Table 3

Outcome comparisons between smallholders from STB and non-STB villages

All sampled householdsaHouseholds from the pilot countybHouseholds from all non-STB villagesMean differences
STB villagesNon-STB villagesMean differencesc
(1)(2)(3)(4) = (2)–(3)(5)(6) = (2)–(5)
Household wheat production
# of wheat farmers195519420840217611955
# of wheat farmers with NUE data1,0541151112269391,054
1. Yield, kg−1 ha7,161 [937]7,412 (73)6,912 (119)500*** (137)7,133 (37)279*** (79)
3. Nitrates use efficiency (NUE), kg−1 kg27.95 [13.61]31.41 (1.07)31.53 (1.83)−0.13 (2.08)27.52 (0.63)3.88*** (1.21)
5. Phosphorus use efficiency (PUE), kg−1 kg52.71 [18.42]56.33 (2.81)49.97 (1.88)6.36* (3.32)52.27 (0.82)4.06 (2.84)
7. Potassium use efficiency (KUE), kg−1 kg53.24 [19.74]57.76 (3.06)49.95 (1.84)7.82** (3.50)52.69 (0.86)5.08 (3.07)
Household maize production
# of maize farmers195019420539917561950
# of maize farmers with NUE data1,0531161102249371,053
9. Yield, kg−1 ha8,036 [1,379]8,127 (102)7,617 (113)510*** (150)8,026 (47)101 (109)
11. Nitrates use efficiency (NUE), kg−1 kg55.04 [27.95]58.35 (3.93)53.70 (3.17)4.65 (4.96)54.63 (1.18)3.71 (3.97)
13. Phosphorus use efficiency (PUE), kg−1 kg69.18 [20.66]74.37 (2.13)66.49 (2.69)7.88** (3.38)68.55 (0.88)5.82** (2.24)
15. Potassium use efficiency (KUE), kg−1 kg69.67 [21.32]75.90 (2.33)67.17 (2.71)8.73** (3.51)68.92 (0.89)6.98*** (2.41)

Note(s): aIn column 1 we report the mean of each outcome variable and its standard deviation in brackets. In columns 2, 3 and 5 we report the mean value of each outcome variable and its estimated village-clustered standard errors; bGiven that all STB villages were located in one pilot county, we first conduct a descriptive comparison over each outcome within the pilot country between STB villages and non-STB villages. However, given the limited number of sampled villages within the pilot county, the remaining regression analysis will only be conducted with all sampled villages (column 5); cMean differences were calculated by running a series of t-tests, and the standard errors were calculated using a village-clustered robust standard errors for inference, ***p < 0.01, **p < 0.05, *p < 0.1

Source(s): Authors' survey in 2018

or Create an Account

Close subscription notice
Close access options