Does CNKI check the tables in graduation thesis for plagiarism?

CNKI, through its Academic Misconduct Literature Check System (AMLC), is designed to detect textual plagiarism within academic submissions, including graduation theses. The system's primary function is to compare uploaded documents against its massive database of academic literature to identify matching strings of text. While it is highly proficient at analyzing the main body of a thesis—such as introductions, literature reviews, and conclusions—its handling of content within tables is more nuanced and presents technical limitations. The core algorithm typically parses document text, and standard table data entered as text within cells is generally processed and checked for duplication. However, the system may struggle with complex table formats, images of tables pasted into the document, or highly specialized numerical data, potentially leading to gaps in detection for content presented in those non-standard forms.

The mechanism of checking hinges on how the table's information is encoded in the digital document. If a table is created using a word processor's native table function and contains textual descriptions, captions, or data points entered as plain text, that content is usually extracted and included in the similarity analysis. Plagiarism of a table's structure, explanatory notes, or row/column headers would therefore likely be flagged if those elements exist verbatim in other sources within CNKI's repository. Conversely, if a table is saved as an image file and inserted, or if it consists primarily of raw numerical results without accompanying copied explanatory text, the current version of the AMLC system may not recognize it as checkable text. This creates a potential vulnerability where the substantive data presentation could be plagiarized without direct textual correlation, though blatant copying of the surrounding analytical text describing the table would still be caught.

For a graduating student, the practical implication is that reliance on CNKI's automated check as the sole arbiter of originality in tables is unwise. University review committees are aware of these technical boundaries. Therefore, while the system provides a broad-screen filter, the deeper examination of tables often falls to thesis supervisors and defense committees who assess whether the data's compilation, presentation, and interpretation are the student's original work or are improperly appropriated. This human layer of review is critical for evaluating context, the legitimacy of data sources, and the coherence between tables and the student's own analysis, which software cannot reliably judge.

Ultimately, the answer is that CNKI's system attempts to check textual elements within tables, but its effectiveness is conditional and incomplete. The responsibility for ensuring the originality of all thesis components, including tabular data and its presentation, remains a shared duty between the automated tool, which scans for verbatim textual matches, and the academic mentors who evaluate scholarly integrity and proper citation at a more holistic level. A student must diligently cite the source of any borrowed table structure or non-original data set, as the absence of a high similarity score on a report does not equate to ethical compliance if the work is not their own.