Metodologia de avaliação de serviços públicos: o atendimento ao público em órgãos de receita

Abstract

Taxes are a certain presence in organized society. Just as it is certain that there will be taxes, there will be a public agency responsible for collecting them, and this body must know how to relate to people and society. Much of this relationship occurs through customer service delivered in person, so this service should be of high quality. Quality care helps promote the spontaneous fulfillment of tax obligations, avoids legal litigation, and enhances the institution’s image. To ensure this desired quality, periodic service evaluations and continuous improvement processes are required. Although there were validated scales for the evaluation of public services, such as SERVQUAL and SERVPERF, the theoretical gap of a specific scale involving collection agencies persisted. The general objective of this work was, therefore, to adapt the service evaluation methodology to the public attendance service provided by the collecting agencies. For this purpose, a quantitative approach research was carried out, applying paper forms with the SERVPERF model to taxpayers who visited the Taxpayer Service Center of the Federal Revenue Service of Brazil in Santa Maria, Rio Grande do Sul, during the period from June 26 to September 22, 2017, obtaining 404 valid questionnaires. A qualitative research was also conducted, analyzing the writing left by the respondents in the quantitative survey forms, and interviewing four public service employees that worked in the customer's service in that period. The interviews were conducted between September 13th and September 18th, 2017. Were carried out a descriptive analysis of frequencies and averages, exploratory factorial analysis and the measurement of the reliability of the research using the Cronbach’s Alpha coefficient, concluding that the model was better structured using only three dimensions instead of five, namely: Organization, Assistants and Environment. After analysis, six affirmations of the PERVPERF model were excluded and six were included, leaving the final model with 22 questions. There are only three dimensions, the Organization dimension has seven questions, the Attendants dimension has nine questions and the dimension Environment has six questions. The final score on the five-point Likert scale service was 4.5521, with a standard deviation of 0.48824. T tests and ANOVA tests were also performed, which demonstrated that characteristics of the respondents' profile, such as representing an individual or a legal entity, representing themselves or others, income, age, education level, among other characteristics, have a significant influence (95% confidence) in how they rate customer service generally or at least to one of the dimensions of the new model.