In this paper, we investigate the effect of face modalities on each other. Analysing the effect of face modalities is a difficult research problem, because of the lack of publicly available annotated databases, in which each sample has labels for each face mode. We selected gender and age face modes to analyse their effect between them. The database is divided into groups uniformly and there is no overlap between training and testing sets. Same amount of training samples is used to train each model to avoid the effect of having different amount of training samples to the results. According to obtained results, utilising the correlation between the face modes provided better results than using the face modes alone.