Despite the inumerous attempts to correlate psychological tests used for evaluating drivers and some types of traffic-related accident rates, up until now, every instrument dedicated exclusively to the quantification of specific skills, mental level or personality has proved itself inconsistent or ineffective. This finding suggests, new ways of assessing drivers, such as measuring human behavior focused on security. These more complex behaviors, according to the Risk Homeostasis Theory (RHT) relate, mainly, to the continuous adaptation to instantly changeable difficulties of the environment. If the main assumption of the theory is correct, in traffic locus, when the competence in a particular expert behavior increases, as the reasoning speed in traffic, the sense of insecurity will decrease and thus, the very caution or attention will fade, keeping the constant level of risk on the wheel.

Given the scenario presented, this work aimed to investigate the existence of risk homeostasis in the context of Brazilian traffic. In this sense, we sought to relate two major psychological skills currently evaluated in aspiring driver's license in Brazil, the attention and reasoning, and also discuss methodological alternatives for future training programs.

Making use of the first data collection performed to the standardization of the instrument SAPIEM (Integral Psychological Assessment System for Drivers Examination), the performance of volunteer drivers was measured in tasks that required their attention or reasoning. The tasks, simply, consisted counting the number of stimuli or deduct the routes traced by moving virtual vehicles displayed on a tablet.

In total, complete data from 89 subjects were collected in three Brazilian cities. These participants exhibited an average score equal to 9:54 (maximum of 10) in the attention task, and 7:56 (maximum of 10) in the reasoning task. The correlation between the two activities was r = -0.13.

Even though the correlation found between the attention and reasoning tasks was low it was still negative, supporting the Risk Homeostasis Theory. In this sense, according to the literature review, in this case it is not recommended the isolated use of any of these properties to measure the risk in a traffic context. It is suggested that an alternative solution is the use of tools designed to evaluate road safety, resulting from a set of compensating skills focused on the recognition of real risks.