86 Chapter 5 The significance tests of the differences showed that for Category L, the difference between TECHL and EMOL was significantly larger than expected (S = 20, E = 6.9, p = 0.001), rating fewer TECH cases as complex. There was a considerable difference in case count between TECHL and SDML; however, this difference did not reach significance (S = 10, E = 4,5, p = 0.157). Bonferroni correction yielded the same results. The other two differences in Category L and all three differences in Category H were insignificant. More results are shown in Table 3. We determined that the three task types (SDM, EMO, TECH) have different L ratings because the TECH and EMO difference is significant within Category L. TECH cases were perceived as less complex than EMO cases. However, we could not establish that this is also true within Category H. SDM did not differ significantly from EMO or TECH cases in all comparisons. Table 1: Case vignette type distributions per preference category. Case vignette type per category Observed sum (S)1) Expected sum (E)2) Difference (S-E) Cumulative probability3) Point probability4) Two-sided p-value5) Bonferronicorrected p-value6) Job satisfaction (N=23 participants) SDML 35 20.7 14.3 1.000 <0.001 <0.001 <0.001 EMOL 31 20.7 10.3 0.999 0.001 0.004 0.024 TECHL 3 27.6 -24.6 <0.001 <0.001 <0.001 <0.001 SDMH 4 20.7 -16.7 <0.001 <0.001 <0.001 <0.001 EMOH 13 20.7 -7.7 0.014 0.008 0.028 0.169 TECHH 52 27.6 24.4 1.000 <0.001 <0.001 <0.001 Complexity (N=15 participants) SDML 15 13.5 1.5 0.772 0.124 0.704 1.000 EMOL 5 13.5 -8.5 0.001 0.001 0.002 0.011 TECHL 25 18 7 0.995 0.008 0.025 0.152 SDMH 17 13.5 3.5 0.929 0.063 0.268 1.000 EMOH 16 13.5 2.5 0.866 0.094 0.456 1.000 TECHH 12 18 -6 0.027 0.016 0.054 0.326 SDM is shared decision making, EMO is emotional, and TECH is technical. L or H in the subscript refers to the low- or high-ranking category, respectively. All significance tests are performed according to the algorithm based on randomization tests, as described under Statistical analyses and in the Appendix. 1) number of times that the task (SDM, EMO, or TECH) is assigned to the position (L or H, as indicated in subscript), added across all persons in the samples. 2) the expected value of ‘sum’ if all persons assign the tasks randomly to positions. 3) probability that the sum is less than or equal to the observed value if all persons assign the tasks randomly to positions. 4) probability that the sum is equal to the observed value if all persons assign the tasks randomly to positions. 5) computed from the previous two probabilities. 6) 6 times the original two-sided probability, as six comparisons are made; p < 0.05 is here considered statistically significant.
RkJQdWJsaXNoZXIy MTk4NDMw