With the continued spread of antimicrobial resistance in animals, it is important to assess its occurrence throughout a microbiome quantitatively in order to evaluate significantly affecting factors e.g. antimicrobial usage. Metagenomics methods makes it possible to measure the abundance of antimicrobial resistance genes in complex samples such as pooled faeces samples from batches of slaughter pigs. This study was performed to determine; the random error in antimicrobial resistance in pooled samples from batches pigs at slaughter, and the measurement error from the metagenomics processes. In four farms, two pooled samples were obtained from a batch of slaughter pigs by two individual samplers, each pooled sample was hereafter processed twice. Hierarchically clustered heatmaps were applied to evaluate dissimilarities between samples. The coefficient of variation was used to calculate the percentage difference between samples from the same farm. The results of the analysis revealed that it was not possible to quantitatively separate the variation arising from sampling and metagenomics processes. They both contributed to the overall measurement error of antimicrobial resistance in batches of slaughter pigs. Sampling of single pigs in 30 randomly selected pig pens within the farms provides a composition representative for frequently occurring antimicrobial resistance genes present within the farms, while rare genes were not dispersed in a similar manner. Aggregating the resistance abundance at gene family or antimicrobial class level will reduce the apparent variation originating from errors in sampling and metagenomics processing.