Within research on the political influence that social network members exert on one another, some studies rely on information obtained directly from different members in the network separately (self-reported measures), while others rely on information obtained from one key informant within the social network (measures based on perception). We investigate the difference between these self-reported and perceived measures by analyzing the correspondence of voting intentions within the family. On the one hand, we examine this correspondence using information obtained from only one family member. On the other hand, we use the self-reported measures obtained from all family members separately. We use data from the Parent-Child Socialization Study (PCSS), a survey conducted among 2,085 mothers, fathers and children in the Flemish region of Belgium (2012). Our analyses suggest that using perceptual measures could lead researchers to different or even opposite conclusions than using self-reported measures from all individual respondents.