The accumulation of deleterious alleles can compromise the fitness of an animal. Genomic data provides a more accurate measure of homozygosity within an individual. Runs of homozygosity (ROH) are homozygous regions in the genome that occur when closely related parents transmit identical haplotypes to offspring. As such, ROH can be used to quantify realized inbreeding. Differences in the length, location, and number of ROH between different breeds can be observed, however population sizes of genotyped animals differ. To make breed comparison possible, it is of interest to characterize and compare ROH in different dairy breeds while taking sample size into account. The objective of this study was to investigate how sample population size affects estimation of inbreeding in a multi-breed study using ROH. A total of 13,274 genotyped animals in 3 breeds (6,560 Ayrshire, 5,667 Brown Swiss, and 1,047 Guernsey) were available. In a first analysis, ROH-based inbreeding was calculated for 500 randomly selected animals of each breed (5-fold cross-validation). In a second analysis, the same approach was used for 1,000 animals randomly selected from each population. A third analysis considered 500 and 1000 animals from each breed individually, and a fourth analysis considered all animals together. Statistical analyses were performed using the R package “detectRUNS” to determine if there is a significant difference in the calculation of inbreeding between different population sizes in a multi-breed study. The results of this study will help us to better understand how to approach multi-breed studies in the future, and to quantify potential breed biases when dissimilar breed population sizes are available.