Law Enforcement Agencies gather intelligence in order to prevent criminal activity and pursue criminals. In the context of Human Intelligence (HUMINT) collection, Covert Human Intelligence Sources (CHIS) provide unique access to criminals and organised crime groups, and their collection of intelligence is vital to understanding England and Wales’ threat picture. However, the elicitation of detailed, accurate, reliable and timely intelligence relies heavily upon the deployment of evidence-based interviewing processes. Therefore, to develop an evidence-base for source handler and CHIS interactions, the present Thesis undertook 5 studies. Study 1 consisted of structured interviews with police source handlers. Rapport was perceived as essential for intelligence elicitation, supported by a range of rapport strategies, with the majority of participants believing that rapport could be trained to some degree. Study 2 comprised source handlers’ perceptions of the interviewing processes employed with informants. Five themes emerged from the interviews, (i) a comparison between interviewing and debriefing; (ii) the PEACE model in intelligence interviews; (iii) the importance of effective communication; (iv) Source Handlers’ use of cognitive retrieval techniques; and, (v) Source Handler interview training. Study 3 examined the impact of a context tasking instruction on intentional memory with mock informants across three conditions: (i) incidental encoding, (ii) intentional encoding or (iii) intentional encoding with tasking instruction, performing a free recall and prompted recall. Results showed that the intentional encoding with tasking instruction condition reported more correct information during the free recall phase compared to those in the incidental encoding condition. A significant increase in incorrect information was reported with the tasking instruction, but at no cost to the overall percentage accuracy. The free recall phase resulted in more accurate recall than the prompts phase. Study 4 gained unprecedented access to real-life audio recorded telephone interactions between police source handlers and CHIS, exploring the impact of rapport on intelligence yield. Overall rapport, attention and coordination significantly correlated with intelligence yield, while positivity did not. Attention was the most frequently used component of rapport, followed by positivity, and then coordination. Study 5 explored the impact of question types on intelligence yield used by source handlers during telephone interactions with CHIS. Source handlers were found to utilise vastly more appropriate questions than inappropriate questions, though they rarely used open-ended questions. Across the total interactions, appropriate questions were associated with gathering the majority of the total intelligence yield. Taken together, an evidenced-based approach shall advance source handler and CHIS intelligence interactions, and information gathering approaches more broadly.