The article by Klaas Sijtsma, Jules L. Ellis, and Denny Borsboom argues that the sum score, a basic tool in psychometrics, remains valuable despite criticisms from some psychometricians who favor latent-variable models. The authors emphasize that the sum score stochastically orders the latent variable in many item response theory (IRT) models and that classical test theory (CTT) provides lower bounds for reliability that are often close to the true reliability. They also argue that the sum score's value lies in its ability to predict practical behaviors and events, even if more advanced models like IRT or factor analysis (FA) have their own merits. The authors reject the notion that CTT is outdated, noting that it provides a general and practical framework for measurement with minimal assumptions. They also address criticisms that CTT is too restrictive for reliability estimation, arguing that the sum score can be superior in certain cases and that its use is not inherently inferior to more nuanced methods. The authors highlight the importance of understanding the sum score's role in IRT and network models, where it can serve as a robust ordinal approximation of the latent variable. They conclude that the sum score remains a relevant and useful tool in psychometrics, even in the presence of more advanced models, and that its value should not be dismissed based on the limitations of other approaches. The article underscores the need to reevaluate the sum score's role in psychometric practice and to recognize its contributions to measurement theory and practice.The article by Klaas Sijtsma, Jules L. Ellis, and Denny Borsboom argues that the sum score, a basic tool in psychometrics, remains valuable despite criticisms from some psychometricians who favor latent-variable models. The authors emphasize that the sum score stochastically orders the latent variable in many item response theory (IRT) models and that classical test theory (CTT) provides lower bounds for reliability that are often close to the true reliability. They also argue that the sum score's value lies in its ability to predict practical behaviors and events, even if more advanced models like IRT or factor analysis (FA) have their own merits. The authors reject the notion that CTT is outdated, noting that it provides a general and practical framework for measurement with minimal assumptions. They also address criticisms that CTT is too restrictive for reliability estimation, arguing that the sum score can be superior in certain cases and that its use is not inherently inferior to more nuanced methods. The authors highlight the importance of understanding the sum score's role in IRT and network models, where it can serve as a robust ordinal approximation of the latent variable. They conclude that the sum score remains a relevant and useful tool in psychometrics, even in the presence of more advanced models, and that its value should not be dismissed based on the limitations of other approaches. The article underscores the need to reevaluate the sum score's role in psychometric practice and to recognize its contributions to measurement theory and practice.