Table 1.
Treatment effects for Ghana
| Ȳ pre | RMSPE pre |
RMSPE pre/Ȳ pre (%) |
RMSPE post |
RMSPE ratio |
LTE | Ranking one- sided | P-value (one- sided) | Ranking two-sided | P-value (two- sided) | |
|---|---|---|---|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) | |
| Child Mortality | 164·87 | 6·19 | 3·75 | 3·92 | 0·63 | –2·30 | 7 | 0·333 | 18 | 0·8571 |
| Diarrhea | 25·13 | 3·07 | 12·21 | 8·95 | 2·91 | –8·85 | 8 | 0·3809 | 14 | 0·6667 |
| Respiratory | 8·82 | 0·05 | 0·56 | 1·12 | 21·11 | –0·84 | 1 | 0·0476 | 1 | 0·0476 |
| Congenital | 4·00 | 0·21 | 0·05 | 0·33 | 15·46 | 0·22 | 20 | 0·9524 | 4 | 0·1904 |
| Asphyxia and Trauma at Birth | 11·95 | 0·25 | 0·02 | 1·06 | 4·12 | –0·95 | 6 | 0·2857 | 12 | 0·5714 |
Notes: The table shows different statistics for Ghana and its synthetic control group. In particular the first column shows the level before the implementation of the International Code of Marketing of Breast-Milk Substitutes. Column (2) shows the root mean square prediction error (RMSPE) between Ghana and its synthetic control group before the implementation. This measure shows a very good fit for all dependent variables in column (3), except for diarrhea. Next the (RMSPE) after the implementation and the RMSPE ratio (after/before) is calculated in columns (4) and (5). Column (6) shows the estimated treatment effect in levels (LTE). This measure is divided by the RMSPE pre (2) and all countries (treated and placebo countries) are ranked based on this ratio. The resulting ranking and P-values for one sided tests are shown in columns (7) and (8). Finally, column (9) shows the ranking of the RMSPE ratio (5) for the same set of countries and the resulting P-value for the two sided test (10).