Lee et al.¹⁾ are to be congratulated for reporting their study on the effects of videotaping students' interviews with patients on interview skills. They are also quite forthright about potential shortcomings in their research; however there is one potential shortcoming that they miss.

This is the question as to what exactly was the intervention that caused the improvement. Their intervention was complex and multifaceted: it consisted of conducting the first interview, watching the videos, reading interview skill evaluation criteria, small group discussion, and finally feedback. It is difficult to tell from the study which component of the intervention caused the improvement. Was it a single component or two components or three components of those described above? This is an important question because the intervention must have been expensive because of its complex and multifaceted nature. Rolling out this intervention to large numbers of students may not be feasible because of the expense associated with its widespread implementation. A follow-up study where different components of the intervention were compared may enable us to discover the effective elements within the "black-box" of the intervention.

As a final point the authors state that the small size of the study was a potential limitation, but here I would disagree. They state that "increasing sample size would make small differences significant" - but that exactly what the researchers should be avoiding. A large sample might show an improvement that was statistically significant, but not clinically or educationally significant. For this reason, smaller samples are often best in this form of research.

Yours Sincerely