Implications for TAMS When SAARF Moved to a New Demographer for the Universe Updates

Implications for TAMS when SAARF moved to a new demographer for the universe updates

On 2July2012TAMS (Television and Audience Measurement Survey)implemented the new 2011 population updates from the IHS Global Insight estimate via the latest 6-months AMPS (All Media and Products Survey)data. These IHS updates, which see the total adult 15+ population growing by 2.7%, had a significant impact on the TAMS results due to the fact that there are some differences in the model they use to determine the population profile compared to the model used by the previous demographers. The industry should carefully examine these population changes, bearing them in mind when dealing with changed audience results as there can be some trend-line breaks.

The new 2011 population updates from the IHS Global Insight estimates have had a significant impact on TAMSUniverses and results.

In Total*, the 15+ population has increased by 2.7%, from 34.02-million to 34.934-million, with a number of demographics changing because of this.

For many media, the new population estimates have also affected audiences in thousands, even if incidence does not reflect any change.

By understanding the effects of the population update, the industry will be able to get a better understanding of what has happened to audience results.

BACKGROUND TO THE POPULATION CHANGE

The previous population updatesreflected mid-year 2010 estimates supplied by the Bureau of Market Research of UNISA (BMR). The next population update was deferred to 2012, in order to take into account the mid-year 2011 estimates from Stats SA.

IHS Global Insight’s demographic model differs from the previously employed BMR model in a number of technical areas:

The cohort component model handles mortality in a different way, specifically with regard to the impact of HIV/Aids. Specifically, the BMR model shows a greater impact.
For regional distribution, slightly different techniques are employed.
The model is checked for consistency with the economic, labour, income and development factors of every region in South Africa.

The resultant population updates are, therefore, different to the previous population estimates for a number of demographics.

*Further down in this article, we explain how the IHS figures are applied to AMPS and then from AMPS on TAMS. Thus at this stage we are just talking about the Total population and not the TV population.

(2010 figures are derived from BMR population estimates, and 2011 figures are derived from IHS Global Insight population estimates.)

SEX / While the changes in male and female proportions have not been significant, the percentages are none the less closer together, indicating slightly more males. However, females still make up the larger proportion of the population.
2010 / 2011
‘000 / % / ‘000 / %
Male / 16 113 / 47.4 / 16 879 / 48.3
Female / 17 907 / 52.6 / 18 055 / 51.7
AGE / There are less 15-24 year olds in the IHS estimate. Numbers for the 25-49 age groups remain similar, while the estimates for 50+ are higher than they were with BMR.
2010 / 2011
‘000 / % / ‘000 / %
15-24 / 10 048 / 29.5 / 9 820 / 28.1
25-34 / 7 749 / 22.8 / 7 905 / 22.6
35-49 / 8 569 / 25.2 / 8 814 / 25.2
50+ / 7 654 / 22.5 / 8 395 / 24.0
Significant increaseSignificant decrease
POPULATION GROUP / Most of the population increase has come from the black community, which has risen by over one million adults. The number of coloureds remains similar, as has the number of Indians. There has been a decrease in the white population by about 300000. Media which target these population groups may see a difference in their audience in thousands.
2010 / 2011
‘000 / % / ‘000 / %
Black / 25 613 / 75.3 / 26 784 / 76.7
Coloured / 2 942 / 8.6 / 3006 / 8.6
Indian / 927 / 2.7 / 936 / 2.7
White / 4 538 / 13.3 / 4 208 / 12.0
Significant increaseSignificant decrease
COMMUNITY SIZE / There have been many shifts in the classification of places. There has been growth in the metros, and two new metros have emerged: Krugersdorp and Germiston. Many places have had their community size either upgraded to city/large town status, such as Knysna and Barberton, while others, such as Brits and Vryheid, have lost this classification and are now classified as small town/village, due to their declining populations
2010 / 2011
‘000 / % / ‘000 / %
$ 250 000+ / 11 541 / 33.9 / 12 700 / 36.4
$ 40 000-249 999 / 4 626 / 13.6 / 4 735 / 13.6
$ 500 -39 999 / 4 148 / 12.2 / 4 223 / 12.1
Total urban / 20 314 / 59.7 / 21 658 / 62.0
Less than 500/rural / 13 706 / 40.3 / 13 276 / 38.0
Significant increase; Significant decrease; $ Movement of places in urban categories
PROVINCE / The Western Cape, Northern Cape, Mpumalanga and Limpopo have remained fairly unchanged in terms of proportion to the total population. The Free State has declined significantly, down by 100000. The Eastern Cape is down slightly. KwaZulu-Natal grew significantly in the previous BMR update, and has maintained these numbers. North-West is significantly down, by 400,000. The greatest difference has occurred in Gauteng, which is up by 1.3-million, which makes sense intuitively, as this is where the country’s growth is. The changes seen on a provincial level reflect the changes in the Community Size populations. The most notable change is the growing tendency towards Urbanization.
2010 / 2011
‘000 / % / ‘000 / %
Western Cape / 3 524 / 10.4 / 3 556 / 10.2
Northern Cape / 718 / 2.1 / 796 / 2.3
Free State / 2 210 / 6.5 / 2078 / 5.9
Eastern Cape / 4 734 / 13.9 / 4 672 / 13.4
KwaZulu-Natal / 7 148 / 21.0 / 7 117 / 20.4
Mpumalanga / 2 340 / 6.9 / 2 487 / 7.1
Limpopo / 3 752 / 11.0 / 3 724 / 10.7
Gauteng / 6 882 / 20.2 / 8 201 / 23.5
North West / 2 712 / 8.0 / 2 303 / 6.6
Significant increaseSignificant decrease
METROPOLITAN AREA / In general, metros as a group have increased in size. All the Gauteng metros have grown – Pretoria and Reef in particular – with the exception of Soweto which has declined by 100000. Durban has maintained but there has been a slight decrease in Pietermaritzburg as is the case for Cape Town.
These changes could affect radio stations and other regional media.
2010 / 2011
‘000 / % / ‘000 / %
Cape Town / 2 298 / 6.8 / 2212 / 6.3
Cape Town Fringe / 332 / 1.0 / 313 / 0.9
PE/Uitenhage / 853 / 2.5 / 885 / 2.5
East London / 386 / 1.1 / 388 / 1.1
Durban / 2 063 / 6.1 / 2096 / 6.0
Pietermaritzburg / 372 / 1.1 / 317 / 0.9
Kimberley / 133 / 0.4 / 139 / 0.4
Bloemfontein / 317 / 0.9 / 358 / 1.0
Vaal / 716 / 2.1 / 840 / 2.4
Greater Johannesburg / 1 934 / 5.7 / 2127 / 6.1
Soweto / 907 / 2.7 / 804 / 2.3
Reef / 2 710 / 8.0 / 3365 / 9.6
Pretoria / 1 344 / 3.9 / 1615 / 4.6
Significant increaseSignificant decrease

The changes described above are to the total adult population of South Africa (latest projections supplied by IHS). These projections then get applied to AMPS (All Media and Products Survey).AMPS is the establishment survey for TAMS (Television Audience Measurement Survey).

An establishment survey is the only way to correctly reflect the relative sizes of all the different TV platforms at a given moment in time as the demographer does not give projections or updates on the “TV population”. Incidence of this is sourced from AMPS®.

The TAMS® TV universe is made up of all households, both rural and urban, with mains electricity and at least one working TV set. TAMS® measures all platforms both analogue and digital and both terrestrial and satellite.

While people are continuously buying new TV equipment, it is sufficient to update the TAMS® universe only once every six months after new AMPS® data is released, usually in January and August. Updating the universe regularly ensures that the TAMS® data remains up-to-date with changes happening in the TV arena.

The latest TAMS universe update was implemented on 2 July 2012.

The tables below reflect changes between the TAMS universe using the previous demographic contractors data applied to AMPS 2011A and with the estimates from IHS Global Insights applied to AMPS 2011B:

TAMS Universe Update 2 July 2012: Move from 2011A to 2011B (6 month AMPS data)
Adults 15+ TV Population
2011A / 2011B / #Difference / % Difference
Total TVPopulation / 29209 654 / 30039 437 / 829 783 / 2.8
Nguni / 11027 342 / 11979 797 / 952 455 / 8.6
Sotho / 9997 586 / 10236 856 / 239 270 / 2.4
English/ Other / 3433 374 / 3194 837 / -238 537 / -6.9
Afrikaans/ Both / 4751 351 / 4627 947 / -123 404 / -2.6
LSM 1 – 4 / 4727 909 / 4889 073 / 161 164 / 3.4
LSM 5 / 5609 036 / 5630 683 / 21 647 / 0.4
LSM 6 / 7031 159 / 7518 738 / 487 579 / 6.9
LSM 7 / 3862 070 / 3876 841 / 14 771 / 0.4
LSM 8 / 2784 117 / 2889 648 / 105 531 / 3.8
LSM 9 / 3007 468 / 3129 350 / 121 882 / 4.1
LSM 10 / 2187 894 / 2105 104 / -82 790 / -3.8
Metropolitan / 11008 610 / 12175 340 / 1166 730 / 10.6
City/ Large Town / 4320 409 / 4452 069 / 131 660 / 3.0
Small Town/ Village / 3795 804 / 3812 746 / 16 942 / 0.4
Settlement/ Rural / 10084 831 / 9599 280 / -485 551 / -4.8
Western Cape / 3415 562 / 3380 152 / -35 410 / -1
Northern Cape / 641 682 / 664 079 / 22 397 / 3.5
Free State / 1985 953 / 1884 835 / -101 118 / -5.1
Eastern Cape / 3414 608 / 3253 853 / -160 755 / -4.7
Kwa-Zulu Natal / 5444 353 / 5541 976 / 97 623 / 1.8
Mpumalnga / 1978 497 / 2212 646 / 234 149 / 11.8
Limpopo / 3406 369 / 3194 212 / -212 157 / -6.2
Gauteng / 6477 456 / 7753 756 / 1276 300 / 19.7
North West / 2445 174 / 2153 928 / -291 246 / -11.9
Significant increaseSignificant decrease

SUGGESTIONS ON HOW TO LOOK AT THE TAMS DATA BEARING THE NEW UNIVERSE CHANGES IN MIND

As different television stations have different target markets, not all of them will have been affected to the same degree by the new universe updates. For example, if a big part of a specific channel’s target market is Nguni speakers, the fact that the new universe update includes a significant increase in the Nguni speaking population will influence the channel more than a channel that does not target this market. The viewers in 000’s will grow while the level of viewing expressed as a percentage may remain stable. Changes in the data, will, of course, also be affected by the broadcast content.

Users should also note that when there is a universe update (whether done by a new demographer or the same demographer) thefigures in thousands that you look at may not be directly comparable with one another. In this instance the figures for the 15+ population in the TV universe went from 29.2 million to 30 million.

Other aspects that users should bear in mind when comparing TAMSdata before and after a universe update are the following:

Make sure you have not eroded your sample size with too many filters. Data on small sample sizes is unstable.
Look at the whole picture i.e. even if you are only interested in LSM 8-10, do a run that has LSM 1-4, LSM 5-7 and LSM 8-10 in it. This way you can look at your “target market” in relation to the rest. In fact you might find out that your “true” target market is actually LSM 5-7 and not LSM 8-10.
Make sure there were no unusual events like the “World Cup Soccer” in either you source or projected period.
Don’t just look at your channels, also look at competing media and other activities
Look at extraneous factors like a bad economic downturn
Remember Margin of Error

For more information on universe updates, please also go to:

The universe update was not the only change to TAMS in the last months. On 6 August 2012, some RIMS were also unbundled and there is a separate article on this.

Please bear both these changes in mind as your source period may be on the old universe or pre-unbundling of RIMS and your projected period might have either one or both changes in it.

END

TAMSPage 1