What a good diet does for body and mind,

So does a healthy data set for analysis and model building!

- Abdi-Khalil Edriss

TYPES OF DATA
I. Data
The purpose of this chapter is to discuss the most common classifications of data, and
to describe when and how each type of data is generally helpful in identifying analysis
techniques. In addition, a thorough discussion on how to design a questionnaire, when
to use open-ended and close-ended questions for data collection, and some insights
into how to plan a survey are discussed in this chapter.

2.1. What are data1?

Data refer to the available raw information gathered through interviews,

questionnaires, observations, or secondary databases. By organizing the
data in some fashion, analyzing them, and making sense of the results, we
may find answers to questions we seek to address.

2.2. Is it true that the type of the data determine the type of
statistical or econometric model?

Yes. The data set usually speaks by itself. We have to be careful what type
of model (static or time series models) we should apply or fit to what type
of data set. We classify data into two major types. Namely, cross-sectional

1
Data is a plural form of datum
~ 21 ~
 (or spot data), time-series (or longitudinal) and panel (hybrid of cross and
time-series) data sets.

i. Cross-sectional data
o Usually contain independent observations
o Exclude time factors or contains no element of time factors, and
hence is named spot data
o Are analyzed through static models such as regression models,
qualitative models, simultaneous models, etc. (Refer to chapter
eight for more).

ii. Time series data

o Usually contain inter-dependent observations
o Includes time factors or patterns
o Are analyzed through time series models such as Autoregressive
Integrated Moving Average (ARIMA), Autoregressive Moving
Average (ARMA), lag or dynamic models for time series data
analysis. (Refer to chapter ten for details).

iii. Panel data

o Hybrid of cross-sectional and time-series
o Repeated data collection on the same observation over similar time
period
o Are analyzed using panel data models such as fixed and random
effect models

II. Methods of Data Collection

There are different ways of collecting data. Some of these include face-to-face
interview of the respondents, questionnaires mailed to the respondents, interview on
the telephone, and electronically distributed questionnaires or on-line interviews.

Generally, interviews could be unstructured or structured, and conducted either face-

to-face or by telephone or on-line. However, face-to-face interview is the more
realistic of way collecting information in most developing countries due to the
inaccessibility (sometimes infeasibility) of telephone and computer (on-line) services
to the majority of the population. Thus, the discussion will focus only on face-to-face
interviews using structured (or closed) or unstructured (or open-ended) questionnaire
in collecting the information on the issues of research interest.

2.3. What is a questionnaire?

A questionnaire is a pre-formulated written set of questions to which

respondents record their answers, usually within rather closely defined
~ 22 ~
 alternatives (hence the name structured questionnaire). Questionnaires are
an efficient data collection mechanism when the researcher knows exactly
what is required and how to measure the variables of interest.
Questionnaires can be administered personally or mailed to the
respondents, or electronically distributed.

Basically, designing a questionnaire and filling it properly is the beginning of a

research trip. It is like filling an application form on your way to obtain your
legitimate passport for a trip to research jungle. This is where you start, and make sure
you start correctly so that you can collect a healthy data set that enables you do a
healthy data analysis, and reach to quality results, discussion, as well as draw proper
conclusion and recommendations.

2.4. What are the requirements of a good questionnaire?

A good questionnaire is complete and concise, reliable, valid and

consistent.

i. Complete and Concise – a good questionnaire seeks to obtain the

information required to meet all survey objectives with as few questions
as possible.

ii. Reliable – a good questionnaire will provide interviewers with reliable

data. Reliability indicates that the same response is obtained regardless
of who asks the question and where the question is asked. The
differences between the interviewers are usually few if reliable
questionnaires are used, especially after training the interviews before
the survey begins.

iii. Valid – a good questionnaire provides valid data by ensuring that the
respondent understands what information is being sought. Validity
implies that the question elicits a true and accurate response that
measures what you are interested in measuring.

iv. Consistent – a verbatim questionnaire outlines questions, exactly as they

are to be asked. The questions must be translated in a way that ensures
both the interviewer and the respondent interpret their meaning the same
way, and thus have the same understanding about what information is
asked for. In this way, inconsistency between interviewers is reduced.
Consistency can also be achieved by good training in the use of the
questionnaires, and following especial instructions designed for the
survey.

~ 23 ~
2.5. What the main criteria for a good questionnaire?

The criteria for a good questionnaire are -

Does it provide the necessary decision-making information?
o The user of the information must approve the questionnaires. If the
questionnaire fails to provide the necessary insights, then discard or
revise
Does it consider the respondent?
Intended respondents
Design of questionnaire
If the respondents find the questions hard to answer, the quality of
the responses will be low

2.6. How do we design a good questionnaire?

Questionnaire design processes are summarized as follows (Courtesy of Professor

Frode Alfnes, Norwegian University of Life Sciences, Norway) –

Step 1: Determine survey objectives, resources, and constraints

o What data are needed?
o If you cannot answer this question => Exploratory design (Focus
groups, interviews etc.)

Step 2: Determine the data collection method

o Face-to-face, internet, telephone, mail, self-administration,
o Limitation: Resources (money, people, skills), requirements of the
respondent, quality of the data, sampling precision, length of
questionnaire, etc.

Step 3: Determine the question response format

o Open-ended questions
o Close-ended questions
o Dichotomous question
o Multiple-choice questions
o Scaled-response questions

Step 4: Decide the question wording

o The wording must be clear
o Avoid biasing the respondent
o Consider the respondent‘s ability to answer the question
o Consider the respondent‘s willingness to answer the question

Step 5: Establish questionnaire flow and layout

o Screener questions
o Begin with a question that obtains a respondent‘s interest
~ 24 ~
 o Ask general questions first.
o Ask questions that require ―work‖ in the middle
o Insert ―Prompters‖ at strategic points
o Position sensitive, threatening, and demographic questions at the
end.
o Allow plenty of space for open-ended responses.
o Put instructions in capital letters.
o Use a proper introduction and closing.

Step 6: Evaluate the questionnaire and layout

o Are all the questions necessary?
o Is the questionnaire too long?
o Will the questions provide the information needed to accomplish
the research objectives?

Step 7: Obtain approval of all relevant parties

o Distribute copies to all parties with authority
o Obtain managerial approval

Step 8: Pretest and revise

o Pretest — trial run of questionnaire
o Revise — do a second pretest
o Focus groups
o Managerial approval — re-obtain

Step 9: Prepare final questionnaire copy

o Precise typing, instructions, spacing, numbering, and pre-coding
must be set up, monitored, and proofread.

Step 10: Implementing the survey

o Supervisor‘s instructions
o Other record sheets
o Field management companies

One objective of any survey design is to obtain accurate information by

minimizing the non-sampling errors that may occur. The following are some of
examples and guidelines to construct an excellent survey questionnaire

i. Question Ordering – attitude toward a question in a survey is very

often set, or changed, by preceding question that bear on the same
topic. Order is important in the relative positioning of specific versus
general question. For example, examine the following question:
a. Will you support an increase in government taxes?
b. Will you support an increase in government taxes for
education?

~ 25 ~
Now it would not be unusual to find more people supporting B if asked in the order
(B,A) than asked in the order (A,B).

o For many survey questions the order of the possible responses (or
choices) to a particular question is as important as the position of the
question on the questionnaire. For example, if a person being
interviewed is presented with a long list of possible choices, or if each
possible choice is wordy or difficult to interpret, a person is likely to
respond with the most recent choice (the last one on the list). But, if a
respondent must choose items from a long written list, then the items
appearing toward the top of the list have a selection advantage.

o Researchers attempting to design a questionnaire should be aware of

the common ordering problems for question and response. They should
attempt to counter potential difficulties by reconsidering the following
techniques.

Printing questionnaires with different orderings for different

subsets of the sample.

Using show cards or repeating the question as often as

necessary in an interview so that the question and possible
answers are clearly understood.

Carefully explaining the context in which a question was asked

in the analysis of the survey data.

ii. Open questions - are inquires in which the respondent is allowed to

freely state an unstructured answer. Here the respondents are allowed
to express some depth and shades of meaning in the answer. Example
of an open question is – what is the most important problem facing
your country today?

This type of question usually creates difficulties in analysis because the

unstructured answer cannot easily be quantified, and may be
impossible to compare across questionnaires.

iii. Closed Questions – by closed questions we mean that each question

will have either a single numerical answer (like marital status and age)
or fixed number of predetermined choices one of which is to be
selected by the respondent. Closed questions are preferred to open
questions for ease of data coding and analysis; however, open-ended
questions are often used in focus group discussion for qualitative
analysis purposes.

~ 26 ~
 Example of a closed question is – which of the following is the most important
problem facing your country? (Check one)
a. Crime
b. Unemployment
c. Inflation
d. Budget deficits
e. Drought

One may see that any closed form of question will limit the respondent‘s response,
and may force a respondent into an answer that would not necessarily be a first
choice. Or, the respondent‘s choices could be forced into predetermined categories.

POINTS TO PONDER
Generally, a good plan for designing a closed question with appropriate alternative is
to use a similar open question on a pretest or pilot survey; then choose as the fixed
alternatives those that most nearly represent the choices expressed in the open
answers.

Response Options – to force people to make decisions on questions they know

nothing about seems inappropriate. Thus a good questionnaire will provide
screening questions to determine whether the respondent has enough information
to form an opinion on certain issue. For example, questions on stricter
enforcement of speed limit laws for drivers, should be stated without a ―no
opinion option‖. Or, questions like ―Do you think the enforcement of traffic laws
in our city is too strict or too lenient? Here no middle ground is offered. The two
choice options force the person responding to think about the direction of the
response, but the interviewer should explain that various degrees of strictness or
leniency could be taken into account.

Wording of Questions – the designer should be concerned about the phrasing of

the main body of the questions. For example, questions such as ―Do you favor the
use of government vehicle for private use on the weekends?‖ should be asked in a
more balanced form such as ―Do you favor or oppose the use of government
vehicle for private use on the weekends?‖

POINTS TO PONDER
Training of field workers or interviewers can greatly reduce the duration of the
interview, as they become more familiar with the questionnaire. This training should
take place before the pre-test of the survey, at which time the length of time required
of each interview should be estimated. Due to distance, fieldwork in rural areas is
always much slower than in urban areas.

~ 27 ~
SAMPLE STRUCTURED QUESTIONNAIRE

QUESTIONNAIRES

The information that the questionnaire collected included:

Household Identification

Household Wealth, Production and Other tenure arrangements during the

1999/2000 Season

Information on Agronomic, Cropping Patterns, Marketing and Exchange

Groundnut Management Practices and Technology Adoption

The Dynamics of Groundnut Production, Profitability and Adoption of Technology in

Subsistence Economy: The Malawi Case

QUESTIONNAIRE – 1999/2000 SEASON

GENERAL INSTRUCTION: CIRCLE THE APPROPRIATE CODE, 1, 2, 3, 4 OR 5 AND

FILL IN THE BLANK SPACES, ACCORDINGLY.

A. HOUSEHOLD CHARACTERISTICS

Household Identification

Region _______________________________
ADD ________________________________
RDP _________________________________
EPA _________________________________
Village ________________________________
Household Number _______________________
Interviewer‘s name ________________________

A. IDENTIFICATION
Note- this part of the questionnaire should be posed to household head. Only if the head is
absent or is expected to be absent for the next two days should the question be posed to
another adult member of the household preferably the spouse of the household head.

A1. Name of the Interviewee ____________________________

A2. Sex 1. Female 2.Male
A3. Age of household head 1. 15-29 2. 30-49
3. 50-59 4. Above 60
A4. Marital status1. Never married 2. Married 3. Widow/Widower
4. Divorced 5. Separated
A5. Marriage system 1. Partlineal 2. Matrilineal 3. Neutral
A6. Education level 1. Primary 2. Secondary
3. Tertiary 4. None
A7. Household head 1. Female 2. Male

~ 28 ~
A8. How many children do you have? 1. One 2. Two 3. Three
4. More than 3
A9. How many people live in this household?
1. 1 2. 2-4 3. 5-6 4. More than 6

B. HOUSEHOLD WEALTH, PRODUCTION AND OTHER TENURE

ARRANGMENTS DURING THE 1999/2000 SEASONS

B.1. Do you have land for cultivation? 1. Yes 2. No if ―YES‖ what is the
holding size of your land? 1. Below 1 hectare 2. 1-2 ha 3. 2.1-5 hectare
4. 5.1-10 hectare 5. Over 10 hectare

B.2. Who owns the land? 1. Self 2. Government 3. Land owner/Estate

4. Village chief 5. Relative 6. Other.

B.3. Main crop grown 1. Maize 2. Groundnut 3. Pigeonpea

4. Tobacco 5. Others, specify ___________

B.4. Of the land cultivated, how much of it is allocated to groundnut production?

1. Less than 25% 2. 26% - 50% 3. 51%-75% 4. Over 75%

B.5. Why do you produce groundnut? For __________

1. Sale 2. Consumption 3. Both

B.6. Who has the Off-farm employment?

1. Father 2. Mother 3. Both 4.None

B.7. Of the income generated, what percentage goes to groundnut production?

1. Less than 25% 2. 25% - 59% 3. 50% - 75% 4. Over 75%

B.8. Where are the sources for the seeds?

1. Previous harvest 2. ADMARC 3. Government 4. NGOs 5. All

B.9. Do you get credit?

1. Yes 2. No If yes, how much did you receive? If No go to B13

B.10. Which organization gives you agricultural credit?

1.MUSSCCO 2. MRFC 3. Commercial Banks
4. Ministry of Agriculture 5. NGOs

B.11. Credit is given in what form?

1. Seed 2. Cash 3. Fertilizer 4. All

B.12. Who gets more credit packages?

1. Men 2. Women 3. Clubs 4. Estate Owners

B13. if you have not received credit, what are the reasons that you have not received credit?
1. Not willing 2. Not important 3. Not attending meetings 4. Unknown

B14. How much groundnut did you produce in 1999/00 season?

1. Less than 50 kg 2. 50-100 kg 3. 100-200kg 4. Over 200 kg

B15. What was your farm capital for 1999/00 season?

~ 29 ~
1. Less MK100 2. MK101-500 3. MK501-1000 4. MK1001-2000
5. Over MK2000

C. INFORMATION ON AGRONOMIC, CROPPING PATTERNS, MARKETING

AND EXCHANGE

C1. How long have you been growing groundnut?

1. 1-2 years 2. 3-5 years 3. 6-10 years 4. Over 10 years

C2. What do you grow groundnuts for?

1. Formal trading 2. Non-formal trading 3. Consumption 4. Other,
specify________________

C3. After harvesting, who is fully involved in the processing of the groundnuts?
1. Only father 2. Only mother 3. Children 4. All

C4. How much groundnut have you produced for 1999/2000 seasons? __________Kg, and
how much of these have you sold? _______________

C5. If the groundnuts or part of the groundnuts are for sale, who controls the income?
1. Father 2. Mother 3. Both

C6. Do you exchange the groundnuts with other crop when trading?
1. Yes 2. No
If ―YES‖, specify which crop(s) ____________________

C7. Do you recognize any nutritional importance of groundnuts?

1. Yes 2. No

C8. Which groundnuts seed multiplication program do you belong to?

1. ICRSAT 2. ACTIONAID 3. ADMARC 4. Others, specify _________

C9. Have you received any crop loans from any of the above programs?
1. Yes 2. No if ―YES‖, specify _______________

C10. If your groundnut is for sale, which market gives you higher prices for your groundnut
production?
1. Private traders 2. ADMARC 3. Local market 4. Fellow farmers

D. GROUNDNUT MANAGEMENT PRACTICES AND TECHNOLOGY ADOPTION

D1. Do you grow more than one groundnut variety? If ―No‖ go to question D5
1. Yes 2. No

D2. Which of the following variety have you grown most?

1. Chalimbana 2. Tchailosi 3. Kalisere 4. CG7

D3. Why do you prefer one variety to the other(s)?

1. High yielding 2. Market appeal 3. Early maturity 4. Taste
5. Drought resistance 6. Disease resistance

D4. Have you grown groundnuts for more than one season?

~ 30 ~
 1. Yes 2. No

D5. Which crop(s) do you grow with groundnut? (Multiple answers allowed)
1. Soya beans 2. Pigeonpea 3. Maize 4. Cowpea 5. None

D6. Have adopted any of the following technologies? (Multiple answers allowed)
1. Intercropping (mixed) 2. Ridging 3. Variety (Improved) crops 4. Rotation

D7. Which organization introduced you to the groundnut technology adoption?

1. ICRSAT 2. ACTIONAID 3. ADMARC 4. Others, specify _________

D8. When did you receive the first training in groundnut technology adoption?
1. Before 1990 2. 1990/3 seasons 3. 1994/6 seasons 4. 1997/99 seasons

D9. Do you know that planting groundnut improves soil fertility?

1. Yes 2. No if ―YES‖ go to D10

D10. Which organization introduced you to the information?

1. ICRSAT 2. ACTIONAID 3. ADMARC 4. Others, specify _________

D11. What is the spacing between the plants in a single row (ridge) that you use?
1. Less than 15 cm 2. 15-20 cm 3. 20-30 cm 4. Over 30 cm

D12. When do you weed your plots after planting groundnuts?

1. Before 30 days 2. 50 days after 3. 90 days after 4. 120 days after

D13. Do you know how to check for maturity?

1. Yes 2. No if ―Yes‖, what signs do you look for? _____________

D14. Where you contacted by Extension Officer for 1999/00 season?

1. Yes 2.No
If ―YES‖, how many times were you visited in the season?
1. Less than 5 2. 5-10 times 3. 11-20 4. 21-30 5. Over 30

D15. Did Field Assistants make you aware of the groundnut technologies?
1. Little 2. Much 3. Very much 4. None

D16. How did you benefit from the groundnut technologies? (Multiple answers allowed)
1. Yield increment 2. Income increment 3. Fertility increment 4. None

2.7. How do we collect data? Or, what are the methods?

Before assessing relationships among two or more variables, one must

measure them in some manner from a population. Among several data
collection methods, we focus mainly on three major types of data
collection methods. These are personal interviews, direct observations and
self-administered questionnaires.

~ 31 ~
 i. Personal Interview

o Requires an interviewer to ask prepared questions and to record the

respondent‘s answers.
o Primary advantage of these interviews is that people will usually
respond when confronted in person, and also the interviewer can
note specific reactions and eliminate misunderstandings about the
questions asked.

POINTS TO PONDER
Aside from cost involved, the major limitations of personal interview is that if the
interviewers are not thoroughly trained, they may deviate from the required protocol
and introduce a bias into the sampled data.

ii. Direct Observations

o Getting data from objective sources that are not affected by the
respondents themselves.
o May not involve measurement on peoples, but from other sources.
For example, laboratory experiments, from records, income
information, counting of objects, etc.

POINTS TO PONDER
The disadvantage of using an observer is the possibility of errors in observations; this
may include over-reporting, under-reporting or no reporting on particular or subtle
issues.

iii. Self-Administered Questionnaires

o Involves getting data from the respondent directly, that is, the
questionnaires are completed by the respondents themselves.

o The self-administered questionnaires do not require interviewers;

the questionnaires are supposed to be mailed or delivered to the
respondents to be completed by the respondents.

o The questionnaires must be constructed carefully if it is to

encourage participation of the respondents or interviewees.

POINTS TO PONDER
The disadvantage of using self-administered questionnaires is that non-response or
lower rate of response. It may introduce bias in the data; in addition, the respondents
may not be representative of the population of interest. To eliminate some of these
biases, investigators frequently contact the non-respondents through follow-up letters,
telephone interviews or personal interviews (though costly in many respects).

~ 32 ~
 III. A Check List for Planning a Survey

2.8. What are the major elements when planning a survey?

Generally, in planning a survey the following checklist should be carefully

considered.

 Objectives – state the objectives (general or specific objectives) of the survey

clearly and concisely. Keep them as simple as possible. Refer to those
objectives regularly as the design and the implementation of the survey
progress.

 Target Population – Carefully define the population to be sampled. For

example, if adults are to be sampled, then define what ‗adult‘ means and state
what groups of adults are over certain age.

 The Frame – select the frame (or frames) so that the list of sampling units and
the target population show close agreement. Keep in mind that multiple frames
may make the sampling more efficient. For example, residents of a city can be
sampled from a list of city blocks coupled with a list of residents within
blocks.

 Sample Design – choose the design of the sample, including the number of
sample elements, so that the sample provides sufficient information for the
objectives of the survey.

 Data collection Methods – decide on the method of measurement, usually

one or more of the following methods: personal interviews, mailed
questionnaires, telephone interviews, or direct observations.

 Selection and Training of Field-Workers (or enumerators) - carefully

select and train the field-workers. After the sampling plan is clearly and
completely set up, someone must collect the data. Those colleting data, the
field-workers, must be carefully taught what measurements to make and how
to make them. Training is especially important if interviews, either personal or
telephone, are used, because the rate of response and the accuracy of responses
are affected by the interviewer‘s personal style and tone of voice.

 The Pretest – select a small sample for a pretest. The pretest is crucial, since
it allows you to field-test the questionnaire or other measurement device, to
screen interviews, and to check on the management of field operations. The
results of the pretest usually suggest that some modifications must be made
before a full-scale sampling or actual survey is undertaken.
~ 33 ~
  Organization of Fieldwork – plan the fieldwork in detail. Any large-scale
survey involves numerous people working as interviewers, coordinators or
data managers. The various jobs should be carefully organized and lines of
authority clearly established before the survey is begun.

 Organization of Data Management – outline how each piece of data is to be

handled for all stages of the survey. Large surveys generate huge amounts of
data. Hence, a well-prepared data management plan is of the utmost
importance. This plan should include the steps for processing data from the
time a measurement is taken in the field until the final analysis is completed.
A quality control scheme should be included in the plan in order to check for
agreement between processed data and data gathered in the field.

 Data Analysis - outline the analyses (descriptive and/or advanced) that are to
be completed. Closely related to the previous note, detailed specifications of
what analyses are to be performed, models and the like. It also involves
discussion and interpretations of the results.

POINTS TO PONDER
Note that building a house without proper plan and architectural design would lead
to disastrous results with lots of resource wastage, and similarly a survey without
proper check list leads to unplanned research with no scientific procedures followed
resulting with biased and inconsistent results consisting high standard errors.

IV. Data Collection Plan

2.9. Submit a draft “Data Collection Plan” describing in detail how

you will collect the data for your research. Provide a sampling
plan identifying the sample population, sample frame, sample
size calculation, and sampling method, and explain precisely
why each question is included and what steps you have taken
to ensure the data collected will be meaningful, reliable and
valid.

DATA COLLECTION PLAN

The Sampling Population and Study Area

The study population is groundnut farmers in the Central Region of Malawi2. The
survey will be conducted in Lilongwe and Salima districts in the Central Region

2
Malawi is divided into three regions: Northern, Central and Southern Regions.
~ 34 ~
 of the country. These two districts lie within Lilongwe Agricultural Development
Division and Salima Agricultural Development Division (ADD)3, which accounts
for 70% of groundnut production in Malawi (Ministry of Agriculture, 1998).
Lilongwe ADD is situated at an altitude of about 600 above sea level while Salima
ADD lies on the lakeshore flood plain at about 200 meters above sea level. The
Central region has a warm to hot weather and cloudy with light to heavy rains,
rainfall ranges from 600-1000 mm per annum, falling in one rainy season from
November to March. This type of rainfall supports crops, such as groundnut,
tobacco and maize that are planted early in the growing season.

The choice of the two districts is necessitated by the need to cover as many
diverse factors as possible that might affect the household‘s decision to grow
groundnuts. These factors are income levels, input and output prices, access to
land and socio-cultural factors related to labour transactions within family groups.

Sampling Design and Instruments

The cross-sectional data will be collected in the form of:

1. Questionnaires - structured questionnaires are used to collect primary

data through interviews with selected households.

2. A checklist is used to collect more information from some key

informants on policy and other issues regarding to the grain legume
sub-sector in Malawi, particularly marketing, before and after market
liberalization. Secondary data will also be collected from Non-
Governmental Organizations (NGOs) that participate in the grain
legume sub-sector research. Some of the major organizations are --
Concern Universal, Action-Aid, World Vision and Self-Help
Development International, ICRISAT and University of Malawi.

3. Participatory Appraisal Methods – this is to collect additional data

from key informants and focus-group discussions. It is used to learn
the decision-making processes and outcomes within the households
and findings from the field surveys that are conducted regularly by
Agricultural Development Divisions (ADDs) will supplement this
study.

3
Malawi is divided into eight ADDs that form different agro-ecological zones. These ADDs lie within the three regions of the
country. The ADDs constitute the primary management unit of extension services. The ADDs are subdivided into Rural
Development Projects (RDPs), which are further subdivided into Extension Planning Areas (EPAs). Extension agents called
Field Assistants supervise at the EPA level.
~ 35 ~
 Sample Frame

The sampling frame (or comprehensive sampling units) for this research is:

i. Groundnut Farmers in Central Region of Malawi (Study Population)

ii. Lilongwe and Salima Agricultural Development Divisions (ADDs)
within the central region
iii. Lilongwe West, Salima East Rural Development Projects (RDPs)
within Lilongwe and Salima ADDs
iv. Mkokera, Mchenchi, Cheseka, Kaphateya, Chafumbewa and
Chinguluwe Extension Planning Areas (EPAs) within the two RDPs. A
total of 200 groundnut farmers will be randomly selected within the
villages in the six EPAs for the interviews.

Sample Size4

Sample Size Calculations

To calculate a good representative sample size for the groundnut productivity
survey, the study focuses on groundnut farmers in the central region of Malawi,
which comprises about 70% (p =0.7) of the total groundnut production in the
country. Hence, for 95% (Z=1.96, 2-tailed test) level of confidence, within ±5% (e
= 0.05) margin of error, and taking into account the proportion of groundnut
farmers in the central region, the sample size n, is given as

Z 2 (1 p) p 1.962 (1 0.7)0.7
n 323
e2 0.052

and adding 5% non-respondents, the sample size is n = 339 groundnut farmers.

Sampling Method

Village and farmer Selection - A multi-stage cluster sampling procedure involving

a combination of purposeful and random sampling procedures are used to draw
the sample. The first step in village selection is a purposeful selection of EPAs
(cluster of groundnut producing EPAs) with relatively high level of groundnut
production5.

4
For detailed sample size calculations, refer to Chapter 4.
5
EPAs with relatively high groundnut production are identified from maps prepared by Famine Early
Warning System (FEWS) office in Agro-Economic Survey (AES), Department of the Ministry of
Agriculture and Irrigation, Malawi.
~ 36 ~
 This criterion is chosen with the need for active farmer participation in mind,
farmers having a commercial orientation and farmers having a subsistent
orientation toward the groundnut crop. Because groundnut is a minor crop in area
terms, it will be important to choose EPAs where the areas allocated to groundnuts
are relatively large in order to improve the likelihood that farmers would be
motivated to participate actively in the research.

Only two EPAs, Chafumbwa from Lilongwe ADD and Chinguluwe from Salima
ADD, are provisionally selected in such a way to capture variation in population
associated greatly in adopting improved technologies, as the alternative of
increasing production from expansion in cultivated area is not available. The rest
of the EPAs are selected randomly from the list of EPAs obtained from the Field
Assistants. These are Mkorera, Mchenchi and Cheseka EPAs from Lilongwe
ADD and Kaphateya EPA from Salima ADD.

The second stage is the probability of selecting sample villages (or clusters) from
the list of villages in six EPAs selected above. A total of ten villages are selected
from the EPAs ensuring that villages with larger proportion have proportionally6
greater chance of containing a selected cluster than small villages.

The final stage is the section of households to be interviewed. Households are

selected randomly from each village selected in the second stage. The lists of
names of the villages are also obtained from the Field Assistants (or extension
workers) who are responsible for the respective EPAs. These names are assigned
numbers, and using table of random number a total number of 34 farmers are
therefore randomly selected from each village. From the 10 villages, a total
sample size of 340 farmers is obtained for the whole survey.

Training Enumerators
Prior to data collection, the enumerators are first trained by the principal
researcher on techniques of administering the questionnaire for collecting
agronomic and socio-economic data. This involves explaining in detail all the
questions to the enumerators. This is done to ensure that the enumerators
understood what each question is asking for, and that the enumerators should be
asking the same questions to the respondents thereby minimizing enumerator bias
and other errors.

Pre-testing of the questionnaires follows the training of the enumerators, and this
is aimed at detecting problems in the wording of questions bearing in mind that

6
This type of sample is self-weighting, which will simplify the analysis and improving the
representativeness of the sample.
~ 37 ~
 the questionnaire is written in English but it is to be administered in Chichewa 7.
Questionnaire pre-testing may also allow enumerators to be exposed to the real
field situations and get used to the questionnaire. After these exercises, all the
necessary changes are made on the questionnaires, and then are administered to
the selected or sampled households in the villages.

Sensitization of the Communities

The project managers in the ADDs and Field Assistants in EPAs in the selected
areas are responsible for sensitizing the chiefs, village committees, the farmers
and related government agencies.

Time of Actual Survey

The actual survey was conducted from July to August 2000 by the principal
researcher with the help of ten enumerators. Data collection exercises involved
administering a structured questionnaire to the sampled household heads or their
spouses where the household heads were not available and participatory focus
group discussions with the stakeholders – chiefs, field assistances, extension
workers, project managers and researchers.

POINTS TO PONDER: Considering good data, the critical aspects of data

collection systems are the coverage, comparability and quality of the data produced.
Coverage refers to the completeness with which the population or events being
studied are in fact enumerated and the range of information obtained about the unit of
observation. In addition, the time reference of the observation, i.e., whether they have
reference to an instant or to an internal in time, may be considered, as well.

Studies and analyses drawing upon data collected in different places, at different
times or utilizing different data collection systems are beset by questions of
comparability. Lack of comparability occurs most often as a result of dissimilarities
of definitions, but may arise from dissimilarity of coverage, differential accuracy and
differential reliability and validity of the data-collecting instruments.

Quality refers primarily to the accuracy of the data, the extent to which the recorded
observation corresponds to the characteristic or attribute of the unit observed, but
also to the validity and reliability of recording or data-collecting instrument or
technique.

Validity is the extent to which the data-collecting instrument records datum is

referring to the attribute or characteristic being studied rather than to some
alternative, perhaps closely related, attribute or characteristic of the unit of
observation. Reliability is the extent to which the particular datum obtained for the
unit of observation is independent of the person or agency.
7
The local language for the majority of the farmers in the area and it is the national language of
Malawi
~ 38 ~
V. Participatory Rural Appraisal
By Mckey Mphepo, Former Lecturer in Extension, Bunda College, University of
Malawi

Over the years Social Scientists have conducted research to collect primary data from
communities in both urban and rural areas. Quantitative data has been collected for
years and used in explaining the setting in particular situations. During the mid-1980s,
there was realization that quantitative data in itself may not be adequate; there is
usually underlying information that is not collected which can provide an explanation
to a situation. Hence, the promotion of qualitative data collection through
Participatory Rural Appraisal is a tool where researchers discuss issues with the
communities in greater detail.

2.9. What is the advantage of collecting qualitative data?

The advantage of collecting qualitative data is that detailed information that
touches on People‘s feelings, values, attitudes and beliefs is gathered. This
information is not easy to collect in a structured, closed ended questionnaire;
hence the qualitative data supplements the quantitative one. Sometimes the
respondent is restricted to a yes/no answer when s/he wants to say yes, but….
The qualitative method of enquiry can elicit that. In most cases participatory
rural appraisal especially its precursor {RRA} is less costly to administer.

2.10. What is the disadvantage of having qualitative data?

The disadvantage of having qualitative data is that the results are basically
descriptive as opposed to numeric values/figures attached to them. Sampling
of respondents is by value of information, while for quantitative data
collection sampling is done by a pre-determined statistical design. The
analysis of the data is manually done or free – form to suit the researcher. On
the contrary, quantitative data is analyzed statistically to meet methodological
criteria. Participatory rural appraisal tends to take a long time, especially when
the researcher really wants the community to fully participate.

2.11. What are the various methods associated with Participatory

Rural Appraisal (PRA)?
Participatory observation is a method that has been borrowed from
Anthropology. The researcher gets involved in carrying out an activity
together with the community members in order to observe how such an
activity is performed. This method is handy where the community is not so
free to be interviewed or has been interviewed so many times that the people
experience a ‗burn-out‘. It is also helpful in finding out sensitive information,
or where people say things to please the researcher.
~ 39 ~
 PRA has various methods or tools which include: focus group discussions,
village mapping, seasonal calendars, and interviews with key informants, as
well as, historical profiles, social or wealth ranking, matrix ranking,
participatory observation and transect walks.

2.12. What is Focus Group Discussion (FGD)?

Focus group discussions are a good tool that a researcher uses to solicit
information from a community. S/he needs to develop guiding questions for
the discussion. These questions must be thought through carefully so that the
required information is collected. During the discussions other follow- up
questions are posed so that the responses are as clear as possible. It is usually
good practice that there are at least two researchers during the focus group
discussion; one facilitates the discussions and the other takes note of the
issues and agreements of the group.

Consensus is important in focus group discussions; the researcher taking

notes has to pay particular attention to what the group has generally agreed. If
there are dissenting views, these need to be noted too, because you do not
want to lose any information. At times one has to divide the groups according
to the social setting so that no category of the community dominates the
other.

For example, in most rural communities men tend to dominate discussions in a

group setting of both men and women. In such circumstances the researcher
would want to have separate groups of women and of men. The village leaders
should also be separated from the rest of the participants in the discussions;
the Village Head‘s opinion tends to be final. It is important that opinions are
solicited from all the sectors of the community; the information should not be
biased towards one sector of the community.

2.13. What is Village Mapping (VP)?

Village mapping is another method of carrying out data collection especially
when the researcher is not familiar with the area. In the process of the
community members drawing the map of the village resources and facilities
are indicated. This gives the researcher opportunity to know what resources
are available in the area and the members of the community become more
aware of their own available resources. Village mapping assist the
Development Agencies to provide appropriate services through diagnosis of
the problems and analysis of the solutions together with the people of the area.

Transect walk involves both the researcher and the community members to
walk across the village(s) to see and note the various resources available. One
could also notice present problems, for example, soil degradation.

~ 40 ~
2.14. What is Historical timelines?
Seasonal calendars involve the people in identifying what activities they
undertake at particular times of the year. The information is essential when
Development Agencies would want to introduce a project; consideration
should be made on what periods of the year people are busy, or have time
available for development activities. For example, introducing ‗Food for
Work‘ project during the growing season would be disastrous; people would
rather concentrate on their own fields than do community work for food.

A lot of information can be collected by interviewing key people in the village

or community. For example, the Village Heads, prominent village persons,
religious leaders, local political leaders can be interviewed and provide
valuable information about the community for development purposes.

Community members can be requested to draw historical profiles on their

village. They can indicate what happened during certain crucial periods in the
history of their village. They can also indicate the natural resources that were
available as compared to what is available at present. These are sometimes
called historical time lines. An analysis of these can lead into problems that
the community is facing that the Development Agency can facilitate to solve
through various interventions.

2.15. What are Matrix ranking?

In a community that is new to a Development Agency, it is important to know
who are the poorest and the richest people. The need may be to provide
development facilities to the poorest segments of the community. Wealth or
social ranking becomes handy in soliciting information on the status of the
various people. In every community there is a different way of ranking the
social status of community members and this information can be solicited in a
participatory way.

Problems in rural areas are many. There is need to rank such problems so that
the most important ones are attended to first. Matrix ranking (pair-wise and
scoring/voting) helps the researcher together with the people to select the
pressing problems to be addressed as a matter of urgency. The community
members can score or vote the highest problem using stones or seeds. The
pair-wise ranking compares each problem with the others at a time on the
basis of the importance or severity of the problem. The most mentioned
problem then ranks the highest and consequently may require the earliest
intervention.

~ 41 ~
2.16. What is Venn diagram?
Venn diagram is yet another participatory method where interrelationship of
individuals or organizations is studied. The further away from the center, the
less the relationship is. The size of the circle around the individual or the
organization also signifies the strength of the relationship.

POINTS TO PONDER
Qualitative and quantitative methods of collecting socio-economic data have
advantages as well as disadvantages. The use of qualitative methods in gathering
data is not a panacea to all the limitations associated with quantitative methods. Uses
of both methods assist in overcoming some of the limitations and ensure that
reasonably accurate data is collected.

~ 42 ~
========================================================
MENTAL GYMNASTICS
CHAPTER TWO
=======================================================

1. Distinguish between the following of pairs of terms.

a) Cross-sectional and Panel data
b) Reliable and Valid questionnaire
c) Data analysis and Data collection

2. True, False or Uncertain. Support your answer.

a) The types of the data being used determine the type of model being applied.
b) Reliable questionnaire is more important than consistent questionnaire.
c) The critical aspects of data collection systems are the coverage, comparability
and quality of the data produced.

3. What is a good questionnaire?

4. Discuss the processes of designing a good questionnaire.
5. Explain advantages and disadvantages of semi-structured and structured
questionnaires.
6. Explain pro and cons of closed questions.
7. Provide a check list for planning a survey.
8. Discuss quality of data concept.
9. What are the most important criteria of designing a questionnaire?
10. Why sample weighting will simplify analysis and improve the representativeness
of the sample?
11. Discuss the advantages and the disadvantages of qualitative and quantitative data.

~ 43 ~
What is the problem associated with underdevelopment
in Africa? The number one problem is that there is
disconnection between African intellectuals, politicians
and its natural resources. The national accumulated
human resource and the natural resources could not
communicate; and therefore, exposed to external
exploitations directly or indirectly that undermine
national development.

~ 44 ~