Visualizing social network data into sociograms plays an important role in communicating information about network characteristics. Previous studies have shown that human perceptions of network features can be affected by the layout of a sociogram [McGrath et al. 1996, 1997]. An empirical user study has been conducted to investigate effectiveness of five different network visualization conventions and impact of edge crossings on sociogram perceptions, using both quantitative performance and preference measures and qualitative questionnaire study. This paper reports results and findings of the questionnaire study. We relate qualitative questionnaire results with quantitative findings and discuss their implications for sociogram design. We found that subjects had a strong preference of placing nodes on the top or in the center to highlight importance, and clustering nodes in the same group and separating groups to highlight groups. They had tendency to believe that nodes in the center o...