Variables that influence the bias of numerical cognition

Purpose: It is known that individuals make decisions in a not entirely rational way, which leads to wrong decisions. People rely on a limited number of heuristic principles and cognitive biases, which reduce the complex task of setting up probabilities and predicting values in judgments. In this context, this paper aimed to identify the relationship between the bias of numerical cognition in managerial decision-making and the variables age, gender, education and mathematical knowledge. Methodology: A self-applying questionnaire which contains made-up situations of a management environment was utilized, aiming to know the answerers' perception of the real magnitude of the numbers. The research was operated in two stages: in the first, the presence of numerical cognitive bias was verified; in the second, the influence of the mentioned variables. The analysis was performed with a sample group of 252 academics (160 from a so-called "managerial" group and 90 from the named "non-managerial" group). For testing the presence of the numerical cognitive bias, ANOVA was used, and, for the statistical inference, the logistic regression, once it allows the explanation of the dependent variable in relation to the independent variables. Results: The results indicate that the numerical cognitive bias is present in both groups (managerial and non-managerial), not showing statistically significant difference, which corroborates the idea that the cognitive bias is inherent to the human beings, regardless its involving in managerial practices. It was also observed that male and older answerers are less prone to the effects of the numerical cognitive bias. Contributions of the Study: The study helps to build a new paradigm for evidences based on perfect rationality and on markets' efficiency for the purpose of understanding the choices and emotions of the decision-makers, elaborating a theoretical-descriptive framework for the decision-making process.