There are no weights given for the 1993 round of the survey, which means the 1993 numbers in the charts below are not strictly comparable to those for the 2003 or 2013 rounds. Weights are needed to go from the sample to the population, so only the 2003 and 2013 numbers were population numbers. The 1993 AIDIS sample itself is much smaller than 2003 and 2013 samples, both of which have similarly sized samples. This means we have little confidence that NSSO surveyed a largely identical group of households between 1993 and the other two later rounds.
To access the slide deck click here
