Martine Kruijtbosch

102 Chap t e r 3. 2 indicate relevant increments beyond five components. Therefore, the PCA-varimax rotation was performed with three, four and five components. The three components explained 29% of the variance in the data and had eigenvalues larger than two. When four components were used for the rotation, the explained variance increased by 5%, and when five components were used, the explained variance increased by an additional 4%. However, interpreting the considerations that correlated when using four or five components did not provide new moral reasoning perspectives. Therefore, we set the number of components to three. Table 4 presents the correlation loadings of the three PCA components for each scenario’s considerations and indicates that the majority of the considerations with eligible correlations represent the same MRPs, either a BO-MRP, a RR-MRP or a PE-MRP. Only one eligible consideration (O2) loaded on more than one component (i.e. loaded on two MRPs), and only two eligible considerations (C5 and O13) correlated with considerations of another MRP. MRP rating and ranking score The MRP rating score percentages in Table 5 suggest that all three moral reasoning perspectives play a role in pharmacists’ reasoning when handling the three drug shortages. The BO-MRP rating score percentage is the lowest in each drug shortage scenario (28.9%, 23.2% and 24.4% for the Contraceptive, Parkinson’s the Osteoporosis scenarios, respectively), and the PE-MRP rating score percentage is the highest (39.0%, 44.4% and 45.3%, respectively). The MRP ranking score percentages demonstrate that the PE-MRP is even more dominant in all three scenarios (60.2%, 72.1% and 68.6%, for the Contraceptive, Parkinson’s and Osteoporosis scenario, respectively). The difference between the BO-MRP and the PE-MRP is larger in the Parkinson’s scenario than the other two scenarios. The same trend was found when the MRP rating and ranking scores were corrected for the considerations that matched the criteria of eligibility and when the considerations that loaded with another MRP (O2 and C5) were given the scores for that MRP (see the Methods section for the criteria of eligibility and Table 4 for the eligible considerations used for this correction. See Appendix, Table 8 for the scores based on these corrections.)