Table 2. Descriptive statistics of the variables used in the regression models (N = 47).
| Mean ROR (%) conditioned by variable value | ||||
|---|---|---|---|---|
| Mean | Std. Dev. | Yes = 1 | No = 0 | |
| Ex-ante estimate | 0.28 | 0.45 | 87 | 85 |
| Analysis characteristics | ||||
| Self-evaluation | 0.28 | 0.45 | 131 | 69 |
| International institution affiliation | 0.26 | 0.44 | 53 | 97 |
| International and academic institution | 0.09 | 0.28 | 11 | 93 |
| Study published | 0.68 | 0.47 | 95 | 66 |
| Research Characteristics | ||||
| ROR reporting period^ | 0.41 | 0.50 | 99 | 65 |
| Multinational scope | 0.04 | 0.20 | 10 | 89 |
| National scope | 0.74 | 0.44 | 70 | 131 |
| Sub-Saharan Africa region | 0.38 | 0.49 | 40 | 114 |
| United States region | 0.51 | 0.51 | 130 | 39 |
| Sorghum only | 0.79 | 0.41 | 102 | 26 |
| INTSORMIL or ICRISAT | 0.49 | 0.51 | 87 | 85 |
| Evaluation characteristics± | ||||
| Pivotal supply shift | 0.17 | 0.38 | 28 | 98 |
| Parallel supply shift | 0.23 | 0.43 | 59 | 94 |
^ the median ROR reporting period is 1995. The reporting period takes the value of 1 if the ROR report is published after the year 1995, and 0 otherwise.
± The ‘pivotal supply shift’ assumes a linear in logarithms supply function and shifts proportionally, whereas the “parallel supply shift’ assumes linear supply function and shifts in parallel.