Statistical Design of a Sampling Method for Quality Control of Research Data

In scientific documents, indexing and quality control are key processes that, when performed correctly, enable effective retrieval in search engines. Although sufficient attention has been given to sampling methods in physical products in scientific resources, there has been little work in the area of data, especially research data. This study provides a framework for sampling quality control processes for data. As a case study, the research data from the National Thesis/Dissertation Information Dissemination Database (PARS) of graduates across the country (GANJ) was selected. 

The results indicated that, given the acceptable quality of many information items in PARS after registration, sampling is a vital task for enhancing the efficiency of the information organization and analysis unit. The OC curve for various designs shows that the proposed designs are effective for assessing the quality level of research data. The framework presented in this study can be localized for various data-driven organizations, especially data-based businesses.