Statistical graphs have been designed for accessible use by visually impaired users. Haptic devices provide an appropriate interface for haptic exploration of statistical graphs. However, haptic exploration of statistical graphs reveals a more local and sequential inspection pattern compared to visual exploration. This difference between haptic exploration and visual exploration is usually attributed to different characteristics of the exploration processes, such as bandwidth of information extraction. To facilitate information extraction from statistical graphs, alternative sensory modalities have been employed. In particular, line graphs have been represented by sound, thus leading to sonified graphs. Despite their demonstrated facilitating effects, sonified graphs have limitations under complex line representations. One method of overcoming those difficulties is to develop a verbal assistance system for haptic line graph comprehension. In the present article, we summarize our studies on designing and developing a verbal assistance system for haptic line graph comprehension. We present the findings in a set of studies conducted with blindfolded and visually impaired participants.