Sample Solution

Even though psychological safety is a notion that lies mostly in the individual perception, it is best observed at team level (Edmonson, 2004). The study dataset contains data collected from team leaders and team members. Given the data clustering (as shown by table), multilevel regression analyses will be performed.
Level Operational (N)
Regression Respondent
2 team leader 90
1 healthcare professionals 514

Prior to performing the regression analyses, data will be tested for ensuring that the normal distribution condition is met. As well, correlation matrices will be used for deciding the control variables to be used.
The mediation will be tested using the causal steps method. As explained by MacKinnon et al. (2012), this method has four steps:
1) testing for significance the relationship between independent (i.e. transformational, respectively transactional leadership) and dependent variable (i.e. psychological safety);
2) testing for a significant relationship between the independent (i.e. transformational, respectively transactional leadership) and mediator (i.e. LMX);
3) testing for significance the relationship between mediator (i.e. LMX) and dependent variable (i.e. psychological safety) when controlling for the independent variable (i.e. transformational, respectively transactional, leadership);
4) assessing whether the relation between the independent variable (i.e. transformational, respectively transactional, leadership) and the dependent varible (i.e. psychological safety) is weaker when the mediator (i.e. LMX) is added (MacKinnon, Cheong, & Pirlott, 2012).
Validity and reliability
Validity consists in measurement assessing what is intended to be assessed for the aim of the research (Field, 2009). In order to achieve internal validity, data collection instrument used previously validated scales for transactional leadership, transformational leadership, Leader-Member exchange, psychological safety. Added to this, factor analysis will be performed. For ensuring external validity, the demographic variables of the study sample will be compared with the demographics of other Dutch hospitals.
Reliability represents data collection instrument’s capability of producing similar results under different contexts with similar characteristics (Field, 2009). Cronbach’s alpha will be computed for evaluating the reliability of the scales used in this research; a value of minimum .7 will be accepted.