����������

A recent study by the ���������� has found that focusing on audio alone improves the performance of the interviewer during interviews, particularly in criminal investigations.

The research, published in The European Journal of Psychology Applied to Legal Context, examined how the interviewer’s cognitive load - how hard the brain is working - influences the performance of the interviewer and the outcome of the interview.

The study emerges from a well-established body of research on cognitive load theory, which suggests that human working memory has limited capacity for processing information simultaneously. While previous research has extensively examined how cognitive load affects interviewees during interrogations, this study examines how different forms of presentation affect the interviewer’s cognitive resources and subsequent performance.

Dora Giorgianni, lead author and PhD student from the ����������’s International Centre for Research in Forensic Psychology, said: “Our research turns a common belief on its head: sometimes, seeing less means knowing more. By focusing on what people say instead of what they look like, interviewers can cut through distractions and see the real story.”

This study, involving 120 participants, investigated how listening only versus watching and listening to an interview affects the interviewer’s mental effort and results.

As part of the study, participants watched a video with audio of a mock suspect, while others only listened to the audio. The suspects were either truthful or lying, and the interviewers had to remember what the suspect said, come up with follow-up questions, and decide whether the suspect was being honest or deceptive.

Researchers found that interviewers who relied on audio only (low cognitive load) asked more total and higher-quality questions and demonstrated greater accuracy in detecting truth or deception than participants who could both hear and see the suspect (high cognitive load). 

Overall accuracy in distinguishing both truths and lies for the low effort group was almost double (61.7 per cent) than for the high effort group (35 per cent). Access to visual cues made interviewers more likely to suspect deception and impaired their ability to assess veracity.  

Further analysis revealed that accuracy in the high cognitive load condition was positively associated with attention to a verbal cue (specifically, the quantity of details in accounts) and a vocal cue (specifically, speech disturbances) rather than visual behavioural indicators. This pattern suggests that even when visual information is available, focusing on auditory elements produces better results.

This indicates that using both audio and visual information at once can overload the interviewer and make it harder to perform well in an interview, for instance to judge whether someone is truthful.

“By focusing exclusively on what suspects are saying rather than how they look, interviewers can dramatically improve their effectiveness both in terms of information gathering and credibility assessment”, added Dora Giorgianni.

The study indicates that focusing on audio-only interviews can help interviewers detect lies from criminals or suspects, enhancing interview effectiveness and aiming to increase the chances of correct convictions.

“This has important implications for training programmes, which often emphasise reading body language and facial expressions,” added Dora Giorgianni. “Our research suggests that blocking out visual distractions and concentrating on verbal content alone may be the key to better interviews and more accurate lie detection.”

Since lie detection cannot be left to technology for ethical reasons, Dora Giorgianni says interviewers must improve their techniques by keeping interviews clear, asking better questions, and focusing on what is said rather than on misleading visual cues.

“The human brain has physical limits that artificial intelligence does not share, but understanding these limitations allows us to develop strategies for improving human performance - especially in situations where human lives are at stake and the lack of transparency in AI systems makes it unethical to use them."