A new method, based on the presentation of initial data in the form of probability circles (PC), has been developed for constructing a calibration curve, which shows a monotonic dependence with respect to the given concentration. The centre of the reference probability circle is defined by its mean value and the radius of the circle is calculated as the value of the standard deviation of the sampling considered. The comparative probability circle is defined by the same corresponding parameters but rotated, relative to the initial reference circle, by an angle, which is related to the Pearson's correlation coefficient (PCC). The two parameters of the PCC and the statistical proximity factor (PCF), which defines the positions of the centres of the circles relative to each other, can be chosen as statistical parameters for the construction of the desired calibration curve. Experiments realized with the mixture of two liquids (chloroform serves as the basic matrix) and acetone (serves as an additive) confirm the efficiency of this new analytical method and demonstrate a possible increase sensitivity for the detection of lower concentration limit by approximately one order of magnitude. This new approach, which is free from model assumptions, and having very clear geometrical meaning, can be applied for different types of spectra and has many potential applications in the construction of calibration curves for different additives embedded within different matrices.