https://phys.org/news/2018-11-terrorism-struggle-primary.html

Unfortunately, a relatively small group of dedicated scholars was not the only obstacle to a timely resolution of the primary data problem. The long and short of it is that terrorism is in many instances simply a difficult subject to gather primary data on. One issue here is that many Western states have been fortunate enough not to suffer from a large number of terrorist attacks. This also means that terrorism is often a relatively 'small-N' problem, which increases data collection challenges. Then there are practicalities like a) finding (former) terrorists who could, for instance, be interviewed, b) convincing them to cooperate and c) making the time necessary to take these steps within research projects that are often conservatively budgeted in time and resources. Moreover, while it's been argued that the potential dangers of engaging directly with (former) terrorists can often be managed, personal safety is still an issue that must be considered, as are the ethical guidelines for working with what are often vulnerable individuals and groups.

Concerns over the field's apparent inability to gather more primary data, despite the problem having been known for decades, arguably peaked in 2014 when Marc Sageman's provocatively argued that research on terrorism had 'stagnated'. At first glance, such a damning statement from one of the most influential terrorism researchers in the post-9/11 period supports continued pessimism about the field. Yet, there was also considerable push-back; other leading scholars were quick to qualify Sageman's piece as overly critical, and pointed to areas of real progress in the study of terrorism. Since then, both established and up-and-coming researchers have begun proposing that, actually, the primary data issue may no longer be quite so problematic as it was. The question, however, is what the data says on the issue. Is ongoing pessimism warranted or has there been a change for the better? With the last in-depth assessment of the field dating back to Silke's groundbreaking work in the early to mid 2000s, a fresh look was needed.