ABSTRACT
This chapter focuses on reliability and its nature. Reliability deals with consistency of measurement. The concern with reliability in assessment in the behavioral sciences is paramount because of the difficulty of producing consistent measures of achievement and psychological constructs. Reliability is a multifaceted concept rather than a singular idea. Like validity coefficients, reliability coefficients are correlations but are interpreted in the context of reliability. The reliability coefficient is computed using exactly the same procedures as for computing the validity coefficient. The degree to which they produce equivalent results is an indication of their reliability. There is a relationship between the length of an instrument and its reliability. The Spearman–Brown formula “prophesizes” the change in reliability of an instrument as a function of a change in length of the test. The general split-half method, and especially the odd-even split, is based on the assumption that the two halves of the instrument are equivalent.
