In my data envelopment analysis (DEA) model, I have 3 outputs, namely (i) No. of Sponsored Projects, (ii) No. of Consultancies and (iii) Total Revenue generated from Sponsored Projects and Consultancies. My DMUs are different universities. It is clear that the third output is the overlapping factor as the revenue is generated due to the sponsored projects and consultancies. If I choose only the third output, the problem is that I cannot take into account the social benefits from a university because it is possible that a university undertakes many projects and consultancies (i.e. indicators of social benefits) while generating less revenue. Again, if I take the first two outputs, then I miss the part of revenue generation because all the projects and consultancies may not generate same revenue. Should I consider the first two outputs or the third output or all the three outputs in my model?
I would use one or the other, and test for robustness of results when changing the third variable
It is generally recommended to include all three outputs in your DEA model, as each of them represents a distinct aspect of the performance of the DMUs (universities).
What I understand is that the first output, No. of Sponsored Projects, reflects the ability of the university to attract funding for research and development. The second output, No. of Consultancies, reflects the university's ability to provide services to external organizations. The third output, Total Revenue generated from Sponsored Projects and Consultancies, reflects the university's overall financial performance.
If you include all three outputs in your model, you can capture a university's social benefits and financial performance. However, you should be aware that the third output may be correlated with the first two outputs, as revenue is generated from sponsored projects and consultancies. This could potentially affect the results of your DEA model.
One way to account for this is to use a model that allows for overlapping outputs, such as the common weights model or the free disposal hull model. These models allow for the possibility that some inputs and outputs may be used in the production of multiple outputs, and can provide a more accurate assessment of the efficiency of the DMUs.
Ultimately, the decision of which outputs to include in your DEA model will depend on the specific research question you are trying to answer and the data that is available to you. It may be helpful to consider the strengths and limitations of each model and the implications of including or excluding certain outputs in your analysis.
Good luck.
- Badges
- Science topic