As a graduation requirement, students often carry out research which must be presented and defended. In this process, related data is gathered using relevant instruments which must satisfy technical properties of validity and reliability. A widespread practice in most institutions is that the researcher validates the instrument through experts’ scrutiny for only face and content, and perhaps construct validity without recourse to reliability. However, most research works contain sections addressing the instrument's reliability, with quoted high indices or coefficients obtained from faulty procedures for establishing reliability. This paper examined the reliability of data collection instruments, the types of reliability, when to use them and methods of computing them with the appropriate statistical tools. It was suggested among others, that the inclusion of reliability procedures during instrument validation for data collection, and internal consistency should be emphasised.