Upload
hoangkhanh
View
215
Download
0
Embed Size (px)
Citation preview
6
CHAPTER 2
RESEARCH METHODOLOGY
2.1 INTRODUCTION
In order to establish the levels of satisfaction of primarily households and businesses
regarding municipal service delivery, surveys were conducted in the jurisdiction areas
of the City of Tshwane (CoT). One survey was conducted amongst households, a
second was conducted amongst businesses and a third amongst embassies. This
chapter elaborates on the research methodology applied during the execution of the
2011 household, business and embassy surveys that largely resembles the
methodology applied in 2009. The discussion will provide a basis for the scientific
foundation of the study and hence the quality, validity and reliability of the data. The
first part of the chapter focuses on the research methodology used for the household
satisfaction survey. This is followed by a discussion of the business survey research
methodology. Finally, the chapter focuses on the research methodology used to
survey the foreign embassies.
2.2 HOUSEHOLD SURVEY RESEARCH METHODOLOGY
The research methodology applied to collect primary data from households residing in
the CoT is explained below.
2.2.1 Survey population
The survey population is defined below in terms of sample units, sample elements and
the geographical demarcation of the survey area.
2.2.1.1 Sample units The sample population of the survey comprised all households residing in the CoT. For
purposes of selecting sample units for interviewing, the following dwelling types
(sample units) were eligible for inclusion:
7
• Formal dwellings such as houses, flats/apartments, townhouses/cluster houses
and rooms (division within a dwelling occupied by a separate household).
• Informal dwellings/shacks on own stand, backyard informal dwellings/shacks
and outbuildings/cottages.
Households occupying any of the above types of dwellings are defined as a person or a
group of persons, who occupy a common dwelling (or part of it) and who provide
themselves with food and other necessities of life. In other words they live together as
a unit. People occupying the same dwelling, but not sharing food and other
necessities, are regarded as separate households.
2.2.1.2 Sample elements
Heads of households were identified as the sample elements for the survey. To qualify
for selection as a respondent a resident should have resided at the household address
for at least six months. Heads of households that were not available/willing for
interviewing were substituted with his/her spouse.
2.2.1.3 Survey areas The survey area comprised the following five regions of the CoT:
• Central Western Region
• North Western Region
• North Eastern Region
• Eastern Region
• Southern Region
8
These regions were subdivided into 76 municipal wards. Each ward was further
subdivided into the following four types of areas by dwelling type:
• formal township area
• inner city
• suburbs
• informal settlement area
The wards and the above strata were used for sample stratification purposes.
Table 2.1 depicts the five CoT regions by ward and the major suburbs initially included
in each region.
TABLE 2.1
REGIONS AND SETTLEMENT TYPE OF THE CITY OF TSHWANE BY REGION AND WARD
Central Western Region
Ward Suburbs 1 Capital Park, Daspoort, Hermanstad, Mountain View, Park Town Estate, Pretoria Gardens, Roseville
3 Philip Nel Park, Pretoria, Proclamation Hill, West Park
7 Lotus Gardens
51 Atteridgeville (25‐29) 55 Andeon, Booysens, Claremont, Kirkney, Elandspoort, Suiderberg, Lady Selborne, Danville
56 Brooklyn, Lynnwood, Colbyn, Hatfield, Hillcrest
58 Arcadia, Sunnyside, Riviera 59 Groenkloof, Lukasrand, Muckleneuk, Nieu Muckleneuk, Hatfield, Sterrewag
60 Pretoria Industrial, Kwaggasrand, Westpark
62 Atteridgeville
63 Atteridgeville (3,6)
68 Atteridgeville (7,17)
71 Atteridgeville (16,18), Blair Athol
72 Saulsville
9
North Western Region
Ward Suburb
2 Dorandia, Florauna, Ninapark, Pretoria North, Theresapark, Tileba 4 Amandasig, Chantelle, Clarina, Karenpark, Rosslyn, The Orchards, Heatherview
9 Winterveld
11 Soshanguve (FF, GG, HH, JJ, LL)
12 Soshanguve (AA, BB, EE)
19 Mabopane (A,E,N,S)
20 Kopanong
21 Mabopane (a,b,x) 22 Mabopane (R,T,U,V) (Nooitgedacht Dam)
24 Winterveld AH
25 Soshanguve (PP,T,V,W)
26 Soshanguve (DD,KK,P,PP,R,S,SS)
27 Soshanguve (CC,P,R,S,W,X,Y)
29 Soshanguve (BB,FF,IA,LL,NN) 30 Ga‐rankuwa Unit 4,15,16,17,23,24,25)
31 Ga‐rankuwa Unit 1,2,5
32 Ga‐rankuwa Unit (3,6‐10,13,21,22)
33 Soshanguve (AA,F,G)
34 Soshanguve (H,L)
35 Soshanguve (K,M,UU)
36 Soshanguve (K,L,M,WW,YY,ZZ)
37 Soshanguve South (1,2,3)
39 Soshanguve South (4,5,14), Soshanguve East
North Eastern Region
Ward Suburb
5 Doornpoort, Montana, Montana Park, Montana Tuine
8 Dilopye (Temba), Suurman
13 New Eersterust
14 Stinkwater, Tswaing Village, Zoutpan
49 Onderstepoort AH, Wolmer, Mont Lorraine AH, Bon Accord AH, Pyramid / Pyramid Estate AH
50 Annlin, Annlin‐wes, Magalieskruin, Sinoville, Wonderboom AH 73 Hammanskraal (1,2), Marokolong, Ramotse, Mandela Village 74 Hammanskraal West 75 Babelegi, Kudube Unit 1,2,3,6,7,10,11,d
76 Majaneng, Mashemong
10
Eastern Region
Ward Suburbs
6 Mamelodi (15) (east)
10 Mamelodi (8,11,12,20,22)
15 Mamelodi (4,5)
16 Mamelodi (6)
17 Mahube Valley, Mamelodi (7,10)
18 Mamelodi (1,2,3)
23 Mamelodi (14) (Moretele View) (South)
28 Mamelodi (26) (South)
38 Mamelodi (24) (West) 40 Die Wilgers, Equestria, La Montagne, Nellmapius, Samcor Park, Willow Park Manor, Meyerspark
41 Brummeria, Silverton, Weavind Park, Persequor, Koedoespoort
42 Ashlea Gardens, Elardus park, Erasmusrand, Menlo Park, Newlands, Waterkloof, Waterkloof Heights, Menlyn, Waterkloof Glen
43 East Lynne, Jan Niemand Park, Eersterust, Silvertondale, Waltloo
44 Faerie Glen, Garsfontein, Wapadrand
45 Constantia Park, Erasmuskloof, Moreletapark, Pretoriuspark
46 Die Wilgers, Lynnwood Glen, Lynnwood Manor, Lynnwood Ridge
Murrayfield, Val‐de‐grace
47 Elardus Park, Erasmuskloof, Irene, Rietvalleirand, Wingate Park
52 East Lynne, Kilner Park, Lindo Park, Waverley, Moregloed
53 Queenswood, Rietondale, Riviera, Kilnerpark, Rietfontein 54 Eloffsdal, Les Marais, Mayville, Rietfontein, Villieria, Wonderboom South, Gezina
67 Mamelodi (13) (North)
Southern Region
Ward Suburbs
48 Celtisdal, Kosmosdal, Monavoni, Olievenhoutbos, Peach Tree
57 Lyttelton Manor, Monument Park, Doringkloof
61 Claudius, Erasmia, Laudium
64 Highveld, Irene, Louwlardia, Rooihuiskraal, The Reeds
65 Pierre van Ryneveld, Cornwall Hill
66 Glen Lauriston, Valhalla
69 Clubview, Hennopspark, Zwartkop, Centurion
70 Eldoraigne, Rooihuiskraal Noord, Sunderland Ridge, Wierda Park
11
2.2.2 Household sample plan design The sample plan design for the household survey is explained below in terms of the
sampling methodology, sample size and sample selection procedure.
2.2.2.1 Sampling methodology In an attempt to facilitate the analysis on a ward level, a sample plan was designed to
meet this endeavour. This required a multistage sampling approach by firstly listing
the size of the population according to the five CoT regions. This allowed for a
representative sample approach whereby the sample sizes were allotted to each
region in proportion to population numbers. The regional population data, as supplied
by the CoT was used to construct the sample plan. Once the regional sample sizes
were proportionately distributed to the total sample of 3 000, the population ward
information provided by the CoT, was used to further distribute the sample among the
wards within each region. For this purpose, a judgemental sampling procedure, guided
by population numbers and the geographic distribution of the ward population, was
applied. Geographic ward maps, as supplied by the CoT, were used in awarding final
sample sizes for the selected wards.
Within each region the following variables were also taken into account when
allocating the final sample sizes to the four identified dwelling types (formal township,
informal settlement, suburb and inner city):
(a) Dwelling typology to distinguish between formal and informal dwellings.
(b) Landline telephone ownership to facilitate the application of both telephone and
face‐to‐face interviews. Telephone interviews were conducted in suburbs and
the inner city, and face‐to‐face interviews in township areas and informal
settlements.
(c) Sample representivity, implying that a minimum of 30 questionnaires were
completed in each ward per region to support statistical analyses.
12
Finally, in order to judge the representation of the sample sizes by region and
settlement type according to the above criteria, the 2001 adjusted census figures of
Stats SA were used.
2.2.2.2 Sample size
Table 2.2 reflects the sample size of 3 000 questionnaires by region. The sample size
per region varies between 960 for the North Western Region to 290 in the Southern
Region. Table 2.2 also reflects the distribution of the sample surveyed by telephone
and personal face‐to‐face interviews.
TABLE 2.2
SAMPLE SIZE BY WARD AND ADMINISTRATIVE REGION
Region Number of wards
Sample size Face‐to‐face Telephone Total
Central Western Region 14 205 320 525 North Western Region 23 890 70 960 North Eastern Region 10 315 115 430Southern Region 8 0 290 290 Eastern Region 21 390 405 795 Total 76 1 800 1 200 3 000
It is clear from table 2.2 that the final sample consisted of 1 200 telephone and 1 800
personal face‐to‐face interviews. The final number of questionnaires returned by
region is reflected in table 2.3.
2.2.2.3 Sample selection procedure
As indicated in table 2.2, the sampling plan was designed to support both telephone
and personal face‐to‐face interviews. For the telephonic interviews, the Pretoria
telephone directory was used as sampling frame to judgementally select households
primarily residing in suburban and inner city areas. The sample selection procedure for
the telephone interviews was designed as follows:
13
(a) Survey areas
Firstly, the survey area (suburb) was identified (see table 2.1) in the telephone
directory. Once the survey area had been selected, the sample unit (household)
and sample element (respondents) was chosen. This selection procedure is
outlined below.
(b) Sample unit selection
Any household (sample unit) listed in the Pretoria telephone directory qualified
for inclusion in the sample. As mentioned, the selection of the survey area as
main sample selection criterion allowed for a representative inclusion of sample
units across the Tshwane area.
(c) Sample element (respondent) selection
The heads of households qualified as the sample elements or respondents for the
survey. To qualify for selection as a respondent a resident should have resided at
the household address for at least six months. Heads of households that were
not available/willing for interviewing could be substituted with his/her spouse.
(d) Time of interviewing
The interviews among households were concluded during the months of January
and February 2011. The interviewer schedule also secured an equal spread of
interviews conducted during day time, evenings and weekends.
Whereas telephone interviews were concluded with households residing in surburbs
and the inner city of Tshwane, personal face‐to‐face interviews were concluded with
households residing in formal township areas and informal settlements. More
specifically, the following selection rules governed the personal face‐to‐face
interviews:
14
• Formal dwellings were selected according to street names and numbers. Any
house in a street with a house number divisible by 10 (eg 20, 170 or 220) was
selected. Only one house per street was eligible for selection. For flats, cluster
homes, semi‐detached homes or duets, any flat, home or duet number divisible
by 10 was eligible for selection.
• For informal dwellings the selection of households was according to stand
number. Any stand number divisible by 10 was selected. For informal
dwellings without stand numbers, the painted census numbers were used. If
there were no numbers, fieldworkers could select any house in the settlement
– followed by every 10th household. The address, as stipulated on the
questionnaire, had to be such that a return visit to the informal dwelling could
be made for control purposes. No more than one respondent per household
was interviewed.
• Furthermore, if there were more than one but fewer than 10 dwellings on a
stand, the ‘A’ unit was selected.
• Substitution. Should the selected household be unavailable (at least twice) or
unwilling to participate, a cluster sample approach was followed. This
procedure allows for the selection of a ‘next door neighbour’ household. The
household located to the right or directly opposite the unavailable/unwilling
household was selected. Only one household per cluster and street was
allowed.
2.2.3 Data collection methodology As mentioned, households residing in suburbs and the inner city were interviewed
telephonically. Security measures applied by residents in these areas often make
personal face‐to‐face interviews difficult to conduct. In turn, the survey of households
residing in formal township areas and informal settlements was conducted through
personal in‐home face‐to‐face interviews. This data collection method was preferred
to prevent the potential of duplicating sample unit selection by area (suburb, inner
city, formal township and informal settlements) and due to the fact that inhabitants of
15
informal settlement areas in particular often do not have access to in‐home landlines.
Ultimately, the latter sample cluster group would not have been eligible for inclusion
should the telephone interview method, using the local telephone directory as
sampling frame, have been solely applied.
2.2.4 Research instrument
Following the first collaboration between the Customer Relations Management
Division of the CoT (who was delegated to oversee the 2009 customer satisfaction
survey as integral part of the City of Tshwane/Unisa MOA) and the BMR, previous
customer satisfaction survey instruments used by the CoT were supplied to the BMR
for possible inclusion in the final 2009 survey instrument. These survey instruments
were used as input to finally design the 2009 and 2011 survey instruments. In fact,
previous BMR survey models designed to measure customer satisfaction with
municipal services were extensively adjusted although the survey scope from previous
studies (2006 and 2007) largely corresponded with the scope of the 2009 and 2011
survey.
Also, it should be noted that the final research instrument for the 2009 study was
developed after a series of workshops in early March 2009. These workshops were
facilitated by the Customer Relations Management Division of the CoT. Separate
workshops were concluded for the household and foreign embassy (first workshop)
and business (second workshop) survey questionnaires. After presenting an initial
draft survey instrument to the CoT task team prior to the workshops, and following
some adjustments based on the CoT inputs received after the workshops, the 2009
survey instrument was finalised and approved by the Customer Relations Management
Division of the CoT.
The 2011 survey instrument closely resembled the 2009 questionnaire to facilitate the
construction of longitudinal trends. Only one question on democratic accountability
was added to the 2011 instrument to enquire on household comprehension of the
16
2011 local government elections. As was the case in 2009, the 2011 instrument was
workshopped and distributed for final approval.
The questionnaires were closely aligned to the aims of the survey (as set out in the
CoT/Unisa MOA and more confined BMR research proposal), namely to measure
customer satisfaction levels regarding broad service issues without going into detail
with regard to each function/department within the CoT. The following main issues
were addressed in the questionnaires:
(a) General perceptions of service delivery of the CoT and its ability to provide
people with a good quality of life as well as changes in service performance
delivery during the past 12 months.
(b) Identification of major challenges and priorities to be attended to during the next
12 months.
(c) Assessment of core household service delivery such as electricity, refuse
collection/waste removal, neighbourhood roads, sanitation/waste water/
sewerage, stormwater drainage/flooding, water provision and street/public
lights.
(d) Assessment of ownership of and satisfaction with pre‐paid electricity and water
meters as well as satisfaction with the installation of pre‐paid meters and
availability of vending points.
(e) Assessment of municipal tariffs/rates with specific reference to electricity tariffs,
pre‐paid electricity tariffs, property rates, water tariffs, pre‐paid water tariffs,
refuse removal, sanitation tariffs, waste removal tariffs and bus tariffs.
(f) Assessment of satisfaction with 30 municipal community service delivery items
such as community halls/recreation centres, fire and rescue services/fire brigade,
emergency medical and ambulance services, municipal bus service, municipal
cemeteries, municipal clinics, municipal hospitals, municipal museums (eg
Pretoria Arts Museum), municipal theatres, Pretoria National Zoo, parks,
pavements/pedestrian walkways, public libraries, public sports facilities,
17
municipal public toilets, garden refuse sites, municipal taxi ranks, informal
trading facilities, public swimming pools, recreation resort/nature conservation
area (reserves), licensing: Learner’s driver’s licence and driver’s licence,
applications/renewals), waste removal, emergency/disaster management, street
sweeping and litter control, traffic lights/signals, grass cutting, road maintenance
(repairs, signage, markings, speed bumps), street trees, government/social
housing (low cost housing). The utilisation levels of each of these 30 service
types were also measured.
(g) Assessment of public safety and by‐law enforcement.
(h) Assessment of billing and payment issues.
(i) Assessment of Customer Care services.
(j) Assessment of involvement in metro consultative and participatory processes.
(k) Assessment of communication preferences.
(l) Perceived levels of corruption in the CoT.
(m) Perceptions of government (Batho Pele) principles.
(n) Assessment of city developments.
(o) Assessment of governance with specific reference to the 2011 local government
elections as well as the level of access to and communication with councillors.
The questions on democratic accountability were primarily aimed at determining
local opinion on the roles of the IEC and the CoT.
2.2.5 Fieldwork
The fieldwork was conducted by several well‐trained fieldworker teams. The majority
of personal face‐to‐face interviewers owed their selection to the fact that they lived in
the survey areas and/or to their experience and continuous involvement in fieldwork.
The need to finalise the fieldwork by the end of February 2011 resulted in a
disproportionally high utilisation of experienced fieldworkers. Strict interviewer
control was exercised and 20% checkbacks were conducted.
18
To support fieldworkers in their task, they were all provided with a detailed training
manual highlighting, inter alia, the following:
• aim and background of the study
• guidelines on conducting the interview
• sampling procedure
• sample size
• coding manual for open‐ended responses
Completed questionnaires returned by fieldworkers were checked by field managers
appointed for the study. Questionnaires not complying with sample requirements
were either returned by the field managers to the fieldworkers for correction or,
where possible, were followed up telephonically. Once the questionnaires had been
completed satisfactorily they were handed to the BMR for central editing. This process
entailed a thorough scrutiny of the questionnaires by the BMR to determine whether
the data recorded were reliable. This was followed by data coding, capturing, cleaning
and storing.
2.2.6 Number of questionnaires returned
Table 2.3 shows the number of questionnaires returned by region in 2011. A total of
3 007 questionnaires were returned, which is virtually equal to the intended sample of
3 000 as outlined in table 2.2.
19
TABLE 2.3
NUMBER OF HOUSEHOLD QUESTIONNAIRES RETURNED BY REGION, 2011
Region Initial sample Realised sample
Total % Total % Central Western Region 525 17.5 517 17.2 North Western Region 960 32.0 969 32.2 North Eastern Region 430 14.3 436 14.5 Eastern Region 795 26.5 792 26.3 Southern Region 290 9.7 293 9.7 TOTAL 3 000 100.0 3 007 100.0
It is clear from table 2.3 that the intended and realised sample sizes per region are very
similar. A maximum of 0.3 % deviation is evident between the initial sample and
realised sample.
2.2.7 Coding, data capturing and tabulation
The BMR handled the data coding, capturing and verification process. All information
captured was in the form of a dataset that was cleaned prior to and during the data
tabulation process. The data were captured and analysed in SPSS format.
2.2.8 Validity of results
Any sample survey is subject to error, and as such, yields useful estimates but no
precise values. The most common errors, namely sample errors, interviewer errors,
and reporting errors are discussed below.
2.2.8.1 Sampling errors
Sampling error occurs when the sample selected is not perfectly representative of the
population. To avoid or minimise administrative sampling errors that relate to
problems in the administration or execution of the sample, Stats SA data were used to
avoid any possible underrepresentation of the population by region and type of area.
The care with which the sample was designed and executed largely ruled out
administrative sampling errors.
20
2.2.8.2 Interviewer errors
Three types of errors can occur as a result of an interviewer’s behaviour, namely errors
when asking questions, errors when recording answers and errors due to cheating.
The level of experience of fieldworkers largely eliminated the first two types of errors.
During checkbacks no cheating was detected.
2.2.8.3 Reporting errors Reporting errors usually stem from memory errors, misunderstanding of questions or
reluctance to answer them. The need to minimise reporting errors in the survey was
borne in mind when constructing the questionnaire and with the selection and training
of interviewers. The questionnaire was also successfully applied in the previous two
CoT surveys.
2.3 BUSINESS SURVEY RESEARCH METHODOLOGY
The level of satisfaction regarding CoT service provision from a business perspective
was also included in the study. This section highlights the survey population, survey
methodology, sampling and number of questionnaires completed with regard to the
business survey. Other methodological aspects of the business survey such as the
development of the research instrument, organisation of the fieldwork, coding and
data capturing are similar to the household survey and will not be repeated.
Furthermore, with the nonprobability sampling method approach used in the case of
the business survey, it was not possible to compute sample errors for the business
survey results.
2.3.1 Sample plan design 2.3.1.1 Business sample population The business sample population is described below in terms of sample units, elements
and survey area.
21
2.3.1.1.1 Sample units
The sample population for the business survey comprised all businesses operating in
the jurisdiction of the CoT. The following two subbusiness populations were included:
(a) Formal businesses
(b) Businesses (mainly informal) established in townships and informal settlements.
(Informal businesses are defined as businesses not registered for VAT purposes.)
For the purpose of the study informal businesses were defined to included categories
of the following:
(i) Small township retailers (or general dealers) and stand‐alone business with a
brick and mortar superstructure often located in a demarcated business area but
also located in the residential areas of townships. These businesses carry a wider
product range than spaza/tuck shops (see definition below) and have more
fixtures often allowing for self‐service to clients. Most important products sold
include soft drinks, bread, sugar, maize meal, cigarettes and tobacco. These
businesses employ fewer than 10 employees on a full‐ or part‐time basis.
(ii) Spaza/tuck shops defined as businesses operating in a section of an occupied
residential home or in any other structure on a stand zoned or used for
residential purposes and where people live permanently. (Small retail businesses
which operate from a residential stand or home). These businesses typically sell
soft drinks, bread, cigarettes and tobacco, milk, sweets and chocolates.
(iii) Shebeens (or taverns) defined as informal and unlicensed liquor traders who
primarily sell liquor to customers in townships.
It should be noted that the informal businesses defined above are businesses not
registered for VAT purposes. It should also be noted that survivalist businesses such as
hawkers were excluded. As most informal businesses operate in the retail industry
22
more informal businesses were anticipated to be included in the final sample.
However, attempts were made to sample informal businesses operating especially in
the manufacturing, construction, business service and transport and communication
industries. Exhibit 2.1 provides an overview of the type of informal business included
in the sample by sector.
EXHIBIT 2.1
INFORMAL BUSINESSES BY SECTOR TYPE ______________________________________________________________Retail:
• Food and vegetables • Cool drinks, cigarettes and sweets • Clothing • Retail trade in second‐hand goods in stores • Sale of motor vehicles • Sale of motor vehicle parts and accessories • Alcoholic beverages (shebeens/taverns)
Manufacturing:
• Food (cooked prepared food) • Bricks • Clothing, textiles and leather goods • Furniture • Brushes and brooms • Manufacture of games and toys • Number plates, signs • Engraving • Jewellery
Construction:
• Construction of buildings (construction of buildings such as houses, flats, farm buildings and industrial and commercial buildings)
• Plumbing, electrical contracting and shopfitting (assembly and/or installation of equipment, such as counters, shelves, cupboards and shop fronts, on the premises of the client in buildings).
• Plumbing (laying of sewerage pipes, the installation of water pipes, wash basins, baths, water heating systems, solar heating systems, sprinkler systems and gutters and sheet metal work in all structures)
23
• Electrical contracting (installation of electrical wiring and lighting in buildings) installation of alarm systems
• Building completion (glazing, painting and decorating, floor and wall tiling or covering with other materials such as parquet, carpets, wallpaper, etc.; floor sanding, finish carpentry, acoustical work, cleaning of the exterior, etc.)
Transport and Communication
• Travel agency and related activities • Telecommunications
Business Services:
• Maintenance and repair of motor vehicles • Repair of personal and household goods • Bed and Breakfast and Guest House Accommodation • Real estate activities • Legal, accounting, bookkeeping and auditing activities • Computer and related activities • Photographic activities • Money lending (Micro lender) • Crèche, motor vehicle driving schools/tutors and music, dancing and other art
schools • Car wash • Hairdressing and other beauty treatment • Funeral and related activities • Washing and (dry‐) cleaning of textiles and fur products • Investigation and security activities
Personal Services:
• Hair salons • Phone shops
__________________________________________________________
Only businesses where the owner/manager could express an informed opinion on CoT
service delivery were included in the survey. Survivalist businesses such as, for
example, hawkers utilising municipal services to a limited extent were therefore
excluded from the sample population.
24
The lack of business registers containing businesses by size class (employment or
turnover) and economic sector largely ruled out the possibility of selecting businesses
by size class or sector.
2.3.1.1.2 Sample elements Interviews were conducted with the manager or owner of smaller businesses and
where possible, with the owner or a senior manager of larger corporations.
2.3.1.1.3 Survey areas All businesses established in the area of jurisdiction of the CoT were eligible for
inclusion in the sample. The bold entries in the Pretoria telephone directory
constituted the sample population for telephone interviews with formal businesses.
Although the businesses included in the sample operated across all five regions of the
CoT, a representative sample per region was not considered, as with the household
survey.
The Pretoria telephone directory excludes the listing of informal businesses.
Consequently, township areas and informal settlements were used as demarcated
areas for sampling informal businesses. Exhibit 2.2 shows the selected regions/areas
for interviews with informal business owners/managers.
25
EXHIBIT 2.2
REGIONS/AREAS SELECTED FOR INTERVIEWING OF INFORMAL BUSINESSES
Region Area Central Western Region Atteridgeville (Units 3,6 7,17, 16,1825‐29,) and Saulsville North Eastern Region Dilopye (Temba), Suurman; New Eersterust; Stinkwater; Tswaing Village; Zoutpan;
Hammanskraal (Units 1‐2) and Hammanskraal West; Marokolong; Ramotse; Mandela Village; Babelegi; Kudube (Units 1‐3, 6, 7, 10‐11 & D); Majaneng & Mashemong
Eastern Region Mamelodi (North, East, West and South and Moretele View) and Mamelodi (Ext 1‐8, 10‐15, 20, 22 & 24) & Mahube Valley
North Western Region Winterveld & Winterveld AH; Mabopane (Ext A,B,E,N, R,S,T,U,V X); Kopanong;Soshanguve (Ext AA, BB, CC, DD, EE, FF, GG, HH, IA, JJ, KK, LL, NN, PP, SS, UU, WW,YY,ZZ& F,G,H,K,L,M,T, P,R,S,V,W,X,Y & Unis 1,2,3); Soshanguve South (Units 4,5,14), Soshanguve East & Ga‐Rankuwa (Units 1‐10,15‐17 & 21‐25)
2.3.1.2 Sampling methodology
For the formal business survey, a systematic random sample was drawn mainly from
the business entries contained in the Pretoria telephone directory. However, this
database only contained contact names and telephone numbers, which largely ruled
out sample selection by sector, region and size. Comparisons according to these strata
should therefore be exercised with caution.
Due to the lack of a comprehensive list of informal businesses operating in the CoT, a
judgemental sampling procedure was used to sample informal businesses according to
the regions outlined in table 2.5(b).
2.3.1.3 Sample size
Tables 2.4 reflect the distribution of the sample of 755 businesses. A total of 519
formal and 256 informal businesses were included in the sample. The sample
distribution of the 256 informal businesses by area is proportional to the size of the
population in each development region based on the population census data (see
exhibit 2.3).
26
TABLE 2.4
SAMPLE DISTRIBUTION BY FORMAL AND INFORMAL BUSINESSES
Type of business n %
Large/medium formal business 185 23.9
Small formal business 334 43.1
Informal business 256 33.0
Total 775 100.0
2.3.1.4 Business survey data collection methodology
Whereas the owners/managers of formal businesses were interviewed telephonically,
interviews with informal business were conducted face‐to‐face. Interviews were
conducted in February 2011 during business hours and during weekends.
2.4 EMBASSY SURVEY RESEARCH METHODOLOGY
This section highlights the survey and sampling designs and data collection
methodology with regard to the embassy survey. Other methodological aspects of the
embassy survey such as the development of the research instrument, organisation of
the fieldwork, coding and data capturing are similar to the discussions presented as
part of the household/business surveys and will not be replicated.
2.4.1 Research design for embassy survey The household survey questionnaire was used to interview a random sample of 20
embassies in the CoT.
2.4.2 Sampling plan for embassy survey As mentioned, a total of 20 embassies were sampled for survey purposes. In this
regard the list of Embassies and High Commissions including Consulates and Liaison
Offices in Tshwane, as published by the Department of Foreign Affairs
27
(www.dfa.gov.za), were used as sample frame. The CoT approved the sample frame
prior to sampling 20 embassies at random.
2.4.3 Data collection method for embassy survey All sampled embassies were interviewed telephonically.
2.5 FOCUS GROUP DISCUSSIONS
A total of 10 focus group discussions were conducted in 10 wards. The 10 wards
selected were those showing the most negative service perceptions in the CoT. The
seven core municipal services were used as a basis for selecting the wards with the
lowest mean ratings for these services. The selected wards are contained in table 2.5.
The table also shows the mean rating (out of 10) for the following core services:
electricity, refuse collection, neighbourhood roads, sanitation/sewerage, stormwater
drainage, street/public lights and water provision.
TABLE 2.5
CoT WARDS SELECTED FOR FOCUS GROUP DISCUSSIONS, 2011
Ward number Region Mean rating for core services (out of 10)
8 North East 3.21 25 North Western 3.80 13 North East 4.26 14 North East 4.95 64 Southern 5.31 9 North Western 5.48 7 Central Western 5.50 73 North East 5.53 21 North Western 5.60 3 Central Western 5.61
The research design made provision for focus group discussions. However, the local
government elections (of 18 May 2011) impacted negatively on the institutional
28
arrangements such as securing the cooperation of ward committees (initially targeted
for the focus group discussions). It proved to be problematic to arrange focus group
discussions (with eight to 10 persons) within the selected wards with the assistance of
community‐based institutions. As an alternative, in‐depth discussions with community
leaders in the selected wards were approved by the CoT.
2.6 CONCLUDING REMARKS
This chapter provided an overview of the research methodology for the 2011 Customer
Satisfaction Survey conducted among 3 000 households, 750 businesses and 20
embassies. The outcomes of the research findings are presented as follows in each of
the consecutive chapters:
• Chapter 3: Household satisfaction survey findings
• Chapter 4: Business satisfaction survey findings
• Chapter 5: Embassy survey findings
• Chapter 6: Summary and recommendations
Prior to a discussion of the outcomes of the various survey results, it is important to
note that the report provides a longitudinal analysis for 2009 and 2011, a major benefit
associated with regular satisfaction surveys. Trend data facilitate the development of a
service performance model for the CoT aimed at guiding strategic planning in terms of
either:
• maintaining the service performance levels of those departments/units that
reflect high levels of satisfaction with service quality
• improving on the service delivery of under‐performing departments/units.
A prerequisite for constructing trend data is the standardisation of the research model
over time. Of particular importance in this regard are the following:
29
• The same satisfaction variables should be measured in repetitive surveys.
• The same rating scale should be applied in repetitive surveys. The latter should
preferably be a 10‐point numerical scale that is sensitive enough to detect even
small changes in service delivery performance.
• The same sampling plan design should be maintained to ensure that variations in
service satisfaction ratings originate from actual changes in service delivery
performance and not from research methodological adjustments.
After taking these criteria into account, a critical comparison of the 2009 and 2011 CoT
satisfaction surveys with the 2007 and 2006 CoT surveys was conducted.
Unfortunately, the outcome of this investigation largely proved that the 2009 and 2011
and previous studies are wholly incompatible for constructing longitudinal trend
analysis. The quantitative incompatibility between the research instruments with
regard to electricity can be used to illustrate the problems with regard to constructing
longitudinal trends:
(a) The 2009 and 2011 research instrument measures the following satisfaction
levels on a 10‐point numerical scale: electricity provision, prepaid meters with
regard to installation/vending points/prepaid meters, control of illegal electricity
connections, billing (including electricity) with regard to the
clarity/correctness/regularity of CoT accounts, methods of payment, staff
attitudes at payment points and Customer Care (including electricity enquiries).
In all the above cases, reasons for any dissatisfaction were requested.
(b) The 2007 research instrument measured the following on a 3‐point (often,
sometimes, never) or 5‐point (ranging from very poor to excellent) verbal scale:
response of municipality to electricity problems, frequency of experiencing
electricity problems, frequency of experiencing meter readings, frequency of
experiencing problems with electricity account. None of these relate directly to
the measurement of service satisfaction.
30
The above clearly shows that different variables with regard to electricity provision as
well as different rating scales were used by the 2009 and 2011 studies compared to
previous studies. Furthermore, the two sets of surveys (2006/07 and 2009/11) also
show notable differentiation in sample sizes, namely 3 000 households and 750
businesses in the 2009 and 2011 surveys against 800 households and 400 businesses in
the 2006 and 2007 surveys. This implies that the 2009 and 2011 model featured an
additional advantage of securing regional analysis that was overlooked in previous CoT
customer satisfaction surveys. Clearly the 2009 and 2011 studies present much richer
information in support of aggregate and regional analysis.
Against this background the following recommendations with regard to the research
methodology for future studies are submitted:
(a) Consistency in the sampling methodology approach to ensure proportionate
representation by region and subregion. This will not only allow validation of
survey outcomes with official census data but could also support operational
planning by the CoT in terms of identifying service delivery priorities.
(b) Consistency in maintaining the core research variables used to measure
satisfaction levels. These core measures include:
• core services
• community/business services
• public safety and by‐law enforcement services
• billing, payment and Customer Care
• communication and interaction
• city developments
31
(c) Similarity in measurement scales used to capture satisfaction ratings. As
indicated, the 2009 and 2011 studies used a 10‐point measurement scale that
allows for more variability that supports longitudinal analysis.
The limitation highlighted above, however, did not preclude a qualitative comparison
by, for example, identifying all the services that require improvement by the CoT (eg
improving conditions of roads, create awareness of the CoT Website, etc). The final
chapter presents a more confined focus in this regard.