This paper provides experimental evidence for considerable education-related differences in processing and comprehension of predicate nominal Object Relative Clauses. The experiment measured high academic attainment (HAA) and low academic attainment (LAA), native and non-native speakers' response time and decision accuracy using an online sentence-picture matching task which compared processing and comprehension of Active Transitive and Subject and Object Relative Clause sentences (e.g., This is the girl that hit the boy, This is the girl that the boy hit). The results support usage-based (e.g., Barlow and Kemmer, 2000; Bybee, 2010; Langacker, 2000) and constraint-based approaches, particularly those that predict that participants' performance is shaped by regularity, frequency and direct experience of the constructions (e.g., MacDonald and Christiansen, 2002). All groups processed Actives and Subject Relative Clauses faster than Object Relative Clauses whilst the LAA native and HAA non-native participants made significantly more errors with Object Relative Clauses than Active and Subject Relative Clause sentences. However, since the results show evidence of non-native speakers outperforming (some) native speakers in an online task tapping knowledge of complex grammar, they are problematic for accounts which posit that only non-native speakers are restricted to ‘shallow’ syntactic processing (e.g., Clahsen and Felser, 2006).