Abstract

The Likert scale is a measurement tool commonly used in various fields including social sciences such as psychology. Although the Likert scale is used with the assumption that the intervals between each scale point are equal, respondents may not perceive the intervals between each scale point equally, depending on the wordings used in scales (e.g., ‘slightly’, ‘very’). This can lead to systematic errors in responses and biases in analysis and estimation of results. In this study, we empirically explored and proposed anchor wordings suitable for use in each point, considering the errors induced by the wordings. Specifically, we investigated what anchor wordings have been commonly used in the Likert scale to measure ‘consent or accordance’ and ‘frequency’. We calculated and proposed what anchor wordings are most suitable for each scale point. In the preliminary survey, we collected and identified wordings of scale points used in recently published studies (January 2020 - December 2022). In the main survey, we recruited participants and had them rate the meaning of each anchor wordings obtained from the preliminary survey using slide scales. The preliminary survey found that 32 anchor wordings of consent and 26 of frequency were frequently used in recent studies. In the main survey, we proposed what anchor wordings are suitable for each scale point on the 4-, 5-, and 7-point scales based on the values that participants responded to what each wordings represent.