Hey!
I have three different data sets:
1.costs (reported by the municipality): DV=continous,
2. the satisfaction (how satisfied the population is with GPs in the municipality): DV=1-8 (likert scale),
3. the capacity (municipality level): DV=continous.
All three data sets have similar background variables of the municipalities; like the number of inhabitants (population), centrality index and what region the municipality is in. I do not have any municipality identifier (do not know what each municipality answered since anonymous data), and therefore I cannot merge the data sets. I will have to do three separate regressions for each of the data set.
I want to investigate what characterizes the different municipalities that have additional costs to the regular GP scheme. Is the municipalities that have the highest costs to the Regular GP scheme also the ones that are less satisfied, and the one with less capacity etc?
Do you have any recommendations or ideas on what regression to use and what statistical tests?
2 is most likely ordinal logistic regression. 1,3 maybe ols but really not enough DV info. to tell.
You might google the query, what type of regression do I use. Best, D. Booth
It is mentioned that you have three sets of data namely related to cost, satisfaction and capacity. Further, you have assumed that the municipalities whose data is with you have similar background information probably characterized by the number of inhabitants and Centrality.
From your description of the question, I understand that additional cost to GP is considered as the dependent variable (?) and Satisfaction and capacity are considered as independent variables.
If I am doing the analysis of the data given to you, based on your description of the problem, I will proceed like this.
1. Identify the municipalities with Low, Medium and High additional costs.Consider them as three different categories of Municipalities.
2. Compare the mean satisfaction score between them first by ANOVA and then by using Post-Hoc multiple comparisons, identify the differences among them.
3. If you have an index to measure the capacity say proportion of households with certain desired quality then compare them among the three categories again using ANOVA.
4. By the step 2 and 3, you get the answer whether the selected municipalities are differing by the groups or not ? If not differing then further analysis need not be carried out and the data can be pooled for further presentation, otherwise draw the conclusion.
5. I do not feel that regression analysis would be a better analysis unless you want to develop a model to predict the category of municipality based on satisfaction index and the capacity index.
6. While doing the ANOVA analysis, remember to note that the interaction being significant or not? In the presence of interaction being significant, you have to be careful in your interpretations.
Alva Nilsen - You can aggregate the datasets into one before testing your hypotheses. Here's an example.
Article Relating Patient Satisfaction to Nurses’ Job Satisfaction, J...
- Badges
- Science topic