The article by Eisinga, te Grotenhuis, and Pelzer (2013) addresses the reliability assessment of two-item scales, a common practice in research due to cost and time constraints. The authors explore the reliability coefficients most appropriate for such scales, including Cronbach’s alpha, Pearson correlation, and the Spearman-Brown formula. They argue that while Cronbach’s alpha is widely used, it is inappropriate for two-item scales as it underestimates true reliability under relaxed assumptions. The Pearson correlation is shown to be an inadequate measure, representing the reliability of a single-item test. The Spearman-Brown formula, on the other hand, is found to be more reliable and less biased, especially when the items are tau-equivalent or congeneric. The authors conclude that the Spearman-Brown formula is the most appropriate reliability coefficient for two-item scales, but they emphasize that using more items is generally better for enhancing construct validity. They also highlight the importance of ensuring local independence between items to avoid inflated reliability estimates.The article by Eisinga, te Grotenhuis, and Pelzer (2013) addresses the reliability assessment of two-item scales, a common practice in research due to cost and time constraints. The authors explore the reliability coefficients most appropriate for such scales, including Cronbach’s alpha, Pearson correlation, and the Spearman-Brown formula. They argue that while Cronbach’s alpha is widely used, it is inappropriate for two-item scales as it underestimates true reliability under relaxed assumptions. The Pearson correlation is shown to be an inadequate measure, representing the reliability of a single-item test. The Spearman-Brown formula, on the other hand, is found to be more reliable and less biased, especially when the items are tau-equivalent or congeneric. The authors conclude that the Spearman-Brown formula is the most appropriate reliability coefficient for two-item scales, but they emphasize that using more items is generally better for enhancing construct validity. They also highlight the importance of ensuring local independence between items to avoid inflated reliability estimates.