Writing rubrics have been used in doctoral dissertations in Turkey to assess student writing for nearly twenty years. This study aims to determine which features are assessed in the rubrics used in doctoral dissertations. Twenty-five rubrics were selected to determine the analysis of validity and reliability, rubric dimensions, features of criterion descriptions, and performance levels. The results showed that researchers did not attach much importance to the validity of rubrics, and some studies provided insufficient information about reliability. Rubrics typically have three and five performance levels. As a result of the content analysis based on the descriptions, nine categories (genre-oriented content, organization, style, presentation, mechanics, vocabulary, sentences, textuality, process) and 104 features were identified. The frequency of these features was 567. According to the emerging categorization in the present study, we observed that the rubrics used in doctoral dissertations were similar to the criteria in the 6 +1 Trait Writing Model. However, the descriptions were found to be more straightforward and concise. The features in the rubrics were similar to one another yet differed depending on the purpose, type, and length of descriptions. Some of the rubrics had descriptions that were ill-suited to the nature of the rubric. The results were discussed, presenting recommendations.