Special Notes New SAARF AMPS Universe DS-CAPI Since 2009, the AMPS Universe is defined as adult males and females aged 15 years or older. (Previously, the universe included adults aged 16 years or older). The SAARF AMPS Jan 11-Dec 11 data release is the fifth 12-month datafile to incorporate this new universe. SAARF AMPS 2009 was the first year that the survey used Double Screen-CAPI (DS-CAPI) technology which was recommended by an international media research consultant in the 2007 AMPS audit. In short, DS-CAPI replaced printed show cards and masthead shufflecards with electronic prompt screens shown to the respondent on a separate laptop. The prompt screens appear automatically on the respondent s laptop screen as the interviewer progresses through the questionnaire. The use of colour screens makes the interview more enjoyable for respondents, and results in improved respondent engagement and co-operation. DS-CAPI further enhances the quality of the AMPS survey to produce even more accurate readership results. Impact of DS-CAPI on SAARF AMPS DS-CAPI has resulted in a new benchmark for Print. As of the SAARF AMPS Jul 08-Jun 09 data release, readership results are not comparable to previous releases. There are differences for many publications in different directions. Readership changes are therefore not an indication of a title s shifting performance in the marketplace. Please see special note on Readership for more detail on how the readership methodology has changed due to DS-CAPI. DS-CAPI has had a lesser effect on the rest of the questionnaire data. Television, Radio, Cinema, Internet and Outdoor are deemed comparable to previous survey results and reflects real trends in these areas. Population 2011 Population figures for SAARF AMPS Jan 11-Dec 11 have been updated using the 2011 population estimates supplied by IHS Global Insight from their Regional explorer suite of models. The update took into account the mid-year 2011 estimates from Statistics South Africa. It is important to note that the methodology employed in the demographic model is different from the previous BMR (Bureau of Market Research) model (2010 and earlier) in a number of technical areas. The resultant population updates are therefore different to the previous estimates for a number of demographics. These changes in the population figures must be taken into consideration when comparing thousands with previous SAARF AMPS surveys. (See Section 4 Universe and sample for more detail). New Household Population As of the SAARF AMPS Jul 08-Jun 09 release, a change to the household weighting procedure on the AMPS Survey has been made. The aim of this change was to correct the imbalance in household size that was identified during an AMPS audit in 2007. The AMPS sample consists of a separate male and female sample with one address at each point being pre-designated as a male interview, and the other a female interview. At each address, all adult members of the specified sex are listed on a grid and one member is chosen according to a random grid. The fact that one adult of a specific sex is selected per sampled household, mathematically leads to an over-representation of households with fewer members of the specified sex, and an under-representation of households with more members of the specified sex. Factors were applied within each sex to down weight households with fewer adult members of that sex and up weight households with more adult members of that sex. This correction had minimal effect on results weighted to the adult population, for example media. However, it had a significant impact on household variables. Users must regard SAARF AMPS Jul 08- Jun 09 as a new starting point for tracking household variables analysed on a household weight and results cannot be compared with previous surveys. Results for household variables in this SAARF AMPS Jan 11-Dec 11 release are, therefore, only comparable with those in SAARF AMPS Jul 08- Jun 09 and releases thereafter. Also note that the Household population has been updated in line with the new IHS Global Insight population estimates for this SAARF AMPS Jan 11-Dec 11 release. Page 7
Readership Store Magazines Whilst the actual wording of the readership questions remained unchanged, DS-CAPI changed the SAARF AMPS readership methodology as follows: Colour mastheads were used instead of black and white; Individual masthead shufflecards were replaced by screens of nine mastheads; Manual shuffling of a pack of masthead cards for rotation is replaced by a computer-chosen rotation of the various newspaper and magazine screens (i.e. newspapers versus magazines; forward versus reverse order; clockwise rotation of mastheads per screen); Masthead prompts were included in all readership questions; Paper prompt lists i.e. for When Last Read and Origin of Copy now appear on screen; Potential title confusion has been taken into account in grouping the nine mastheads on a screen; Daily and Weekly newspapers are measured together with no regionalisation; All supplements for a newspaper are shown on one screen and in colour. DS-CAPI has resulted in a new benchmark for Print. It is important to note that this new measurement for SAARF AMPS Jul 08-Jun 09 onwards, should not be compared with previous AMPS readership results. During 2009, the decision was made to measure each individual publication separately and not to have any titles grouped together in the interview. This was implemented on the Jul-Dec 2009 AMPS survey and measured in the same way for the Jan-Jun 2010 AMPS survey. This meant that magazines that were published in two different languages had two prompts, i.e. one for each version. For the Jul-Dec 2010 AMPS survey, the SAARF Councils decided to do away with the double prompting of multilingual publications with identical or almost identical mastheads. This methodology was continued for the Jan- Jun 2011 and Jul-Dec 2011 AMPS surveys. On the database, those publications with combined ABCs are released as one combined figure and no separate figures are published. From Jul-Dec 2009, newspaper supplements are no longer regionalised. All supplements are shown nationally. The criterion for reporting of publications is that the said publication must have a minimum of 40 respondents that qualify as average issue readers. SAARF issued this press release regarding the store magazines measured on the SAARF AMPS questionnaire on Friday, 19 th March 2010: The SAARF AMPS Scrutiny Committee has decided to withhold the AMPS 2009AB store magazine readership results, to give the advertising, media and marketing industries the opportunity to carefully evaluate this sector s results. At an AMPS scrutiny meeting held on Wednesday, 17 March 2010, SAARF informed representatives of the SAARF Councils that it was concerned about the reliability of the store magazines readership figures in AMPS 2009AB (fieldwork period Jan Dec 2009). In 2009 the PMSA requested that separate mastheads be shown for all individual magazines. The store magazine prompts were expanded in line with this decision. Now, instead of measuring 10 store magazines as in previous surveys, AMPS 2009AB measures 23 store publications. The result of this change to separate masthead prompts has been an increase in readership for many of the store magazines, against circulation trends. A possible contributing factor is that respondents are confusing marketing material with the actual store title. Many of the stores in question produce prominently-branded catalogues and inserts, in similar formats to their magazines, which are distributed not only to club members but to the general public. AMPS contractor, Nielsen, supports this view, saying that where the chain s name is prominent in the store magazine s title, or is visually dominant in the masthead, readership figures have grown substantially. SAARF is in the process of contacting the publishers of magazines in this sector, to ensure they are aware of the issue and that they contact the PMSA with their suggestions and comments. The PMSA will also take up this issue at its next research committee meeting. After discussing the way forward with its stakeholders, it will revert to the SAARF Print Council with recommendations on future actions in this regard. In the meantime SAARF and Nielsen will study the outcome of the latest AMPS wave and come up with proposals on how these publications should be handled on AMPS, if at all. Subsequently, the SAARF Councils agreed to pilot a new readership methodology for store magazines on the SAARF AMPS Jul-Dec 2010 and Jan-Jun 2011 surveys. The filter question for publications read or paged through in the past 6 months has been adapted to introduce store magazines separately from newspapers and other conventional magazines. An explanation is given of what is meant by store magazines and the showscreens differ from those for newspapers and conventional magazines i.e. the store magazines are listed alphabetically by store name with a picture of its masthead alongside it. Page 8
Store Magazines (Cont.) In May 2011, the 6-month (Jul-Dec 2010) store magazine pilot results were evaluated. The SAARF AMPS scrutiny Committee looked for decreased results, but for most of the store magazines the total readership was above 2009 levels. It was suggested at the Scrutiny meeting that the data for these publications be filtered by own/household copy to downweight the data to more realistic levels. Furthermore, it was proposed that the showscreen format be changed back to 9 mastheads on a screen, with mastheads rotated as for other publications, but keeping the upfront 6-month filter question for store magazines a separate question. The Print Council accepted the proposed changes. The Jul-Dec 2011 questionnaire was adapted. The SAARF AMPS Jul 10-Jun 11 release and the current SAARF AMPS Jan 11-Dec 11 release include store magazine readership results filtered for own/household copy. Changes to Publications Echo HUSTLER and loslyf Saltwater girl magazine SARIE KOS Vuk uzenzele Sunday Times - Money & Careers Sunday Times - Lifestyle Magazine The Herald - Leisure Refer to Section 6 The Interview for details for all publication changes for SAARF AMPS Jul-Dec 2011. It was agreed by the Print Council to measure Echo as a weekly supplement of The Witness on SAARF AMPS from 2007. It must however, be noted that Echo also has a big bulk free distribution on its own. The readership of Echo obtained from SAARF AMPS will therefore, be underread. HUSTLER and loslyf were both measured as six-weekly magazines in the January-June 2011 survey. Their publishing frequency was changed to monthly and was measured as such on the July-December 2011 survey. Due to this change, the Number of Issues (B2) data has not been released on this 12- month SAARF AMPS Jan 11-Dec 11 datafile and note that the AIR is a combination of the previous and current 6-month s AIRs. Saltwater girl magazine changed from being published six-weekly (Jan-Jun 2011 survey) to being published every alternate month (Jul-Dec 2011 survey). For this reason, the data from Number of Issues (B2), has not been released for this magazine on the SAARF AMPS Jan 11-Dec 11 datafile and note that the AIR is a combination of the previous and current 6-month s AIRs. SARIE KOS changed its publishing frequency from quarterly to alternate monthly in April 2011. The magazine was measured as a quarterly publication on the Jan-Jun 2011 survey and as an alternate monthly magazine on the Jul-Dec 2011 survey. For this reason Number of Issues (B2) has not been released for SARIE KOS on the SAARF AMPS Jan 11-Dec 11 datafile, and note that the AIR is a combination of the previous and current 6-month s AIRs. Vuk uzenzele changed from an alternate monthly magazine to a monthly magazine in May 2011. For the Jan-Jun 2011 survey, it was measured as an alternate monthly magazine and for the Jul-Dec 2011 survey as a monthly magazine. Due to this change in frequency, Number of Issues (B2) has not been released on the SAARF AMPS Jan 11-Dec 11 datafile and the AIR is a combination of the previous and current 6-month s AIRs. The supplement Careers in the Sunday Times has changed to Money & Careers. The data reflected in the SAARF AMPS Jan 11-Dec 11 release is a combination of Careers (as measured on the Jan-Jun 2011 survey) and of Money & Careers (on the Jul-Dec 2011 survey). The supplement Lifestyle and the Sunday Times Magazine (both measured as separate supplements on the Jan-Jun 2011 survey) were replaced with the new supplement Lifestyle Magazine (as measured on the Jul-Dec 2011 survey). The figure released on the SAARF AMPS Jan 11-Dec 11 datafile for Sunday Times s Lifestyle Magazine is a combination of Lifestyle, Sunday Times Magazine and the new Lifestyle Magazine. The Herald s weekly supplement tgif, as measured on the Jan-Jun 2011 survey, was replaced with the supplement Leisure for the Jul-Dec 2011 survey. The publisher agreed that the figure released on the 12-month SAARF AMPS Jan 11-Dec 11 database be a combination of the old tgif and the new Leisure supplements. Page 9
Titles not reported in SAARF AMPS Jan 11-Dec 11 Titles only found in the SAARF AMPS Dec 11-6m Datafile Publications that were not common to both SAARF AMPS Jan-Jun 2011 and SAARF AMPS Jul-Dec 2011 or did not have sufficient average issue readers or ceased publication during fieldwork (Jan-Dec 2011), are not reported in the Rolling database: avocado/avokado Bignews DESTINY MAN FAIRLADY food HUISgenoot Pols/YOU Pulse intiem Mamas & Papas Move! BestParent people puzzles PRETORIA NEWS WEEKEND Privé (Sun International) PSYCHOLOGIES MAGAZINE SOUTH AFRICA SA India SL MAGAZINE SoccerLife FourFourTwo SoccerWeek The Zimbabwean VISI Wiel Wine MAGAZINE YOUR BUSINESS ZiGZAG Titles that were newcomers to the SAARF AMPS Jul-Dec 11 questionnaire have been included in the SAARF AMPS Dec 11-6m datafile if they qualified on a minimum of 40 average issue respondents: Mamas & Papas SA India DESTINY MAN YOUR BUSINESS Radio Listenership All questions pertaining to Radio listening patterns were deleted from the SAARF AMPS 2003 questionnaires and only measured through the SAARF RAMS Radio Diary. However, due to the new SAARF RAMS data release structure, the past 4 week, past 7 day and yesterday radio listening questions were re-incorporated into the SAARF AMPS 2004 questionnaire. Users should take note of this when tracking SAARF AMPS Radio data. Community service stations are reported as Total Community Radio in SAARF AMPS electronic reports. This allows users to ascertain the part that community radio plays in the media mix. Community Radio Stations with 40 or more mentions are released individually on the database. For further analysis of community radio, users are referred to the SAARF RAMS release databases. Refer to Section 6 - The Interview - for details of the radio station changes for SAARF AMPS Jul-Dec 2011. Page 10
TV Viewership As of the July-December 2010 AMPS survey, the lists of television channels have been separated out for households receiving terrestrial television only, subscribing to DStv and subscribing to TopTV. All three lists have an option to state any other TV viewed. The electronic reports include breakdowns for Total DStv, Total TopTV and Total Community TV. From the SAARF AMPS Jul 10-Jun 11 release, the Total DStv and Total TopTV codes include all channels available on these services including the terrestrial services: SABC 1, SABC 2, SABC 3, e.tv and M-Net (for DStv) that were previously excluded from these totals. Total Community TV currently includes Soweto TV, CTV (Cape Town TV), Bay TV and Tshwane TV. The Total DStv figure in this Jan 11-Dec 11 AMPS release is therefore only comparable with the previous Jul 10-Jun 11 DStv codes. For further analyses of satellite channel viewing, users are referred to the SAARF AMPS database. The figures for SABC 1, SABC 2, SABC 3, e.tv, and M-Net reflected in the electronic reports and on the database still reflect total viewership for these stations regardless of the platform through which they are viewed. Refer to Section 6 The Interview for details of all TV stations changes for SAARF AMPS Jul-Dec 2011. Cinema Three questions were added to SAARF AMPS Jul-Dec 2001 questionnaire to obtain the number of visits in the last 3 calendar months. This inclusion is to extend measurement of cinema visits to the Dec/Jan holiday period, which is not covered during the SAARF AMPS fieldwork period. Average 4- week cinema attendance over past 12 months is included in the electronic reports. From SAARF AMPS 2006, the recency scale includes an option for yesterday. Out of Home In adapting the Out of Home section to DS-CAPI for 2009, the prompt material for the various Out of Home advertising mediums changed from a single black and white card with pictures of the 9 out of home types to 9 individual colour screens. From SAARF AMPS Jan 10-Dec 10, additional breakdowns are shown in the electronic reports for Any Out of Home advertising seen yesterday, up to 7 days ago and up to 4 weeks ago. From SAARF AMPS Jan-Jun 2011, the question wording and frequency options for Out of Home media have been revised. The list of Out of Home media increased from 9 to 11 types. In addition, the screen visuals of the different types of Out of Home media were updated. This SAARF AMPS Jan 11-Dec 11 shows the first full 12-month impact of the changes. Please keep this in mind when comparing the current data with previous releases. Financial Services / Products In the SAARF AMPS Jan-Jun 2001 questionnaire, those respondents who said their house/flat/dwelling was owned, were asked whether or not there was a mortgage bond/home loan on the home. As from the SAARF AMPS Jul-Dec 2001 questionnaire, this filtered question was deleted, and instead, Home Loan/Mortgage Bond was included on the show card for financial services. All respondents were asked to indicate which of these services they personally have or make use of. This data is therefore not comparable with data prior to the SAARF AMPS Jan 01-Dec 01 release. From SAARF AMPS 2007, the question asking which financial services have been opened in the past 12 months (Question K1), has been adapted to be less specific and only ascertains whether or not any account was opened in the past 12 months. The SAARF AMPS Feb 06-Jun 07 database and electronic reports (and data released thereafter), therefore, only include the one figure for opening any financial services account in the past 12 months. The same holds for policies, investment plans and insurance (Question K10). The SAARF AMPS Feb 06-Jun 07 database and electronic reports (and data released thereafter), only include a figure for any financial product (long term policies/investments/insurance) taken-out in the past 12 months and one figure for short term insurance taken out in the past 12 months. From the SAARF AMPS Jan-Jun 2009 survey, micro loan has been combined with other personal loan and the account type is referred to as Personal loan. From the SAARF AMPS Jan-Jun 2010 survey, ATM Card has been removed from the list of financial services and Savings/Transmission account has changed to Savings account. From the SAARF AMPS Jan-Jun 2011 survey, the Investments/Subscription shares and/or paid-up shares financial service description (K1) has changed to Investments/Deposit accounts. The word current has also been added to the Cheque account option. Please keep this in mind when analysing and comparing with previous survey data. (See K1 and K10 of the questionnaire in the Appendix of this report). Page 11
Activities 4 weeks / 12 months FMCG (Product) Categories Most of the activities covered on the SAARF AMPS 2006 questionnaire moved to the 2007 Products and Activities self-completion questionnaire. The database and electronic reports only include those activities that remained on the SAARF AMPS questionnaire (see question J1 of SAARF AMPS questionnaire in Appendix of this report). From SAARF AMPS 1992 to 2000, FMCG categories were included in the main SAARF AMPS questionnaire with respondents classifying themselves into usage/purchase categories - Heavy, Medium, Light, Never from 1994 onwards. A self-completion questionnaire measuring FMCG categories on unit-bases was introduced from SAARF AMPS 2001. The number of units used/purchased is obtained and for each product category, an ever measure has been calculated as well as the component light, medium and heavy responses. The responses are divided into best thirds to form the light, medium, heavy categories. In addition, some categories had branded questions on a user-pay basis. From SAARF AMPS 2005, the self-completion questionnaire has been expanded to cover over 100 FMCG product categories and includes three branded questions for each category. The light, medium, heavy categories were re-examined for the SAARF AMPS Jul 08-Jun 09 release. The following product categories had changes to their light, medium, heavy categorisation: Fresh milk Dog food dry Hair styling products Baby food Iced tea Infant cereal Mineral water Toilet paper Sanitary pads Soup Tampons Chicken fresh/frozen Bread Vegetables frozen Cat food wet Potatoes frozen Cat food dry Flavoured milk Dog food wet Herbs/spices Refer to Section 6 The Interview for details of all changes made to the FMCG (Product) categories for SAARF AMPS Jul-Dec 2011. On the 12-month SAARF AMPS database, only comparable (across the two 6-month surveys) product categories (and brands) are released. Age Language There are two age questions in the SAARF AMPS questionnaire one on exact age and the other a grouping into 8 categories. However, the weighting cells are 15-24, 25-34, 35-49 and 50+ (not the individual groups) and analysis by any other age groupings will not necessarily reflect the correct population statistics. In addition, there is a proportion of respondents who refuse to give their exact age and therefore adding the individual exact ages will not necessarily equate to the groups in the grouped age question. Since SAARF AMPS 94, respondents were able to claim more than one home language. From 1994 to 2000 the language categories for reporting purposes were grouped into English/Other, Afrikaans/Both, Nguni/Other and Sotho/Both with no duplication between categories. The decision was made from SAARF AMPS 2001 to use all 11 official languages, with an additional category in which all other languages are grouped as a breakdown in the report. For reporting purposes, the language categories are as follows: English South Sotho Tswana Afrikaans North Sotho Venda Xhosa Ndebele Swazi Zulu Tsonga Other Each language code comprises only those respondents claiming that specific language as a home language. In order to remove duplication in home language, two census language questions replace the home language question in the SAARF AMPS Jul-Dec 2002 questionnaire. Firstly, the language personally spoken most often is obtained followed by any other languages personally spoken at home. For SAARF AMPS Jan 02-Dec 02 12-months database and reports, these two questions were combined and matched with the SAARF AMPS Jul 01-Jun 02 language data. Due to the methodological change in the home language questions, SAARF AMPS Jan 02-Dec 02 was not comparable to previous surveys. From the SAARF AMPS Jul 02-Jun 03 12-month database, users have been able to analyse home language unduplicated. Page 12
Community With the annual update to the population estimates, a number of places changed their community size definitions from SAARF AMPS Jul 10-Jun 11. SAARF Lifestyles Children s Primary Purchase Decision Maker Mothers with Children SAARF Attitudes From SAARF AMPS 2005, SAARF Lifestyles were derived separately for LSM s 1-5 and LSM s 6-10. From SAARF AMPS Jan 07-Dec 07 onwards, a further entirely separate analysis has been done for the total adult population covering LSM s 1-10. For further details, see Section 10 SAARF Lifestyles. Two codes for Primary Purchase Decision Makers for babies up to 23 months and for children 2-15 years old were created and released from the AMPS Jan 07-Dec 07 database. Due to the inclusion of 15 year olds in the SAARF AMPS adult universe in 2009, the Primary Purchase Decision Makers for children 2-15 years changed to Primary Purchase Decision Makers for children 2-14. See Section 3 Definition of Terms for a further explanation. The description of the breakdowns: Women with babies up to 23 months and Women with children 2-15 years were changed to Mothers with babies up to 23 months and Mothers with children 2-15 years in the SAARF AMPS electronic reports from SAARF AMPS Jan 07-Dec 07. Note that only the description changed, not its definition. Due to the inclusion of 15 year olds in the SAARF AMPS universe in 2009, Women with children 2-15 years was changed to Women with children 2-14 years. See Section 3 Definition of Terms. In order to extend the opportunities for a greater understanding of the South African consumer, an attitude section was added to the self-completion questionnaire from SAARF AMPS 2007. Attitude Topics were identified, defined and measured using 4 different four point scales. Respondents were subsequently classified into those who were above or below the mean on each Attitude Topic. They were further grouped into 5 attitude types who hold similar views: Now Generation, Nation Builders, Distants, Rooted and Global Citizens. The Distants attitude type can also be accessed by those who are in LSM s 1-5 and LSM s 6-10. The individual SAARF Attitude statements are also available for analysis. A detailed write-up of SAARF Attitudes appears in Section 11 of this Technical Report. Page 13