When horses are housed in groups, these groups are often built based on human convenience rather than horse-specific social compatibility. The resulting social dynamics could influence the occurrence of species typical behaviours to fulfil ethological needs, e.g. lying behaviour specifically in subordinate horses. Therefore, we assessed whether group structure relates to lying behaviour in group-housed horses. The sample included 112 horses of 22 groups. Horses were observed on two consecutive days using scan sampling to determine each horse’s nearest neighbour every 5 min. Agonistic behaviour was recorded continuously. During the observation days and in the night in between, horses wore loggers that measured acceleration (1 Hz), from which we extracted total lying duration and number of lying bouts per 24 h. The variables regarding agonistic behaviour and nearest neighbour were summarized by a principal component analysis. The first two principal components (PCs) explained 57% of the variance. Based on their factor loadings these PCs were named “dominance” and “compatibility”. In two linear mixed effects models, we then assessed the influence of these PCs on total lying duration and number of lying bouts. Group was considered as a random effect. Increasing dominance (F1,88=1.02, b=4.91, 95% CI [-4.77,14.58], p=0.32) and decreasing compatibility (F1,88=1.10, b=-6.58 [-19.05,5.88], p=0.30) correlated with increasing total lying duration; however, this could not be statistically supported. We could not find relationships between dominance (F1,88=0.80, b=0.14 [-0.17,0.44], p=0.37) nor compatibility (F1,88=0.14, b=0.07 [-0.29,0.43], p=0.71) with number of lying bouts that could be statistically supported. These findings suggest that social dynamics may influence lying behaviour, though this effect was not clearly evident under the observed conditions. Given the shared leadership typical of horse groups, no single social role appears to consistently affect recumbent resting. After social stabilization, agonistic interactions remained part of normal group maintenance without clearly altering lying behaviour.