2) Analytic Strategy
To test the hypotheses suggested in the previous section, we establish an empirical model as an ordered logistic regression model. This empirical model is useful in investigating the statistical association between predictors and outcomes measured in ordinal variables. By utilizing this model, for example, we can predict whether a hospital with citizen engagement as its resources would be more likely to perform better. The dependent variables used are the overall quality of care and patient satisfaction. The independent or predictor variables are physical, financial, human, and social resources. We also include patient engagement as an essential predictor of hospital performance.
The main models of our empirical analysis have each of those two performance measures as dependent variables. First, we estimate the models using all samples. Second, we conduct sub-sample analyses to investigate whether the roles of each resource differ across groups with and without patient engagement.