, the proportion of does having injuries was high in part-time housing group (53% of 32 does). The platform were used for 32% of the observation time. Does were observed more frequently on the platform during the last two weeks of the experiment just before weaning their litter (66% and 47% of the total observation, P < 0.05), confirming that the platform offered does the possibility of escaping from their kits once they leave the nest box. In conclusion, a partitioned enriched housing with burrow for animal refuge did not permit to reduce fighting between group-housed does, it caused high level of injuries and reduced reproductive performance.