CODE FOR SELF REGULATION FOR USE OF THE IRS DATA

A     A. Objective

The IRS is a collaborative study serving the research needs of the members of the MRUC. As such, it is desirable that the IRS data is used responsibly by all the member-subscribers and hence the need for a self-regulating code for use of the IRS data. The Code is to be read along with the IRS End Users License Agreement released with IRS 2019Q4 which is binding on all Subscribers. The Code for self-regulation is to provide for

a)      Common understanding of research terms, data definition and data analysis norms as provided by MRUC, including the statistical rules governing the IRS (Appendix I);

b)      Adherence to data analysis and reporting norms for truthful representation of data to clients, readers of the publication and the public at large (Appendix I);

c)      Redressal mechanism for complaints received in respect of data misrepresentation, abusing/disregarding set statistical norms and/or any other data usage provided by this Code.

d)      Any claim or comparison for publicity/promotion, hoardings or communication in any other form which is in violation of the protocols laid down in para B below will be subject to adjudication as per the redressal mechanism.

Disclaimer

The IRS 2019 Q4 data is licensed to the IRS Subscriber for a period of two years and as such the IRS data can only be used by registered subscribers of the data. MRUC retains the right to initiate legal action against any access or unauthorized use of the data by non-subscribers and any unauthorized sharing of the data by any subscriber. 

      B. Code of Self-Regulation

            a) Comparisons with previous rounds

IRS 2019 Q4 data comprises 4 quarters data of 2019 (Q1 '19 + Q2’19 + Q3’19 + Q4’19).

Note that this round release contains the more recent data collected and across quarters is essentially from 2019. Unlike earlier rounds there is no break in the data collection. Please consider the field work duration of previous report/s while comparing the data.

b)     Use of IRS2019Q4 data in publicity and promotions

In all cases of publicity/promotion, hoardings or communication in any other form, where Readership / Listenership / Viewership data may be compared with two or more publications / radio stations / TV channels / any other media, users should abide by the following protocol:

(i)     Any claim for “leading”, “No.1” or to establish top position by any parameter/s should ONLY be based on a like to like comparison, i.e. the same set of readership / listenership / viewership data to be compared amongst publications / radio stations / TV channels / any other media, sourced only from IRS 2019Q4.

(ii)   Comparison should be restricted to publications of similar publishing frequencies, i.e., Daily newspapers to Daily newspapers, weekly newspapers to weekly newspapers and magazines to magazines.  

(iii) Comparison should be between similar readership frequencies, and restricted to any one of the pre-defined metrics – i.e. AIR to AIR, Upto 3 days to Upto 3 days, Upto 7 days to Upto 7 days and TR to TR.

(iv)  For any specific demographic, geographic, product profile category the comparison should be done with the same set of comparable readership / listenership / viewership numbers for all publications / radio stations / TV channels / any other media.

The applicable parameter/s along with the specific geographical area (City, State, Country) must be clearly mentioned (in the same font size as the headline) in the publicity material.

c)      Variant readership

(i)        The “main” edition and the “variant” edition should be treated as two different newspapers 

(ii)      Publishers should refrain from issuing misleading communication wherein readership numbers of Main+Variant are reported as a single readership number for that publication.

(iii)    Publishers can use the Main + Variant readership numbers together, provided they report the Main & Variant numbers separately and not only as one single consolidated number in their communication. For eg, if for a given publication, Main Issue Readership = 100 readers; Variant Issue Readership = 20 readers, then the publication can mention in their communication that their readership numbers are : Main Issue =100; Variant Issue =20; Main+Variant Total = 120.

(iv)     Media planners should verify readership being used for comparison between publications, i.e., Main or Variant, and ensure like-to-like comparisons are used for selection of publications

(v)       Readership numbers of Main issue of one publication should be compared only with the Main issue readership numbers for other publications.

(vi)     For any cost comparisons, it is essential to verify whether the right rates, from the Main issue or Variant issue rate cards, are being used for all selected publications to ensure no incorrect comparisons are made.

C.      C. Redressal mechanism

Any violation under this Code by any subscriber/member/user, will be adjudicated by a Disciplinary Committee (DisCom) constituted by the MRUC.

a)      Any member and/or subscriber affected by the release of any publicity material violating the code provided shall write to the MRUC DisCom within seven days of the date of publicity with the following:

(i)     Covering letter from the Owner/CEO of the Complainant giving

1)      Name and Address of the Complainant

2)      Name and Address of the Opposite party

3)      Details of the nature of violation along with corresponding claim of the Complainant

(ii)   Actual copy in the case of printed publicity (newspaper, magazine or flyer), colour photograph in case of any Outdoor medium and a recording in case of Audio-visual

(iii) A DD of Rs. 10,000/- issued in favour of MRUC

b)      The DisCom shall deliberate on the Complaint within seven days of receipt of the complaint and convey its decision to the Complainant and the Opposite party within seven days of its decision

c)      If the Opposite Party is found guilty of contravening theCode, DisComwillproposefollowing corrective action against the offender, which may include:

•      The said publication to print a corrigendum on the same page and size of the publicity in case of publicity/campaign in anewspaper and/or magazine (text of corrigendum as advised by MRUC) 

•      In case of any other form of publicity/campaign, then the corrigendum to be published on page 3 in size 20 x 3 cc as per draft of the corrigendum advised by MRUC

•      Corrigendum to be published within a period of 15 days from the date of decision ofthe DisCom

•      MRUCmay advise all affected competing publications to publish the same corrigendum on no cost basis.

•      MRUC may also inform all its members about the said violation.

d)      MRUC also retains the right to take suo moto cognisance of any violation of the guidelines under this code.

e)      The decision of the MRUC Disciplinary Committee will be final.

APPENDIX I

Guidelines for Responsible Use of IRS Data

1)      Statistical rules governing IRS 2019Q4

      a) IRS 2019Q4 fieldwork period: December 2018 to March 2020

     b) The estimations made and provided in the IRS are based on random sampling methodology and all such estimates operate within 20% (approximate) sampling / non-sampling error¹ level for any reporting breaks with 90% confidence level for 10% incidence.

         c) Estimates for a title are suppressed if the unweighted² all-adult readership for the title for the period of    the survey, in the respective reporting unit of IRS, falls below 30.

        d) For such publications/titles; a separate database has been is provided which has limited geographic variables along with print readership variables such as TR and AIR.

2)      Data classification

      a) Household vs. Individual data

i)        The IRS reports Household data and Individual data separately

ii)      Reading Household Consumer Goods and Durables

•      Penetrationof Durables ownership and purchase of consumer goods is captured at household level and should not be used for universe sizing in the Individual database.

•      Individual datashould be considered only for consumer profiling. It should be read as the targeted Individuals who own those durables or purchase the respective consumer goods, in their household.

      b) Readership data

i)        Readership metrics

Readership for Daily Newspapers is now reported by the following metrics-

·         AIR - Yesterday Readership

·         1 to 3 days

·         1 to 7 days

·         Total Readership – last 4 weeks

ii)      Variant Readership

·         IRS 2019Q4 captures readership for both - main paper as well as variants of the main paper for select markets.

·         Variants have been identified basis differences in mastheads / presence or absence of supplements and/or price point differences (as reported in ABC or by details as provided by respective publication houses)

Causes for Variations in Sample Surveys

All estimates based on a sample survey are subject to ‘sample variation’. However tightly controlled, the results from one sample of people will differ somewhat from another sample of people drawn in exactly the same way.

Any characteristic (e.g. % owning a cell phone, % reading a newspaper etc.) observed in any sample survey or sample surveys conducted at two different time periods, could show different results. These observed differences can be of two types:

a)      Real change has occurred in the characteristic being measured. Such as cell phone ownership may have gone up.

b)      No real change has actually happened, but the survey shows some differences.

Obviously, there is no issue in the first case. Let us now focus on the second case. The observed differences can be due to many reasons. These can be classified into three broad groups:

                                i.            First:  Sampling Error - Sampling error represents the uncertainty in survey estimates that occurs because we observe data on a sample in the population rather than on every unit of population. Any sample survey, including the best and the largest in the world, will have variations in estimates, simply because it is a sample survey. One can only minimize sampling error by designing samples to provide the most precise estimates at available resources. There is no way to avoid this error other than to conduct a Census.  Sampling error is often expressed as standard errors or in simple terms 'Margin of Error‘ (at a design confidence level, for estimates). The magnitude of Margin of Error depends on the incidence of observed characteristic and the sample size. (Refer below the paragraph on Sampling Variations)

                              ii.            Second: The survey design (the theoretical parameters such as representativeness and accuracy of the household selection frame, in our case the electoral rolls, etc.) will have a role to play in this.

                            iii.            Third: Non-Sampling Errors - Lastly, errors creeping in due to non-response, respondent’s understanding of questions, wrong or incomplete responses from respondents,  respondent selection processes not followed accurately, interviewer mistakes and errors in data punching or processing etc.

Our objective, is to

a)      Operate within the ranges defined by globally accepted random sampling methods (i.e. within ‘First point’ above)

b)      Create a robust theoretical design to minimize errors referenced under ‘Second’.

c)      Control processes as much as possible such that errors occurring due to ‘Third’ are kept to a minimum.

Sampling Variations:  

The level of sampling variation (Margin of Error) is what the survey designers have agreed to accept (indicated by the survey’s reporting standard) for any survey estimate. The IRS reporting standards define that the estimates be reported within 20% Margin of Error at 90% confidence level for an incidence of 10%.

Let’s explain a few related and important points on this:

(i)     90% confidence means that if a survey were to be conducted 100 times, on 90 occasions the variation would be within range defined by the reporting standard. Please note that this means that in 10% cases the estimate may well be beyond the defined range.

(ii)   IRS estimates are not point estimates, but a range estimate and the range depends on the 'Margin of Error' associated with each estimate. And as per IRS design statistics, there are 90% chances that the actual estimate lies within this range (lower and upper confidence limit).

(iii) IRS reporting standards define that the estimates be reported within 20% Margin of Error at 90% confidence level for an incidence of 10%. Hence, a higher incidence will have a lower Margin of Error and a lower incidence will have a higher Margin of Error.

×