The IRS data is reported at the Household level and at an Individual level.
Household Data: The household data is collected by interviewing the householder. Information in the household section is focused on all household details from the household composition, durables owned, household items purchased and other key demographic variables. A Normal Household is defined as a group of people who normally live together and take their meals from a common kitchen unless the nature of their work prevent any of them from doing so. Persons in a household may be related or unrelated to each other. There may be one member households, two member households or multi-member households. The link in understanding whether it is a household or not, is a common kitchen.
The Householder is defined as: A person who takes the decision on purchase of day to day household products such as groceries, toothpastes, soaps, detergents etc. with respect to what to purchase, when to purchase and how much to purchase is a Householder. This person has to be staying in the household and can be a male or a female.
- The householder need not always decide on the brands.
- The householder need not always physically go to the shop to buy these products; the householder may only be suggesting the requirements while someone else implements it.
Individual Data: Individual data is collected from a systematically randomly selected person who is 12 years or older and stays in the household. The Individual Questionnaire is mainly focused on capturing readership of publications, TV viewing, Radio listening, Mobile usage, Internet usage, Cinema going habits and personal usage of selected products.
Data Capture: We have introduced DS-CAPI (Dual-Screen Computer Assisted Personal Interview) as a new way of capturing data for the IRS. The DS-CAPI method eliminates the need for printed masthead booklets and instead uses a second device (dual screen) to display mastheads and other stimulus to the respondent. The DS-CAPI data capture methodology has these benefits:
- Simpler for the interviewer to navigate through the questionnaire.
- Simpler for the interviewer to manage the large amount of stimuli material better – as all the stimuli automatically appear on a separate screen, in front of the respondent.
- Simpler and more comfortable for the female respondent, as the method allows her to sit at a comfortable distance from the interviewer, hitherto not possible in the Pen-and-Paper or CAPI methods.
- A short interview length - both Household and Individual interviews not more than 30 minutes in order to ensure good quality data.
Coverage and Reporting: The sample size covered in IRS 2013 is approximately 235,000 households and is reported at these breaks:
- All India
- Individual States
- Socio-Cultural Regions (SCRs)
- 92 Independent & 99 Clustered Districts
- Population Strata, in Urban & Rural respectively at an All India level
- All towns with a population of 5 Lakh and above
- Delhi, Mumbai & Kolkata – reported in zones (4 zones in each city)
Data Fusion: To restrict respondent and interviewer fatigue, the Individual and Household interviews are designed to be completed in 30 minutes each. This is enabled by efficient data fusion – a globally used ascription methodology.
Quality Assurance: Quality of data collected is of utmost importance and is ensured by enforcing multiple checks:
- On-the-ground Quality Checks: A face-to-face follow-up interaction with the respondent is conducted to monitor and maintain the quality of data collected. All field managers are empowered with real time tools to monitor progress and quality.
- Telephonic Quality Checks: A dedicated team makes telephone calls to the interviewed respondents and checks for compliance. This is to assure that the interviewers correctly followed norms.
- Real Time Workforce Tracking: Through which interviewers are tracked while they are on field, assuring better control over data capture.
- Audio Recording Back-Checks:A dedicated team listens to audio recorded questions from the interview and ascertains if questions were rightly asked and responses correctly coded.
District Clustering: We have considered a Hybrid approach for district/s reporting, which is reporting some Districts independently, and others in clusters. The objective of this exercise was to cluster Districts in a meaningful manner from a Readership perspective. We believe that this innovative approach adds another dimension beyond clustering Districts by either population or literacy.
Methodology adopted for District Clustering:
- We examined a large number of variables from the Census 2011, including household variables like type of house, electricity, telephone connection, water source as well as population, literacy and urbanisation.
- We took circulation as the surrogate variable for readership and identified variables which had a strong correlation with circulation. The three variables that had the strongest correlation with circulation were population, literacy and urbanisation (% urban population of total population).
- As these variables had high correlation coefficients, we combined these variables through Principal Component Analysis (PCA) and constructed a linear combination of these three variables, which in a manner speaking the ‘importance’ of the District with respect to readership perspective.
- We ensured that the clustered districts are geographically contiguous and belonged to the same SCR.
- The distribution of number of districts independently reported and reported at a cluster level are – 92 Individual districts and 99 Clustered Districts.