This paper argues that Cronbach's alpha, commonly used as a reliability and internal consistency measure, has significant limitations. Alpha is not a reliable estimate of test reliability and does not accurately reflect the internal structure of a test. It is a lower bound to reliability, often an underestimate, and is unrelated to the test's internal consistency. Alpha is frequently misinterpreted as a measure of internal consistency, leading to incorrect assumptions about the test's validity. The paper highlights that statistics based on a single test administration provide limited information about individual test performance accuracy. It also discusses the availability of better reliability estimates, such as the greatest lower bound (glb), which provides a more accurate reliability estimate. The paper concludes that alpha is not a valid measure of internal consistency or unidimensionality, and that better reliability estimation methods should be used. The paper also emphasizes the importance of considering the limitations of single-administration data in assessing individual measurement accuracy. Overall, the paper argues that alpha is not a reliable or useful measure of test reliability or internal consistency and that better alternatives should be used.This paper argues that Cronbach's alpha, commonly used as a reliability and internal consistency measure, has significant limitations. Alpha is not a reliable estimate of test reliability and does not accurately reflect the internal structure of a test. It is a lower bound to reliability, often an underestimate, and is unrelated to the test's internal consistency. Alpha is frequently misinterpreted as a measure of internal consistency, leading to incorrect assumptions about the test's validity. The paper highlights that statistics based on a single test administration provide limited information about individual test performance accuracy. It also discusses the availability of better reliability estimates, such as the greatest lower bound (glb), which provides a more accurate reliability estimate. The paper concludes that alpha is not a valid measure of internal consistency or unidimensionality, and that better reliability estimation methods should be used. The paper also emphasizes the importance of considering the limitations of single-administration data in assessing individual measurement accuracy. Overall, the paper argues that alpha is not a reliable or useful measure of test reliability or internal consistency and that better alternatives should be used.