Improving Survey Data Interpretation: A Novel Approach to Analyze Single-Item Ordinal Responses with Non-Response Categories

Abstract

Questionnaire data plays a key role in social research, especially when evaluating public attitudes using Likert-type scales. Yet, traditional analyses often merge some ordinal categories and exclude responses such as Don’t Know, No Answer, or Refused—risking the loss of valuable information. This study introduces BS-TOSIE (Belief Structure-Based TOPSIS for Survey Item Evaluation), a novel method that preserves and integrates all response types, including ambiguous ones. By combining the Belief Structure framework with the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method, BS-TOSIE offers a structured approach to ranking and evaluating individual survey items measured on an ordinal scale, even in the presence of missing or incomplete data. Response distributions are transformed into a belief structure vector, enabling comparison against ideal and anti-ideal benchmarks. We demonstrate this approach using data from the Quality of Life in European Cities survey to assess perceptions of local governance in European cities. This study analyzes changes in citizen satisfaction with local public administration across five key dimensions—timeliness, procedural clarity, fairness of fees, digital accessibility, and perceived corruption—in 83 European cities between 2019 and 2023. The findings reveal persistent regional disparities, with Northern and Western European cities consistently outperforming those in Southern and Eastern Europe, although some cities in Central Europe show signs of improvement. Zurich consistently received high satisfaction scores, while other cities, such as Rome and Palermo, showed lower scores. Unlike traditional methods, our approach preserves the full spectrum of responses, yielding more nuanced and interpretable insights. The results show that BS-TOSIE enhances both the clarity and depth of survey analysis, making a methodological contribution to the evaluation of ordinal data and offering empirical insights into public perceptions of local city administration.