Capacity building for research: promoting inclusive development of agricultural value-chains, 1-3 September 2014

Sampling, questionnaire and interview design

Risti Permani, Global Food Studies, University of Adelaide

adelaide.edu.au
Recall yesterday’s topic: Governance
analysis

Mapping Demand
and
actors supply

Coordina Participa
tion tion

Rules and Enforce

regulations ment

Knowledge Services

University of Adelaide 2
Recap yesterday’s topic:
Example when we observe ‘services’ aspects
of governance, this is what we want to get

Types of Provider Available for Degree of How to

services different VC satisfaction improve
actors with the
services

University of Adelaide 3
• Improving data collection techniques
will enhance the accuracy, validity,
and reliability of research findings.

University of Adelaide 4
Outline
1. Different types of data collection
2. Basic primer of sampling
3. Questionnaire design
4. Interview design
5. Data analysis

University of Adelaide 5
1. Different types of data
collection
Representativeness and objectivity
The extent to which we can
Observations using direct assume that what we observe
measurements vs subjective with the tool represents what
judgements that rely on the happens in the country;
beliefs of the appraiser depending on: sample size and
sampling techniques

Source: LSMS-ISA (World Bank)

University of Adelaide 7
How to select methods
• Redundancy with other studies
• Logistics: costs, time available, staffing, access to records
• Sampling: framing availability, sampling size
• Depth of information
• Generalisability
• Sensitivity of issue

 In most cases, methods are used in sequence:

• Provide explanation
• Used in instrument development

University of Adelaide 8
Total survey quality

Regardless of the
methods chosen,
quality research
includes collecting
quality data.

University of Adelaide 9
Training on household survey
• Useful resource: http://go.worldbank.org/FZIIVCYCQ0
• Adopt a similar approach to the World Bank’s LSMS-ISA
training workshop
– The Living Standards Measurement Study - Integrated Surveys
on Agriculture (LSMS-ISA) is a $19 million household survey
project established by the Bill and Melinda Gates Foundation
and implemented by the Living Standards Measurement Study
(LSMS) within the Development Research Group at the World
Bank.

University of Adelaide 10
Data collection: By participants

Household survey

Middle-segment of value
chains

University of Adelaide 11
Interviews
• Interviews to gather opinions, perception, attitudes.
• Interviews to gather background information:
– Expert knowledge
– Facts, descriptions of process
• Some interviews may include both aspects

University of Adelaide 12
Types of interviews

• Free-flowing; vary much by the respondent

• Can take a long time
Unstructured

• Used often in policy research

• Using a guide with questions and aspects that must be covered.
Semi- • To ensure complete and consistent information across different interviews.
structured

• Questions are fixed and asked in a specific order.

• Similar to a survey
Structured

University of Adelaide 13
Example

• I’m here to understand how you operate your farm.

Unstructured

• Could you describe how you improve your farm performance?

• PROBE: How do you monitor your farm performance?
Semi- • PROBE: What do you do to increase feed quality?
structured • PROBE: Do you receive any support from governments or NGOs?

• Do you monitor your farm performance:

• [ ] Yes
• [ ] No
Structured • [ ] Other

University of Adelaide 14
2. Sampling
Sampling
• World Bank: “Sampling is said to be hard and better left
to experts”.
• RAND: “Get sampling help!”
• Risti: Agree, will only do (a very brief) introduction to
some sampling methods.

University of Adelaide 16
 Generalisability claims
- Random and
structured

Inferences
- Non-probability
sampling

University of Adelaide 17
Non-probability sampling
• Convenience sampling eg grab or opportunity sampling;
– useful for pilot testing;
– cannot scientifically make generalisations about the whole
population from this sample.
• Quota sampling
– Segmented into mutually-exclusive sub-groups, then judgement
is used to (non-randomly) select sample from each segment.
• Purposive sampling
– A limited number of people that have expertise or knowledge in
the area being researched.
• Snowball sampling
– Locating informant-rich key informants

University of Adelaide 18
Simple random sampling

If sampling frame is
large, this method is
not practical.

Source: http://faculty.elgin.edu/dkernler/statistics/ch01/1-4.html

University of Adelaide 19
Stratified sampling

Non-overlapping
groups;
Individuals within
each group
(stratum) should be
similar in some
way;
Proportional
sample from each
stratum.
Source: http://faculty.elgin.edu/dkernler/statistics/ch01/1-4.html

University of Adelaide 20
Systematic sampling

First order our

individuals, then
select every k-th;
k=N/n; where
N=total population
and n=sample size

Source: http://faculty.elgin.edu/dkernler/statistics/ch01/1-4.html

University of Adelaide 21
Cluster sampling

We use it when our

population is
already broken up
into groups
(clusters)), and
each cluster
represents
population;
We assume clusters
do not differ
significantly from
each other.

Source: http://faculty.elgin.edu/dkernler/statistics/ch01/1-4.html

University of Adelaide 22
Multistage sampling
• The process of taking random samples of preceding random
samples.
• Multistage sampling is used frequently when a complete list
of all members of the population does not exist.
• Reduce costs.

University of Adelaide 23
Errors
1. Non-sampling error: results from the survey process
– Incomplete frame
– Non-response
– Interviewer error
– Misrepresented answers
– Data checks
2. Sampling error: results from using a sample to estimate
information regarding a population

University of Adelaide 24
 3. Questionnaire Design

University of Adelaide 25
Steps

DRAFTING
LITERATURE REVIEW QUESTIONNAIRES: PILOT TEST:
and INITIAL Questions, types of
CONSULTATIONS WITH Fieldwork to test draft
variables, formatting, questionnaires, FINAL SURVEY
STAKEHOLDERS: length, characteristics observe ‘respondent
Checking policy of respondents, policy fatigue’ and then revise
relevance relevance, plan for data
analysis

University of Adelaide 26
The factors of a good questionnaire (1)
 The questionnaire is asking the right questions.
• Initial consultations; pas surveys
 Questions are asked to the right respondents.
• The preferred respondent is the best informed respondent; plan to
record ‘proxy answering’
• In general small children should not be asked questions directly.
• Education to all respondents (except young children), but food
consumption should be asked to the housewife.
• The wording should make clear the group of people to which the
question refers to:
o Do you, or anyone in your household, own or operate a non-farm
enterprise?
 Respondents understand the questions.
 The flow of the questions is adequate.
 The respondents are willing to answer questions.
 The answers are recorded correctly.

University of Adelaide 27
The factors of a good questionnaire (2)
 Respondents understand the questions.
o Avoid ambiguous wording
o Ask one question at time.
o Be clear who the question refers to – either an individual
respondent, the household, a firm, etc.
o Make the reference period explicit.
o Avoid jargon or overly academic phrased
o Language should be straight-forward (avoid double negatives)
o Questions should be culturally sensitive and appropriate
o Even the most carefully developed question is only as effective as
its translation.
 Interviewers will be instructed to read the questions
exactly as they are written on the questionnaire

University of Adelaide 28
One-minute group discussion #1
• Exercise 1: We want to measure the current net present
value of the respondent’s land. Which of the following
wordings do you think is more appropriate?
a) How much is your land worth?
b) What is the current net present value of your land?
c) How much would you sell your land for today?
d) If you were to sell your land today, what price do you believe
you could receive?
e) How much did you pay for this land?

University of Adelaide 29
One-minute group discussion #2
• Exercise 2: Is there anything wrong with the wording of
any of these questions?
1. How much do you spend on food consumption?
2. When did you start farming?
3. Do you have experience getting loans from a bank,
cooperative or other institutions?
4. Are you aware that being a member of a cooperatives is
not improving your welfare and productivity?

University of Adelaide 30
The factors of a good questionnaire (3)
 The flow of the questions is adequate.
o The flow affects the time, interviewer’s ability to engage the
respondent’s interest; and the quality of the information being
collected.
o Questions should be grouped by topic and logically sequenced.
o Putting ‘skips’
 The respondents are willing to answer questions.
o Examples: income, disease, etc.
o No rules to deal with sensitive questions: but it is generally
preferable to leave them for the final stages of the interview.
o Record the presence of other people

University of Adelaide 31
The factors of a good questionnaire (4)
 The answers are recorded
correctly.
o ‘Don’t know’
o Question numbering
o Linear vs grid format Source: Vietnam’s VVLS Handbook instruction for
interviewers 1997/1998 (115 pages), available at:
o Interviewer manuals http://lsms.adeptanalytics.org/course/fscommand/s
ession2/ref/VietNam_interviewer_manual.pdf
o The role of manuals

University of Adelaide 32
Pilot testing: Reality checking

Some considerations Summary

• Should be time-watched. • To check:
– A rule of thumb: 1. Wording
questionnaires should not ask
2. Format
more than 90 minutes of
questions to a single 3. Flow
respondent in a single sitting. 4. Duration
• Should be done by experienced
interviewers or researchers?
• On a random sample or
purposive?
• A iterative process?

University of Adelaide 33
4. Interview design
How to do a (good) research interview?
1. Knowledgeable (familiar; test the materials)
2. Structure

Introduction
Main body of End of interview
Topic of interviews interviews Thanks, additional
Ethics, informed points, queries, ethics
consent

University of Adelaide 35
How to do a (good) research interview?
(cont)
3. Clear (questions are simple, easy, short, don’t use
jargon)
4. Gentle (give the interviewees time to finish)
5. Sensitive (listen!)
6. Steering (needs to know what s/he wants to find out)
7. Critical (not to disagree but more to ‘challenge’
consistencies, etc).
8. Remembers (what has been said previously)
9. Interprets (clarifies and extends the meaning).
10. Balances (don’t talk too much, but not too little)
11. Ethically sensitive (ensure the interviewee appreciates
what the research is about)

University of Adelaide 36
Semi-structured interviews
• Might be best to follow ‘a guide’: key topics to be covered and
the order.
– Introduction  Ground rules  Questions and probes  Thank you
and next steps
• Use of probes:
– FOR CLARITY:
• Can you be more specific?
• What is your best estimate?
• If you had to pick one answer, what would you choose?
– FOR COMPLETENESS:
– OTHER PROBING TECHNIQUES:
• Repeat the question
• Echo their response
• Pause a second
• Baiting eg “I’ve heard some people said that ….”

University of Adelaide 37
The interview team

INTERVIEWERS NOTE TAKERS

• Gain cooperation • Record accurately
• Listen • Note subtleties, non-verbal
• Be neutral behaviours
• Maintain confidentiality • Understand their role in the
session
• Understand when clarification
is needed
• Be cost-effective

University of Adelaide 38
Overcoming common barriers
• I’m not interested”: Explain purpose of research and remind
them it’s an opportunity for their opinions to be heard.
• “I’m too busy. How long will this take?”: Be honest about time
and suggest you can schedule for a more convenient time .
• Fear of being inadequate: Provide reassurance that we want
their opinion, that there are no right or wrong answers.
• Have a negative reaction to research: Reiterate why the
research is important, provide information on legitimacy of
the research.
• Confidentiality concerns, questions too personal: Explain that
answers will be combined with answers from others and
presented in aggregate form; no names will be reported

University of Adelaide 39
5. Data Analysis
• Assuming that (good quality) data have now been
collected, what’s next?
• What will your type of research? Is this something you
need to think about before you conduct the fieldwork?

University of Adelaide 41
Quantitative versus qualitative research
Criteria Qualitative research Quantitative research
Purpose To understand and To test hypotheses, look at
interpret social interactions cause and effect and make
predictions
Group studies Smaller and not randomly Larger and randomly
selected selected
Form of data collected Qualitative data such as Quantitative data based on
open-ended responses, measurements using
interviews, participant structured and validated
observations, field notes data-collection instruments
and reflections
Type of data analysis Identify patterns, features Identify statistical
and themes relationships
Objectivity and subjectivity Subjectivity is expected Objectivity is critical
Results Particular or specialised Generalisable findings that
findings that is less can be applied to other
generalisable populations
University of Adelaide 42
Successful data analysis
Whether quantitative or qualitative, successful data
analysis requires
1. understanding a variety of data analysis methods;
2. planning data analysis early in a project and making
revisions in the plan as the work develops;
3. understanding which methods will best answer the
study questions posed, given the data that have been
collected; and
4. once the analysis is finished, recognising how
weaknesses in the data or the analysis affect the
conclusions that can properly be drawn.

University of Adelaide 43
Three-minute group discussion

TWO MINUTES:
• Choose one of these three aspects and write down a set of
questions (maximum 3 related questions; good questions
please ) that will help you as a researcher understand a
key issue that farmers are facing:
a) Farm performance
b) Governance
c) Adoption of technology
ONE MINUTE:
• Then swap with the group sitting next to you and
comment each other

University of Adelaide 44
References
• Harrell, M.C. and Bradley, M.A. (2009). Data collection
methods: Semi-structured interviews and focus groups,
RAND Corporation, 2014(Available from URL:
http://www.rand.org/content/dam/rand/pubs/technical
_reports/2009/RAND_TR718.pdf
• LSMS-ISA (2014). Building capacity, The World Bank,
2014(Available from URL:
http://go.worldbank.org/FZIIVCYCQ0

University of Adelaide 45
• Terima kasih.

• Any questions?

• Email: [email protected]

University of Adelaide 46