Objective: Primarily to evaluate the completeness of the description of the clinical reasoning process in RCTs with patients with non-specific neck pain with an argued or diagnosed cause i.e. an impairment or activity limitation. Secondly, to determine the association between the completeness of the clinical reasoning process and the degree of risk of bias. Data sources: Pubmed, Cinahl and PEDro were systematically searched from inception to July 2016. Study selection: RCTs (n = 122) with patients with non-specific neck pain receiving physiotherapy treatment published in English were included. Data extraction: Data extraction included study characteristics and important features of the clinical reasoning process based on the Hypothesis-Oriented Algorithm for Clinicians II (HOAC II)]. Data synthesis: Thirty-seven studies (30%) had a complete clinical reasoning process of which 8 (6%) had a ‘diagnosed cause’ and 29 (24%) had an ‘argued cause’. The Spearmans rho association between the extent of the clinical reasoning process and the risk of bias was −0.2. Conclusions: In the majority of studies (70%) the described clinical reasoning process was incomplete. A very small proportion (6%) had a ‘diagnosed cause’. Therefore, a better methodological quality does not necessarily imply a better described clinical reasoning process.