This chapter summarises research on leading questions, showing that they undermine witness accuracy and pose a high risk for the quality of evidence of vulnerable witnesses. It then considers a series of Court of Appeal decisions which together signal the need for a changed approach from advocates and further empower judges to control cross-examination of vulnerable witnesses. These cases have greatly strengthened the ability of intermediaries to challenge conventional cross-examination strategies.

The chapter also reviews the provision of relevant training for the judiciary and advocates and regulation of the advocacy professions. The Judicial College has wholeheartedly incorporated the Court of Appeal’s approach; the professions have yet to adopt an equivalent advocacy training requirement and regulation of advocacy standards has been strongly resisted. The response of those responsible for advocacy training has been slow at best, resistant at worst and, until recently, has lacked leadership. By default, intermediaries and judges (themselves still absorbing the new judicial training) find themselves playing a case by case educational role with advocates.

Cross-examination and research

The objective of cross-examination in our adversarial system is to persuade the opposition’s witnesses to change their version of events or to discredit their evidence (for a helpful discussion, see Hanna et al, 2010). Accordingly, questions that lead the witness by suggesting the answer and inviting the witness to agree are the preferred form. Advocacy training encourages cross-examiners to control responses and substitute their case for the witness’s damaging testimony (for example, Hobbs, 2002, 2003; Hampel et al, 2008).

Several decades of research across common law jurisdictions shows that leading questions, along with other cross-examination techniques such as complex vocabulary and multi-part questions asked out of chronological order, mislead and confuse children and vulnerable adult witnesses (for example, Ellison, 2001; Hanna et al, 2010).