The rapfish sensitivity analysis will produce policy strategies that can be carried out to improve the sustainability index. This output can be considered a policy recommendation in addition to describing the existing conditions of the sustainability index.
Multidimensional scaling sensitivity analysis is actually a simulation by giving a good value to each attribute (assuming other attributes are fixed). Changes in the sustainability index value are then recorded and compared with other attributes when the same thing is done.
Conducting this sensitivity analysis is actually similar to determining attribute leverage. However, attribute leverage works by removing one attribute at a time (assuming other attributes remain fixed), then calculating the change in the R squared (not the sustainability index value). However, based on experience, the results are usually not much different between sensitivity analysis and attribute leverage.
Rapfish sensitivity analysis step
Sensitivity analysis is conducted in each dimension. For example, the technology dimension has five attributes: seed technology, planting technology, fertilization technology, pest management technology, and post-harvest technology.
Take a look at the picture below:
The table above describes a technology dimension of the multidimensional scaling value. The initial MDS value is 58 (medium sustainability). To remind you, here are the categories of sustainability index values:
Very poor sustainability (0–20)
Less sustainability (20–40)
Medium sustainability (40–60)
Good sustainability (60–80)
Very good sustainability (80–100)
The first sensitivity analysis step is to change the value of seed technology (value 3) to its good value (value 5) while the value of other attributes remains, then re-run MDS-rapfish. In this example, the result is 63 (yellow shading).
The next step returns the value of seed technology to the actual number (value 3), then changes the value of planting technology (value 2) to its good value (value 5), which then obtains an MDS result of 60 (yellow shading). And so on until the last attribute is replaced with the value of good with the value of other attributes fixed (cateris paribus assumption).
Next, we look at the largest MDS value. That is fertilization technology. Fertilization technology has the highest priority for improvement because it has a large impact on the sustainability index. Next, HPT technology is the last priority.
We use this attribute prioritization to find scenarios for how the sustainability index value can be increased. If we replaced the attribute values with their good values one by one, we would now do the replacement in stages. The first priority attribute replacement results in an MDS value of 71. Next, we also change the second priority attribute’s value without returning the first priority attribute’s value.
Take a look at the image below:
Improving the fertilization technology attribute is enough to give the MDS a good score (71). Meanwhile, to raise the sustainability index value to the excellent category, it requires a scenario of at least changing three attributes: fertilization technology, post-harvest technology, and seed technology. The three attribute changes will result in an MDS value of 83.
So the policy recommendation that can be offered is how to change these attributes from the initial value to the good category of each attribute.