Adjacency Matrix (AM) is one of the commonly used techniques to visualize networks. While an AM provides a clean and compact representation for dense networks, several studies have shown that it is not suitable for path-related tasks. Several visualization techniques have been proposed to address this limitation. This thesis goes in the same direction by investigating the influence of the rotation of the matrix and the visual encoding of links on the user performance at different network analysis tasks. To this end, a crowd-sourced user study was conducted to evaluate different variants of AM across several network analysis tasks and different network properties. The results reveal that the accuracy is not significantly affected by the rotation of the matrix, as well as the visual encoding of links, across all tasks. This means that users achieve similar accuracy on all visualizations, indifferent of the rotation of the matrix or the visual encoding of links. Several isolated cases exist, where both, the VCD and ArcM record a visibly higher accuracy than the AM. However, the difference in accuracy is deemed not significant in all those cases. For example, visibly higher accuracy is recorded on the VCD as well as on the ArcM than on the AM for the task concerning the detection of mirror symmetry, across most networks. However, the difference in accuracy is deemed not significant across those networks. The results also reveal that the answer time is a critical factor. The rotation of the matrix, as well as the visual encoding of links, lead to significant degradation of answer time across most connectivity-related tasks. As for the path-finding task, the answer time is not significantly affected by both, the rotation of the matrix and the visual encoding of links.