The prototypical analysis of evocations is one of the main techniques that allows the exploration of the structure of social representations from verbal data. The present study aims to evaluate the effect of variations in the size of the sample in the configurations of the results of the prototypic analysis from free evocations. Subsamples were randomly selected from a database with 469 participants with progressively smaller sample sizes (200, 100, 50 and 25 cases), and compared the results settings of the first quadrants with respect to the reference sample. The results show that the samples of 200 and 100 cases have patterns close to the total sample in terms of agreement with the composition of the first quadrant. In addition, samples with 50 and 25 participants are not recommended because they have high variability and low coincidence with the larger sample.