The assessment of vocational workplace-based qualifications in England relies on human assessors (raters). These assessors observe naturally occurring, non-standardised evidence, unique to each learner and evaluate the learner as competent/not yet competent against content standards. Whilst these are considered difficult to measure, this study aims to provide information on the operational reliability of assessor decisions for three workplace-based vocational qualifications in Hairdressing and Electrotechnical Engineering. It is shown how existing assessment records, typically generated locally by the training providers, were collected and used in a reliability study. A suitable methodology for estimating the inter-rater reliability and internal consistency was then used with this empirical data. In general, the reliability estimates measured may be considered high. This research study shows that it may be possible to measure reliability for these qualifications using commonly used measurement methods, although several characteristics of the assessment system should be considered in their interpretation. (Contains 2 tables and 1 note.)