This paper compares reference values assigned by isotope dilution mass spectrometry with participants' consensus values in a small-scale proficiency test (PT) on the determination of residues of organochlorine pesticides (a-endosulfan, b-endosulfan and endosulfan sulfate) in a tea sample. The PT was specifically organized to benchmark the technical capability among the local participating laboratories in Hong Kong providing pesticide testing service to their customers and to identify possible areas for further improvement. Participant's performance, based on z-score, was evaluated using consensus values and standard deviation that was estimated from the Horwitz function. Two out of the seven participating laboratories were identified to produce either questionable or unsatisfactory results (in terms of z-scores), showing that the performance of the majority was satisfactory. However, the number of questionable or unsatisfactory results increased to three if the consensus values were replaced by assigned reference values derived from gas spectrometry-isotope dilution mass spectrometry with higher metrological traceability than those used in obtaining the consensus values. Though the said PT program was fit for its intended purposes at that point of time, revisiting the program explicitly demonstrates the significant importance of assigned reference values in achieving reliable assessment in reality. Unless the bias of consensus values is known with metrological traceability, PT scheme providers should be cautious in assessing the competence of participating laboratories using consensus values particularly in small-scale inter-comparison programs. The importance of pre-wetting dry tea samples to improve recovery of pesticide residues is also noted.