The law of small numbers bias is a representativeness heuristic that often leads individuals to draw extensive conclusions from small samples while underestimating the generalizability in larger ones. This study investigated whether individuals overestimate perceived discrimination in small employment samples and underestimate it in large ones. A pre-registered scenario-based experiment was conducted, where participants (N = 874) estimated probability of discrimination versus chance in skewed hiring distributions. We manipulated employment sample size (filling four vs. 100 positions) and ethnic majority (hiring more immigrants or Swedes) using a 2x2 between-subject design. A tendency for people to overestimate perceived discrimination by underestimating the impact of chance in small employment samples was revealed. Conversely, in large employment samples, people tended to underestimate perceived discrimination by overestimating the impact of chance. Hence, results aligned with the law of small numbers. Furthermore, participants were more inclined to attribute an event as discriminatory when organizations hired more Swedes than immigrants, reflecting (accurate) prototypes of discrimination. This study's implications are discussed concerning the repercussions of underestimating and overestimating perceived discrimination in hiring situations. Future research suggestions are also provided.