Google

A split-half correlation is a statistical technique used to assess the internal consistency of a multiple-item measure. This involves dividing the measure's items into two distinct halves, calculating a total score for each half, and then finding the correlation between those two sets of scores. For example, comparing the even- and odd-numbered items of the Rosenberg Self-Esteem Scale might yield a strong positive correlation, such as $$+0.88$$, demonstrating high internal consistency. In general, a correlation coefficient of $$+0.80$$ or higher is considered indicative of good internal consistency. A strong correlation indicates that the measure is consistently assessing the target construct.

Split-Half Correlation

Cronbach's $$\alpha$$ (alpha) is the most frequently used statistic by psychological researchers to evaluate the internal consistency of a multiple-item measure. Conceptually, it represents the mean of all possible split-half correlations for a given set of items; for example, there are $$252$$ ways to split a $$10$$-item measure into two halves. A high value is generally interpreted as indicating good internal consistency across the measure, with a statistical threshold of $$+0.80$$ or greater generally considered good. In addition to internal consistency, Cronbach's $$\alpha$$ is also commonly utilized to assess inter-rater reliability.

Learn Before

Related