Table 2. Policy-mixes defined by principal component analysis followed by a partition-clustering analysis.
The three retained factors after rotation account for 66% of total variance. Scores in bold indicate PCA loadings higher than 0.3 or grouping frequencies bigger than 50%.
| Factor analysis | Cluster analysis | ||||||
|---|---|---|---|---|---|---|---|
| Variables | F1ProArbol | F2NPA&PSA-H | F3PROGAN | C1ProArbol | C2 PA&PSA-H | C3PROGAN | C4Agr. only* |
| PROGAN | -0.098 | 0.059 | 0.855 | 49% | 72% | 84% | 50% |
| PROCOREF | 0.272 | -0.067 | 0.347 | 71% | 59% | 79% | 0% |
| PRODEFOR | 0.434 | -0.219 | 0.041 | 72% | 28% | 27% | 0% |
| PSA-CABSA | 0.460 | -0.026 | -0.077 | 59% | 26% | 6% | 0% |
| PSA-H | 0.327 | 0.414 | -0.259 | 67% | 59% | 13% | 0% |
| NPA | -0.126 | 0.810 | 0.101 | 0% | 100% | 0% | 0% |
| % communities in categories | 28% | 12% | 33% | 27% | |||
* Ag. only: only agriculture and cattle programs (PROCAMPO and PROGAN).