94 Chapter 6 Figure 6.2: Research questionnaires and assignments timeline Statistical analysis SPSS version 28.0.0.0 was used to process and analyse the results. Missing data were imputed with the ‘average of the available items’: Within one participant, the scores of the questions that refer to the same motivation type as the missing value, were used to calculate an average. This average score replaced the missing value. We used linear mixed effects models with random intercept and fixed effects for group (uniprofessional versus interprofessional) and time, to assess the longitudinal change in autonomous motivation scores and controlled motivation scores within each group. Timepoints were included as factor to account for non-linear effects over time. To study autonomous and controlled motivation scores over time within the interprofessional group, we repeated the linear mixed model analyses with interprofessional participants as reference category. Interaction effects between group and time were used to assess whether autonomous motivation scores and controlled motivation scores trajectories differed over time between the uniprofessional and interprofessional participants. Simple linear regression analysis was performed to investigate the relationship between autonomous motivation at T3 (predictor) and the ICCAS post-score (dependent variable). A power analysis was conducted for both the variables motivation and skills to determine the sample size. With an anticipated mean of 3.5 for motivation for the UPE-group and 4.0 (SD±0.5) for the IPE group, alpha=0.05 and a power of 0.8, the calculated sample size was 16 participants per group. An anticipated mean of 4 for interprofessional collaboration skills for the UPE-group and 5 (SD±1) for the IPE group, alpha=0.05 and a power of 0.8, the calculated sample size also was 16 participants per group. For the increase score of the ICCAS we anticipated a mean of 0.5 for the UPEgroup and 1 (SD±0.2) for the IPE group, alpha=0.05 and a power of 0.8. The calculated sample size for this analysis was 3 participants per group.
RkJQdWJsaXNoZXIy MTk4NDMw