Estimating missing values is known as data imputation. Traffic data imputation has a long history, and such a convention can be traced back to the starting date of traffic monitoring systems in 1930s. Literature review indicates that many highway agencies in North America and Europe estimate missing values in their collected data. However, no studies have been found to assess the accuracy of these imputation methods. This study indicates that the imputation methods used by highway agencies are intuitive in nature and could result in large imputation errors under certain circumstances. The large imputation errors could lead to significant deviations in the resulting operation plans and designed structures. Therefore, it is necessary to evaluate the accuracy of various imputation methods employed by highway agencies. The assessment is carried out in this study based on data from the provincial highway agencies of Alberta and Saskatchewan in Canada. Study results clearly show some imputation models can consistently provide more stable and accurate estimates than the others. It is believed that study results are helpful for traffic engineers to review their imputation practices and hence, improve their data quality.