2016 aapor mari harris

Session: Maintaining Data Quality During Fieldwork in International CAPI Surveys

Survey planning and logistics in Africa – catching up and setting the bar high in the 21st Century Mari Harris Director Ipsos South Africa Saturday, 14 May 2016

For many years, all surveys in Africa were done via the rather long, cumbersome and laborious route of pen-and-paper (PAPI) data collection; maps photocopied from map books were taken along on trips and once the sampled area was reached, prescribed random sampling procedures could take a long time to implement. Paper copies of questionnaires and show cards took up lots of space in cars – and were subject to lots of “disasters”.

Although interviewers still travel to remote areas of the continent, CAPI and the exponential growth of mobile technology have changed the rest of this picture completely: • the road maps are accessible on satnav, making

trip planning easy; • GIS is used to create electronic maps of the

sampled areas; • questionnaires and show cards are scripted and

available on mobile phones or laptop computers – no more manual data entry;

• much less data errors and data cleaning; • no logic mistakes and a much faster turnaround

time… • and more educated and motivated interviewers!

DISCUSSION POINTS

• Africa and technological advances

• Complicated selection procedures made easy

• Inherent Data Quality

AFRICA AND TECHNOLOGICAL ADVANCES

FROM DARK CONTINENT TO IMPORTANT PLAYER

The African economy has trebled from 2002 to 2015

Literacy levels are improving; more girls are attending school

Africa is home to almost a billion people

© 2015 Ipsos.

8 © 2015 Ipsos.

8 © 2015 Ipsos.

8 © 2015 Ipsos.

8 © 2015 Ipsos.

8

SUB-SAHARAN AFRICA GROWTH

Angola 4.5% 3.9%

Ethiopia 8.7% 8.1%

Ghana 3.5% 5.7%

Ivory Coast 8.2% 7.6%

Kenya 6.5% 6.8%

Mozambique 7.0% 8.2%

Nigeria 4.3% 4.0%

South Africa 1.3% 0.7%

Tanzania 7.0% 7.0%

Uganda 5.2% 5.5%

Zambia 4.3% 4.0%

• Commodity cycle • China • Drought • Corruption • Fall in Oil prices

• Tourism • Demographics • Institutional strengthening

2015 2016

!

AFRICA

Population Internet Users Internet Users Penetration Facebook

(2015 Est.) 31-Dec-00 30-Nov-15 (% Population) 15-Nov-15

Angola 19,625,353 30,000 5,102,592 26.00% 3,300,000

Cote d'Ivoire 23,295,302 40,000 5,230,000 22.50% 1,800,000

Ethiopia 99,465,819 10,000 3,700,000 3.70% 3,700,000

Ghana 26,327,649 30,000 5,171,993 19.60% 2,900,000

Kenya 45,925,301 200,000 31,985,048 69.60% 5,000,000

Mozambique 25,303,113 30,000 1,503,005 5.90% 1,200,000

Namibia 2,212,307 30,000 470,000 21.20% 470,000

Nigeria 181,562,056 200,000 92,699,924 51.10% 15,000,000

South Africa 54,777,809 2,400,000 26,841,126 49.00% 13,000,000

Tanzania 51,045,882 115,000 7,590,794 14.90% 2,700,000

Uganda 37,101,745 40,000 11,924,927 32.10% 1,800,000

Zambia 15,066,266 20,000 2,711,928 18.00% 1,300,000

SSA INTERNET USER STATS 2000 - 2015

http://www.internetworldstats.com/stats1.htm

Do you own a cell phone? Is

it a smart phone?

Smart but not cheap While the prices for smartphones are falling fast, they’re still unaffordable to many on the continent. About a third of Africans live on less than $1 a day, which means even if a smartphone falls to $50, that’s 50 days of wages. Data and airtime costs are also a major hurdle. High data costs are thought to be a main reason the digital divide is proving so difficult to narrow.

Most smartphones need a charge every day or maybe every second day. For those living in areas with limited access to an affordable, reliable electricity supply, that’s a deal breaker.

Network coverage

Network coverage in Africa is still patchy in many areas. While smartphones’ potential for bringing educational opportunities to rural Africans has been much talked about, it’s impossible to stream an instructional video on a smartphone if you don’t even have a strong enough signal to make a phone call.

The high cost of electricity

12 © 2015 Ipsos.

RESEARCH CONSIDERATIONS

Thousands of local languages + English, French, Portuguese.

More than 3,000 unique ethnic

groupings. Adult literacy average

33.3%.

Limited demographic data sources, old census data.

Poor maps. Lack of incidence data,

market data, etc.

Over 60% of population living in rural areas, this impacts on resources,

project timings and costs. Travel time – long flights

and local traffic jams. Visa processing

difficulties.

Facilities need careful vetting to ensure

compliance with global standards.

Internet and communications can be

problematic.

LANGUAGE / CULTURAL BARRIERS

DATA VACUUM LOGISTICS FACILITIES AND

INFRASTRUCTURE

13 © 2015 Ipsos.

COMPLICATED SELECTION PROCEDURES MADE EASY

14 © 2015 Ipsos.

RANDOM SELECTION PROCEDURE

The requirements of selecting a stand/plot, a household and then a person in the household did not change, however, technology has speeded this process up considerably.

This procedure could take 20-30 minutes in the

past, now it is done in a few minutes.

15 © 2015 Ipsos.

RANDOM SELECTION PROCEDURE

16 © 2015 Ipsos.

RANDOM SELECTION PROCEDURE

17 © 2015 Ipsos.

18 © 2015 Ipsos.

RANDOM SELECTION PROCEDURE

19 © 2015 Ipsos.

20 © 2015 Ipsos.

21 © 2015 Ipsos.

22 © 2015 Ipsos.

23 © 2015 Ipsos.

INHERENT DATA QUALITY

24 © 2015 Ipsos.

PHYSICAL MANUAL DATE ENTRY

• High error rates. • Data cleaning a tedious and specialised

process. • What to do if the “wrong” person was

interviewed (in a remote village 500 miles away).

25 © 2015 Ipsos.

26 © 2015 Ipsos.

27 © 2015 Ipsos.

28 © 2015 Ipsos.

29 © 2015 Ipsos.

? • Technological advances brought about a revolution i.t.o. data collection

and data quality.

• In Africa we “leap-frogged” from PAPI to CAPI – in most countries straight to mobile.

• We are also making strides in other areas: CATI (to be grown from SA); CAWI (penetration figures are still low).

• This paper highlighted these strides and illustrated that – although survey planning and logistics still warrant detailed attention – the continent is benefiting from a range of technological developments.

• We still have a long way to go, but it is important to take stock of how far we have come as we prepare for the next growth spurt.

30 © 2015 Ipsos.

31 © 2015 Ipsos.

THANK YOU!