Letters 30 August 2021

Jesim Pais and Vikas Rawal have indeed done a commendable job in assessing the various phases of the Consumer Pyramids Household Survey (CMIE’s Consumer Pyramids Household Surveys: An Assessment, 15 August 2021).

However, they made a glaring mistake in reporting by reshuffling the composition of the total sample size in rural and urban sectors. While in reality there were 1,10,975 households in the urban sector and 63,430 in rural sector (see Table 6, Wave Jan-Apr 2020, Survey Design and Sample, Vyas, March 12, 2020), the figures reported by the authors are about 1,11,000 in rural and 63,400 in urban as in “Section1. CPHS”. The authors have thus missed the fact that the design has a bias towards the urban sector and hence to the better-off. In fact, besides leaving out the nomadic population, there is a definite under-representation of the rural sector in the CPHS.

Second, in bringing out the confusion between houses and households, an assumption is made that a house has only one household which may not be true as a multistoried house in an enumeration block has many households. Then what was the instruction to the investigator? Did they enumerate all the households or one was picked at random? If the latter, they should have noted down the number of flats for using subsequently in tabulation for estimates. No clear indication is given by the CPHS documents about such cases.

Salil Sanyal is a retired statistician and a former UNDP consultant.