This study examined the reliability of data obtained from online surveys, which are increasingly used as a survey method in marketing research and public opinion polling. In this study, an online survey and an offline survey were administered to the same panel members using an identical questionnaire with a two-week interval, and the reliability between the two datasets was assessed using test-retest reliability. The analysis results revealed that online survey data exhibited a very high correlation with offline survey data, confirming that its reliability is at an acceptable level. However, for Likert-scale items composed of multiple sub-items, it was found that reliability may vary depending on their formatting or their position within the questionnaire. Therefore, it was demonstrated that in online surveys, relatively greater attention must be paid to questionnaire design and formatting compared to offline surveys.