Martine Kruijtbosch

100 Chap t e r 3. 2 Data analysis SPSS 25 was used for all data analysis. Control considerations One of the 13 considerations per drug shortage scenario was a control consideration (i.e. a meaningless consideration, labelled ‘M’), which was formulated to correct for respondents providing potentially unreliable answers. 32 If respondents ranked such a control consideration more than one time, their surveys were excluded from analysis. Validation of moral reasoning perspectives First, a principal component analysis (PCA) was performed to check the rating scores for construct validity, searching for a confirmation of the three MRPs in three distinguishing components. The rating scores were thus checked for factorability with the Kaiser-Meyer- Olkin’s measure (KMO, ideal value < 0.6). Subsequently, correlations between variables were tested with Bartlett’s test of sphericity (index p < 0.05). Varimax rotation was used to extract the components to increase interpretability of the data. The components were further examined by their percentage of variance explained, their eigenvalues (eligible value > 1) and their component statement loadings (eligible value ≥ 0.35, or a difference of > 0.2 between the correlation values if a statement loaded highly on more than one component). 49,50 Measuring moral reasoning perspectives: the MRP rating and ranking score In order to measure the relative influence of the three moral reasoning perspectives (BO, RR and PE), rating and ranking scores were calculated for each MRP. MRP rating score: For this score the meaningless considerations were excluded, leaving 12 considerations per scenario, of which four represented the BO-MRP, four the RR-MRP and another four the PE-MRP. Each consideration in a scenario was given a numerical weight from four (the consideration played a very strong role) to zero (the consideration played no role). Hence, for each MRP in a scenario a respondent could have a maximum MRP rating score of (4 times 4) 16 points if all four considerations of one MRP (BO, RR or PE-MRP) played a very strong role in handling the drug shortage, and a minimum MRP rating score of 0 points if none of these four considerations played any role. For each respondent the BO-MRP, the RR-MRP and the PE-MRP rating score were expressed as a percentage of the maximum score. Finally, these three MRP score percentages were normalised to 100% to present their relative percentages. MRP ranking score: A numerical weight from four (first ranked) to one (fourth ranked) was given to each ranked consideration. 51 Ranked meaningless considerations were assigned