A influência do item "não se aplica" na estimação do grau de maturidade de empresas quanto as tecnologias de gestão através da TRI

Abstract

With the current competitive companies seek constant updating of its technologies, models of organization and management, among others. For the good performance of activities logistics of an organization is necessary to evaluate its performance in the market and one of the logistics performance evaluation possibilities is benchmarking. The Monitoring System Productive Arrangements, which aims to the current competitive benchmarking, It has a corporate database in relation to the management tools applied in each one. Where questionnaires are used to collect data is unlikely that all respondents answer all items. In this study, because of the diversity of links and supply chains of companies, these items unanswered items were considered "not applicable ". However, some items were also considered in this category because they were added to the questionnaire later, generating this kind of response for companies respoderam before this change. This study sought to evaluate the influence of the item "not applicable "in the estimation of the degree of maturity of registered companies in SIMAP. The data used in the work are from 46 items evaluated in 238 companies. Data were analyzed using Item Response Theory, particularly the model of gradual scale. The five practical approaches to assess the impact of this item were: are considered all items; are excluded items (33, 34, 35, 36, 37, 38, 39) wherein at least 20% of the the companies considered as not applicable (remained 39 items); excludes items (6, 17, 33, 34, 35, 36, 37, 38, 39) wherein at least 15% of the companies considered as non- applicable (remained 37 items); excludes those items (6, 11, 17, 20, 32, 33, 34, 35, 36, 37, 38, 39) in which at least 10% of the companies considered as not applicable (remained 34 items); excludes those items (4, 5, 6, 8, 11, 13, 17, 18, 19, 20, 29, 30, 32, 33, 34, 35, 36, 37, 38, 39) in which at least 5% of the companies considered as not applicable (26 items). The approach to all items (46 items) showed the best estimates, as revealed smaller standard errors for the power of item discrimination, item difficulty and degree maturity of the companies. In spite of all these approaches present similar values for estimates and maturity, these differences were responsible for the change in ordering the discrimination power of the items and the position of the companies on the scale of maturity. Thus, the results shown in the study had an indication that the withdrawal items considered "not applicable"is associated with the loss of efficiency of the estimates.