The Likert scale is widely used in industry and research, but its suitability for managerial practice is rarely discussed. Researchers in various fields of business use those in analyses such as comparative statistics, t-tests, regression, and other statistical models. Whether it could lead to meaningful and reliable conclusions for decision makers without potential biases might be the subject of debate. This study makes an attempt to describe how numeric or quantified verbal answers of respondents relate to their actual attitudes. It proposes a simulation model using a beta distribution of respondents’ answers and the power function of utility values. The objective is to evaluate the robustness of approaches to treating Likert scales using the differences between the mean and median of responses. Numerical experiments reveal how alternative distributions of responses in real business can lead to discrepancies, which could have implications for statistical methods involving Likert scales and their modification used in management research. The choice of analysis method should align with the purpose of the research, as different interpretations can arise from using means versus other statistical measures. This guidance is as crucial for accurate interpretation in business as in other contexts.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Statistical Analysis of Responses to Likert Scale Questions in Management Science: Simulation Model Using Utility Function and Beta Distribution

  • Berdymyrat Ovezmyradov,
  • Akbar Jumayev,
  • Bunyod Utanov,
  • Ekaterina Gromova

摘要

The Likert scale is widely used in industry and research, but its suitability for managerial practice is rarely discussed. Researchers in various fields of business use those in analyses such as comparative statistics, t-tests, regression, and other statistical models. Whether it could lead to meaningful and reliable conclusions for decision makers without potential biases might be the subject of debate. This study makes an attempt to describe how numeric or quantified verbal answers of respondents relate to their actual attitudes. It proposes a simulation model using a beta distribution of respondents’ answers and the power function of utility values. The objective is to evaluate the robustness of approaches to treating Likert scales using the differences between the mean and median of responses. Numerical experiments reveal how alternative distributions of responses in real business can lead to discrepancies, which could have implications for statistical methods involving Likert scales and their modification used in management research. The choice of analysis method should align with the purpose of the research, as different interpretations can arise from using means versus other statistical measures. This guidance is as crucial for accurate interpretation in business as in other contexts.