What problems were addressed? The multiple mini-interview (MMI) is a multi-station interview consisting of eight to 12 stations, each of which lasts 8-10 minutes. A standardised scenario at each station evaluates a single facet of competence or a personal trait through the interviewer-interviewee interaction. It provides a valid, reliable and defensible approach to medical student selection. However, the MMI is insufficient for evaluating interpersonal skills, which are crucial for the medical practitioner.What was tried? This study added a group interview to seven MMI stations in medical student admission interviews at the National Cheng Kung University (NCKU), Taiwan, in order to strengthen the evaluation of applicants' interpersonal skills, and to construct a more holistic and valid medical student admission interview.In 2011, NCKU interviewed 122 people who applied for admission to its medical programme. A total of 33 interviewers participated in the interviews. They all received at least 7 hours of training that aimed to establish an interview blueprint and scoring criteria, which would increase inter-rater consistency.Based on the mission and philosophy of the Department of Medicine at NCKU, the seven stations evaluated the following competencies and personal traits: empathy; respect for life; crisis management; initiative; insightfulness; integrity, and communication skills. At the group interview, a 5-minute video was played, after which four standardised questions were used to stimulate group discussion. Each group consisted of six or seven interviewees and three interviewers who monitored and evaluated the interviewee peer discussion.Based on statistical analyses, the correlations between each of the seven MMI stations and the group interview were all positive, ranging from 0.15 to 0.42. The correlation coefficients were all significant (all p < 0.05), the only exception being that for station 5 (insightfulness) (p = 0.10). Cronbach's alpha for the seven MMI stations was 0.54. When data for the group interview were added to the analysis, Cronbach's alpha for the MMI combined with the group interview rose to 0.63.What lessons were learned? The combination of the MMI with the group interview improved the validity of the medical student selection interview. In test item selection, the primary concern is to design and choose items that increase internal consistency first, and convergent and discriminatory validity secondly. 1 The interviewee-interviewee peer discussion in the group interview is quite different from the interviewer-interviewee interaction in MMI stations. The positive correlations between the MMI stations and the group interview, then, can be taken as a good indication of convergent validity. It means the MMI stations and the group interview measured similar competencies in accordance with the mission and philosophy of the Department of Medicine at NCKU.Furthermore, the inclusion of the group interview with the MMI raised Cronbach's alpha from 0.54 to 0.63. This proves that the combining of ...