Fig. 5

Results of correspondence analysis of hospitals that did not perform cardiac surgery. Figure 5 illustrates the results of correspondence analysis of hospitals that did not perform cardiac surgery. Measures that are close together indicate measures with similar ratings and measures that are close to ratings represent combinations that occur more frequently. η₁² and η₂² near the axes denote the eigenvalues of the first (horizontal) axis and the second (vertical) axis respectively. The total sum of the eigenvalues are 0.034 = 0.023 + 0.011. The contribution ratio of the first axis, γ₁ is 0.67 = 0.023/0.034 and of the second axis, γ₂ is 0.33 = 0.011/0.034, which means the first axis explains 67% of the variance in the data and the second 33%. In other words, the the first two axes account for 100% of the variance of the data. The circles correspond to the clusters in Fig. 3. Communication measures in the left circle were placed close to each other around “High” and far away from “Medium” and “Low,” indicating that scores were higher for communication measures than for the other seven measures in the right circle that are closer to “Medium” or “Low. Revised from Fig. 4 in Okuda M, Yasuda A, Tsumoto S. Factors of Patient Satisfaction based on distant analysis in HCAHPS Databases, IEEE International Workshop on Data Mining for Service (DMS2014), Shenzhen, China, December 14, 2014 [36]