Objective: The objective of this study was to compare the cognitive performance of elderly who are enrolled in a physical activity (PA) program with those who are not, considering some sociodemographic variables that are related to the practice of cognitive activities. Method: 59 elderly people participated in this study divided into control group (CG) and gymnastics group (GG). The Vienna Test System® was used to estimate the attention and concentration capacities, simple reaction time and peripheral perception. To test the normality of the data it was used the Shapiro Wilk Test. The association between PA and socioeconomic variables was verified by Chi-Square Test and Fisher Exact Test. To compare of differences between groups the Mann Whitney U-test for variables with non-parametric distribution and the Student T-test for independent samples for other variables were used. Result: There were no significant differences in the performance of the groups, except in the capacity of attention and concentration in which CG presented best performance (p=0.01). Factors as schooling (p=0.02), income (p=0.001) and regular practice of handcrafts (p=0.06) can explain the best performance of CG. Conclusion: This study found no evidence that the practice of physical activity can generate some cognitive benefit in the elderly when compared to the elderly who did not practice regular physical activity. However, we perceive the existence of other aspects that influence on the cognition, such as schooling, sociocultural level and the reading habit, that have a significant importance degree in the analysis.