social currency that nourishes a high-quality leader-member exchange; that transformational leadership enhances followers’ receptivity for role expansion and extra-role actions via processes of social and/or personal identification; and that leader-member exchange gives leadership a more personal meaning (Wang, Law, Hackett, Wang, & Chen, 2005). In a case-based analysis, Roussin (2008) found that dyadic leadership discovery – “leader joining in honest, revealing, and (potentially) trust- building conversations with individual team members” – was more effective in increasing team psychological safety and team performance (Roussin, 2008). In a study aimed at understanding whether LMX mediates the relationship between Research&Development team leadership and employees’ organisational commitment, Lee (2005) found that transactional leadership had no associations or negative associations with different dimensions of LMX (Lee, 2005). A subsequent study of the author established that transactional leadership is negatively influencing innovativeness due to the leader-member exchanges that it encourages (Lee, 2008).
H3: LMX partially mediates the relationship between transformational, respectively transactional, leadership on the one hand and psychological safety on the other hand.

Conceptual framework

Strategy
Study Design
The research is a cross-sectional quantitative study. Data used for this study are part of an extended database; only part of the collected data is of interest for the present study. The study population is represented by employees of an academic hospital in the Netherlands. The respondents are leaders, respectively members, in teams at operational, tactical or strategical level.
The organisation and respondents
Data were collected from employees of a Dutch academic hospital having more than 1,000 beds and 10,000 employees. Respondents were either a team leader or member in a team whose leader was enrolled in the study. Data collected from leaders and team members at operational level are of interest for this study.
Data collection methods
Data were collected through a 111 item questionnaire which included both validated and non-validated scales. The questionnaire collected data on leadership styles, Leader-Member exchange, psychological safety, individual well-being, team reflexivity, team effectiveness, and respondents’ background. Of interest for this study are the questionnaire items on transactional leadership, transformational leadership, Leader-Member exchange, psychological safety, and respondents’ background. Except for respondents’ background, the answers were collected by the use of items extracted from validated scales; these answers were collected in form of 1 to 7 Likert scales (1 = “strongly agree”, 7 = “strongly disagree”). Data on respondents’ background were obtained using close-ended question or open-ended question framed to obtain answers represented by discrete values or 1 word.
Data on transactional leadership was collected using 5 items of Podsakoff, et al.’s (1990) scale. The items targeted behaviours such as providing feedback, acknowledging or not performance, and praising work that is better than the average (Podsakoff, MacKenzie, Moorman, & Fetter, 1990). For data regarding transformational leadership, 12 items of the same scale were used. The items addressed behaviours like role modelling, encouraging teamwork, and openness to new opportunities (Podsakoff, MacKenzie, Moorman, & Fetter, 1990). Information on LMX were collected by using 7 items of Graen & Uhl-Bien’s (1995) instrument. The items were aimed at dimensions of leader’s relationship with members (Graen & Uhl-Bien, 1995). Data on psychological safety were collected using items from Edmonson’s (1999) instrument; items targeted behaviours related to team climate (Edmonson A., 1999). Respondents’ background consisted in items on gender, age, education, leadership position, number of team members (if applica