Correlation and regression models applied to public bidding procedure by brazilian universities: analysis of electronic bidding between 2011 and 2016

To investigate the relationships between the variables that influence the electronic auctions held in the three federal universities, the research used descriptive statistics, correlation tests and linear regression to analyze the data. For UFSM (p=0.047) and UNIPAMPA (p=0.026), there was a signific...

Descripción completa

Detalles Bibliográficos
Autores: Gomes, Thiago Eliandro de Oliveira, Zanini, Roselaine Ruviaro
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2023
País:Brasil
Institución:Universidade Federal de Santa Maria (UFSM)
Repositorio:Revista Práticas de Administração Pública
Idioma:portugués
OAI Identifier:oai:ojs.pkp.sfu.ca:article/69209
Acceso en línea:https://periodicos.ufsm.br/pap/article/view/69209
Access Level:acceso abierto
Palabra clave:Public bidding
University
Electronic auction
Correlation
Regression
Licitação pública
Universidade
Pregão eletrônico
Correlação
Regressão
Descripción
Sumario:To investigate the relationships between the variables that influence the electronic auctions held in the three federal universities, the research used descriptive statistics, correlation tests and linear regression to analyze the data. For UFSM (p=0.047) and UNIPAMPA (p=0.026), there was a significant correlation for the number of small companies bidding services and for the number of large bidders in the case of UFRGS (p=0.019). In addition, there was a decrease in the number of companies participating in the bids of the three institutions. With tests to compare the equality between the intercepts and the angular coefficients, the sample evidence of coincidence of the estimated straight for the models was observed, indicating parallelism (p < 0.05) and that the temporal evolution is similar between the institutions. Such results are granted by the addition of analytical models through measurements over the years in the electronic auction of different institutions. On the other hand, the regression model adjusted by expense type has a disadvantage, as it failed to detect changes in the average number of trading sessions. As a buyer, however, the public administration benefited from the proven effectiveness of price reductions in the bidding process.