Which choice describes a downside of using cross-validation?

Using cross-validation is a method that enhances the reliability of model evaluation by splitting the dataset into subsets. The selected answer highlights an important aspect of this process: the computational burden it introduces.

When cross-validation is employed, particularly k-fold cross-validation, the model must be trained multiple times, once for each subset (or fold) of the data. This means that if a model is trained on a large dataset with many folds, the amount of computation required increases substantially. Each training and evaluation cycle can significantly add to processing time and resource usage, making it a labor-intensive approach compared to simpler validation methods, like a single train-test split.

While the other aspects such as bias, dataset size requirements, and acceptance of the method are important considerations in research methodology, they do not directly address the specific characteristic of increased computational demands that arises from the repeated training involved in cross-validation. Therefore, focusing on the computational implications provides a clear understanding of this downside.