Despite the increase of literature on seabird plastic ingestion in recent years, few studies have assessed how plastic loads vary according to different sampling methods. Most studies use necropsies of seabirds with a natural cause of death, e.g. beached or predated, to determine plastic loads and monitor marine debris. Sampling naturally dead seabirds may be biased as they have perished because of their intrinsic factors, e.g. poor body condition, high parasite loads, sickness or predation, affecting estimates of plastic loads. However, seabirds killed accidentally may be more representative of the population. Here, we used the short-tailed shearwater Ardenna tenuirostris to test different sampling methods: naturally beached fledglings and accidentally road-killed fledglings after being attracted and grounded by artificial lights. We compared plastic load, body condition, and feeding strategies (through using feathers' δ13C and δ15N isotope niche) between beached and road-killed fledglings. Beached birds showed higher plastic loads, poorer body condition and reduced isotopic variability, suggesting that this group is not a representative subsample of the whole cohort of the fledgling population. Our results might have implications for long-term monitoring programs of seabird plastic ingestion. Monitoring plastic debris through beached birds could overestimate plastic ingestion by the entire population. We encourage the establishment of refined monitoring programs using fledglings grounded by light pollution if available. These samples focus on known cohorts from the same population. The fledgling plastic loads are transferred from parents during parental feeding, accumulating during the chick-rearing period. Thus, these fledglings provide a higher and valuable temporal resolution, which is more useful and informative than unknown life history of beached birds.