This study is a conceptual replication of Ellis' 'Measuring implicit and explicit knowledge of a second language: A psychometric study', published in "Studies in Second Language Acquisition" (2005), aiming to establish the importance of including belief statements (hypothesized to increase processing demands) in the design of Elicited Imitation (EI) tests. In a between-groups comparison, 103 participants completed one of two versions of Erlam's EI test (2006), one which required participants to make a 'belief' judgement, and one which did not. Both contained ungrammatical stimuli. Participants were scored for repetition of grammatical stimuli and correction of ungrammatical stimuli. An Exit questionnaire aimed to establish how aware they were of ungrammatical stimuli. Participants also completed a timed grammaticality judgement test (TGJT), an untimed grammaticality judgement test (UGJT), a metalinguistic knowledge test (MKT), and another Elicited Imitation (EI) test designed as a measure of global proficiency. Performance on each EI version was examined in relationship to performance on these measures. There was no variation between performance on either EI test version. The article underscores the need for ongoing research investigating design features of EI tests.