Learning and Mislearning 75be made. The measurement of water flows, of input supplies and applications, of
areas under HYVs and transplanted, and of crop yields through crop-cutting are
all subject to methodological problems, and often wide margins of error. Although
there is no reason to suppose this research was exempt from these difficulties, and
although the WAPCOS report makes no mention of them, the critique which fol-
lows assumes complete accuracy of measurement and reporting. Full acknowledge-
ment must also be given to the difficulties recognized by the researchers. The 1979
monsoon was poor and less water than normal was available, especially on the
HBP. Low rainfall affected crop yield directly through water scarcity, and indirectly
through increased pests and diseases. Cultivators were free to use whatever inputs
they wished, weakening some aspects of the comparison of trial chaks with control
chaks. The Government of Madhya Pradesh could only construct channels down
to subchaks for three outlets – two on MRP and one on HBP. Finally, of these
three, the one on HBP received very little water, and one of the two on MRP had
a severe attack of gall midge.
In its conclusion, the WAPCOS (1980) report stated that it had briefly
reviewed ‘Exhaustive, most valuable and reliable data as collected’, and then sum-
marized its findings. Some findings were also presented in a subsequent paper
(Chadha, 1980). Those especially relevant to the policy decision for smaller
subchaks were:
1 acceptance: subdividing of a large outlet chak into smaller subchaks as irriga-
tion units found general acceptance from the cultivators (WAPCOS, 1980,
p30);
2 time taken to irrigate: ‘the time taken for irrigating the entire command of an
outlet was 5 to 14 days as against 20 to 45 days for normal outlets without
subchaks’ (Chadha, 1980, p388. Chadha, 1981, p70 specifies this as the time
for providing first irrigation);
3 uniformity of yield: the variation in yield within an outlet command was less
when it was divided into smaller units (WAPCOS, 1980, p30);
4 yield: with smaller subchaks yields were 70 to 137 per cent higher compared to
the normal outlets (Chadha, 1980, p388).
The syntax used in both sources implied that the benefits were the result of the
experimental treatment of subdivision and rotation.
The procedures and analysis throw doubt on the validity of these conclusions.
To understand why, we must look at the detail of the procedures and analysis
employed. Three methodological defects can be noted.
First, two of the three trials were eliminated from most of the comparisons
because they performed badly, but none of the 11 controls was eliminated. In
mitigation, none of the controls was reported to have suffered as severely as these
two trials from their respective problems – a gall-midge attack, and receipt of very
little water respectively. All the same, no analysis was made of the reasons for the
gall-midge attack or the little water received or whether these were related to the