Does data curation matter in citation and co-citation analysis? Evidence from a top service journal

Bibliometric scholars have primarily evaluated massive data without refining any potential typing and/or spelling errors, resulting in two constraints: misinterpretation of findings and misleading future research in the knowledge domain. Thus, this study aims to introduce the data curation approach in order to reduce these restrictions. Utilizing a renowned service journal (Journal of Service Research) as the study sample, we first acquired all published papers and then constructed raw and clean datasets. We ran citation and co-citation analyses on these datasets separately. Our investigation reveals that clean data yielded more trustworthy and valid results than raw data with redundant references. This study provides an answer to how and why data in bibliometric analysis needs to be cleaned. It thus contributes to the literature by suggesting a new route for scholars to improve the accuracy and reliability of their bibliometric findings.