On 2016 Jul 08, UFRJ Neurobiology and Reproducibility Journal Club commented:

The study demonstrates that the presence of a conspecific animal during a fear conditioning task can influence fear behavior in mice, which is an interesting finding. However, we point out that, since training in the pair-exposed animals was performed with a conspecific, while testing was performed with an isolated animal, there is a possibility that the reduced freezing in this group when compared to single-exposed animals was merely due to a contextual change (i.e. presence of a conspecific or not), and not necessarily to social interaction. After all, the training and testing sessions were clearly more similar to each other in the single-exposed group, a fact that could account for the higher freezing with no need to postulate any kind of social modulation of fear. Thus, we feel that an important control missing from the study is a group in which pair-exposure occurred both in the learning and retrieval sessions. Alternatively, the demonstration that a non-social contextual element during training (e.g. an inanimate object) does not alter freezing would also strengthen the case that the observed difference is indeed socially mediated.

On 2016 Jul 08, UFRJ Neurobiology and Reproducibility Journal Club commented:

The study demonstrates that the presence of a conspecific animal during a fear conditioning task can influence fear behavior in mice, which is an interesting finding. However, we point out that, since training in the pair-exposed animals was performed with a conspecific, while testing was performed with an isolated animal, there is a possibility that the reduced freezing in this group when compared to single-exposed animals was merely due to a contextual change (i.e. presence of a conspecific or not), and not necessarily to social interaction. After all, the training and testing sessions were clearly more similar to each other in the single-exposed group, a fact that could account for the higher freezing with no need to postulate any kind of social modulation of fear. Thus, we feel that an important control missing from the study is a group in which pair-exposure occurred both in the learning and retrieval sessions. Alternatively, the demonstration that a non-social contextual element during training (e.g. an inanimate object) does not alter freezing would also strengthen the case that the observed difference is indeed socially mediated.