Evolutionary algorithms (EAs) are optimization algorithms inspired by biological evolution and have been widely applied for solving various optimization problems. When EAs are applied to solve optimization problems, it is very important to set adequate parameters taking into account the characteristics of an objective function. Among the characteristics of the objective function, landscape modality is particularly important. For example, if the landscape is unimodal, the efficiency of EAs can be improved by tuning control parameters to local search. In contrast, if the landscape is multimodal, the robustness of the EAs can be improved by tuning control parameters to global search. However, in a black-box optimization problem, the function is unknown. In this study, we propose a novel method that estimates the modality of landscape being searched: simple unimodal or not. In our method, two rankings according to the fitness and distance are given for each individual of EAs. Next, Spearman's rank correlation coefficient is calculated using the two rankings. The value of rank correlation coefficient becomes different, depending on the landscape modality. Using this feature, we classify the type of function modality being searched by population. Through experiments using basic benchmark problems, we show that the proposed method can correctly perform modality estimation.