Sea lice audits were performed by the Atlantic Veterinary College on commercial aquaculture sites in New Brunswick, Canada, in 2011. Although the primary objective was to verify that farms were reporting similar lice counts to third-party counts, more detailed comparisons were made to identify when lice counts were more likely to differ between the audit team and farm employees. A total of 28 sea lice audits were conducted on 16 sites between June and December 2011. During each audit, 10 cages were evaluated per site where possible, with ten fish per cage being evaluated by an audit technician and a further ten by a farm employee. Data analysis included descriptive statistics of lice counts by stage and limits of agreement plots. A random effects negative binomial model that accounted for clustering of cages within sites was applied to assess the effect of counter type and season on lice counts by stage. The results indicate that farms counts were generally in agreement with audit counts. However, when the average counts for chalimus and preadult (male and female) and adult male lice stages were high, farm counters were more likely to report a lower value. Higher lice counts were observed during autumn compared to summer especially for the adult female stage. Finally, there was a significant clustering effect for site and cage, with most of the variation attributable to site.