Quality in the process of scientific production in accounting in Brazil

• Ano de defesa: 2016
• Autor(a) principal: Oliveira, José Renato Sena
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Accounting - Research; Contabilidade - Pesquisa; Graduate studies; Pesquisa científica; Pós-graduação; Produção científica; Scientific production; Scientific research
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/12/12136/tde-12072016-102123/

This study aims to analyze the building process of scientific production in accounting in light of the quality attributes of good research. In so doing, it attempts to identify the stages of the research process in which these characteristics are revealed, to compare the literature on research attributes with the attributes identified in respondents\' practices, and to present an approach based on the attributes of good research to judge the quality of scientific production in the accounting field. This research uses the Modified Delphi Technique, which indicates that the first inputs came from the literature. The expert panel consisted of 37 faculty members from 19 Brazilian graduate programs in accounting who were recommended/recognized by the CAPES Foundation. Based on the literature, especially Brinberg and McGrath (1985), Spencer et. al. (2003), and Mays and Pope (2006), an orientation matrix was developed with 53 attributes/relationships related to general quality criteria and nine key features. Experts gave each proposition a grade from 1 to 10 based on their level of agreement regarding adherence to their research practices. The Delphi was applied in two rounds using online questionnaires with customized access. The findings reveal that most of the respondents obtained their doctoral degree in accounting at a national institution other than the one at which they currently work and that more than 70% of the respondents have been working as teachers or coordinators in graduate programs for fewer than seven years. With respect to the respondents\' research experience, most serve on journals\' editorial boards, act as journal referees, and at some point have obtained research funding from development institutions. Approximately one-third have received research productivity grants and almost one-half either currently serve or has served as a journal editor. Approximately 3/4 of the propositions achieved a strong level of agreement, and the following 10 propositions achieved a Delphi relative score of more than 90%: voluntary participation of subjects, goal/problem shown precisely, confidentiality of participants\' data, conclusions versus aim, results comparison with other studies, checklist of findings versus purpose, literature review versus main concepts, theory to support propositions, useful strategy for purpose, and previous findings versus hypotheses. Those attributes that demonstrated low or moderate levels of agreement involved items that may compromise the quality and integrity of research, including those related to ethical principles, demonstrations of how researchers addressed errors and biases, and disclosure of the impact of the research team\'s participation on the results. The approach chosen meant discussing the relationships between criteria and fundamental features versus levels of agreement, followed by presentation of a logic model to evaluate the research process. The research concludes that certain current practices employed by Brazilian researchers in accounting - combined with the local institutional environment - contribute to reduce the quality of accounting research. This position is supported by the high dispersion of answers on various items and the low acceptance of attributes related to ethics, both of which are mandatory requirements under Brazilian law. Additionally, the low level of agreement on issues regarding the criteria of rigor and internal validity/credibility or defensibility - in addition to items related to rigor, integrity, and feasibility - reached only the moderate level.