“…We found that 22 (or, just over half of the) articles rely on new data, be it from surveys, interviews (with criminals and law enforcement), court records, case studies, underground markets or honeypots [3], [7], [8], [10], [11], [24], [29], [31]- [36], [43]- [45], [48], [49], [52], [57]- [59]. The other studies primarily use existing literature to form their arguments and often rely on research from other fields, or established approaches mapped to the cybercrime domain [2], [4], [5], [9], [12], [13], [15], [18], [21], [25], [27], [46], [50], [53]- [56]. This balance is interesting to note as it highlights the extent of articles on profiling that are not working with primary data.…”