A Note on Gender and Age Indicators in CPHS
by Gayatri Dewan
The latest note in the CPHS documentation series was released on the 18th of September, 2021. Authored by Mahesh Vyas, the note details how CPHS captures gender and age information of individual members of sample households, methods used for checking and dealing with inconsistencies in responses, and an analysis of trends in the age distribution and gender ratio of the Indian population. The note also employs a critical lens to discuss present limitations in the collection of this data, and inferences arising from the same.
CPHS captures gender at the level of the male/female binary. Responses regarding gender are taken at face value, without judgement. Since CPHS is a panel survey of multiple waves, it is easy to spot gender change in the database. These could be either the correction of a past entry or genuine gender change. The survey methodology involves checks to reduce errors in capturing gender, which are described in the note.
Using the CPHS data on gender, the note provides estimates of the gender ratio. An advantage of CPHS’s fast frequency data over the Census is that it allows the study of inter and intra-year movements in the ratio between two Censuses.
The CPHS data displays a significantly worse gender ratio compared to the 2011 Census, indicating a considerable deterioration in the ratio in the 2010s. In addition to analysing the national averages of the ratio across 2014-20, the note also presents a summary of urban-rural and state-wise gender ratios.
Between 2014-2020, the rural and urban gender ratios experienced both ups and downs, and eventually they converged. The rural ratio improved mildly, whereas the urban ratio fell tremendously.
With respect to state-wise ratios, the Census and CPHS data tell a similar story. The data show that there has been a conspicuous worsening of the ratio in the northern belt containing Punjab, Haryana, Rajasthan, Madhya Pradesh, Jharkhand, and Bihar, which are also India’s less developed states. Conversely, Kerala, Puducherry, Goa, and Tamil Nadu in the South have had better outcomes.
The second section of the note details the manner in which Age is recorded in CPHS. Many people in India do not know their exact age, and often popular festivals are used as markers of an individual’s year of birth, which can lead to some inaccuracies in reporting. Despite these limitations, the age of individuals in CPHS is expected to be correct within the bounds of a few years, given certain checks present in the survey mechanism which are specified in the note.
Within the age section, the note contains an elaborate discussion of the changing distribution of the Indian population across various age groups. Starting with the under-5 age category, there appears to be a notable drop in 2014, compared to the 2011 Census, followed by a progressively declining proportion across the years. The note mentions possible survey-based reasons for the decline, and efforts to correct the same. CPHS suggests that birth rates in India might have declined significantly during the 2010s.
Conversely, there has been an increase in the proportion of the 15-24 category, leading to a shift in concentration of the population pyramid from the base (younger age groups) to the working age group. On the other hand, the proportion of 25-34 year olds has been falling since 2014. This could be due to migration out of the household for work, or young couples moving out to start a new household, which leads to them dropping out of the sample. A similar, detailed analysis regarding the trends in the proportion of middle-aged and senior citizens can be found in the note.
The last section of the note, titled Age-Gender Pyramid, contains an overview of changes in India’s population pyramid over the last decade, as well as certain reasons why the CPHS population pyramid may be limited in reflecting the demographic transition of the Indian population due to its panel nature.
Lastly, the note ends with a discussion on the gender ratios across various age groups, and their transition between 2014-19. Interestingly, the gender ratio appears to peak in the 30-34 age category. Yet, it declines again for ages 40s through early 60. While these trends are found in the Census data as well, there is a divergence in the seniors’ category, which may point to women in the senior age group missing from the CPHS count.
This note acts as an introductory resource for all those interested in India’s changing demography. It would be a useful brief to those who would dive deeper into the vast CPHS database with its 23 waves of data since 2014. The note shines a light on what is possible and the areas where researchers need to be careful in deploying the data for demographic analysis. The note points out that CPHS is not a database on demographics. But, demographic data is an automatic outcome of a large all-India household panel survey. Demographic data, of course, is also critically important to all other studies of Indian households. This note, therefore, is relevant to all users of CPHS.