Total uncertainty can be discussed in terms of bias and reproducibility. We have studied bias and reproducibility in the critical dimension measurement scanning electron microscope and measurements have derived expressions of them, and discussed the parameters which determine them by making a comparison with measurement data.