Purpose: This study intends to evaluate the service quality of academic digital libraries(DLs)in China. By utilizing tetra-class model which concentrates on categorizing services according to their contributions to user satisfaction, this paper attempts to visually categorize the specific DL service elements to reveal their present performances, and then further explain the categorizing variations among different groups of users to discover the user preference.Design/methodology/approach: This paper carries out a survey to evaluate user experience on 27 typical DL services summarized from our investigations of representative Chinese university DLs. Based on the five-point Likert-type scale evaluation, the users’ attitudes toward specific service element are divided into negative and positive dimensions. Afterwards,a correspondence analysis is applied to calculate the contributions to satisfaction and dissatisfaction of each service element based on tetra-class model. As a result, the DL service elements of Chinese academic libraries are classified into four categories(i.e. Basic, Secondary,Plus, and Key). Finally, we compared the categorizing variations.Findings: The results show that the DL service elements of Chinese academic libraries are all distributed in Basic and Key services regarding information retrieval and informationorganizing; 80% of the interaction services elements are Plus services, while 50% of the Secondary services are information-providing services. The results also reveal that service categorization is obviously influenced by the users’ education background, especially their disciplines. Furthermore, the users who are older, more highly-educated, or studying in higher reputation universities are more likely to evaluate DL services as either critical or useless.Research limitations: Tetra-class model cannot reveal the interplay among the DL service elements. In addition, the user segmentation in our studies is limited to the sample structure.Practical implications: This empirical study focuses on the evaluation of DL services of academic libraries in China, the analyses of their current performances could provide useful reference for the assessment of other types of Chinese DLs. Moreover, the consideration of user characteristics(gender, age, and education background, etc.) in the DL evaluation would help librarians improve DL services to meet the users’ various needs in teaching and doing scientific research.Originality/value: Different from the frequently-used factor analysis which focuses on the relationship among factors and user satisfaction, this paper tries to use and compare element distributions of different user segments while focusing on various service objectives. Factor analysis shows some flaws as used to measure the element with selected indicators, for it ignores the fact that the indicators which measure the same factor would have different degrees of impacts on user satisfaction. However, the tetra-class model can better visually analyze the performance of each DL service element from its contributions to satisfaction and dissatisfaction, which would help librarians to better understand users’ need and offer DL services more efficiently. Purpose: This study intends to evaluate the service quality of academic digital libraries (DLs) in China. By utilizing tetra-class model which concentrates on categorizing services according to their contributions to user satisfaction, this paper attempts to℠ categorize the specific DL service elements to reveal their present performances, and then further explain the categorizing variations among different groups of users to discover the user preference. Design / methodology / appach: This paper carries out a survey to evaluate user experience on 27 typical DL services representative Chinese university DLs. Based on the five-point Likert-type scale evaluation, the users’ attitudes toward specific service element are divided into negative and positive dimensions. Afterwards, a correspondence analysis is applied to calculate the contributions to satisfaction and dissatisfaction of each each service element based on tetra-class model. As a result, the D Finally, we compared the categorizing variations. Findings: The results show that the DL service elements of Chinese academic libraries are all distributed in Basic and Key services regarding information retrieval and informationorganizing; 80% of the interaction services elements are plus services, while 50% of the Secondary services are information-providing services. The results also reveal that service categorization is obviously influenced by the users’ education background , especially their disciplines. Furthermore, the users who are older, more highly-educated, or studying in higher reputation universities are more likely to evaluate DL services as either critical or useless. Research limitations: Tetra-class model can not reveal the interplay among the DL service elements. In addition, the user segmentation in our studies is limited to the sample structure. Practical implications: This empirical study on the evaluation of DL services of academic libraries in China, the analyzes of their current features could provide useful reference for the assessment of other types of Chinese DLs. Moreover, the consideration of user characteristics (gender, age , and education background, etc.) in the DL evaluation would help librarians improve DL services to meet the users’ various needs in teaching and doing scientific research.Originality / value: Different from the frequently-used factor analysis which focuses on the relationship factors and user satisfaction, this paper tries to use and compare element distributions of different user segments while focusing on various service objectives. Factor analysis shows some flaws as used to measure the element with selected indicators, for it ignores the fact that the indicators which measure the same factor would have different degrees of impacts on user satisfaction. However, the tetra-class mod el can better visually analyze the performance of each DL service element from its contributions to satisfaction and dissatisfaction, which would help librarians to better understand users’ need and offer DL services more efficiently.