SESSION 7

SAMPLING & ACCESS: CHOOSING &

REACHING YOUR RESEARCH PARTICIPANTS

Dr ZHENG Weijia
Offi ce ： IB941
Email ： zheng.weijia@bfsu.edu.cn
Session Overview
• Introduction to concept of sampling
and its rationale
• The underlying theoretical assumptions
of sampling decisions
• How to sample and why
• Issues in generalization of findings,
especially in qualitative research
projects
• Access issues
What is Sampling?

• The “who?” or “what?” of your research

• Sampling involves selecting some of the
elements of a wider population to take part in
your research
• It is connected to epistemological questions
like ： What counts as valid knowledge? How
similar is human behaviour?
• Is related to your chosen method (and therefore
your epistemology)
Why sample?

• Practicality
- often a researcher can’t survey everyone (time,
access)
• To make inferences about a wider group of
people or things based on the behaviour/
responses of those in your selected sample
Sampling:
Important in ALL projects
• Not just research involving people
- E.g., secondary analysis of financial indicators
- Which sources and why?
• A logical, strategic choice about where the best
place to find the indicators of your concepts that
comprise your variables…
• Also relates to your choice of case study:
- Why this or that company?
- N.B. Sampling also goes on inside the case study
What’s the relationship between
a population and a sample?

A. The sample is the chosen population

B. The sample is a subset of the population
C. The sample must represent the population
D. The sample is the way you choose the
population
Definitions

• Population
- the entire group of people/ events/ things you want
to investigate
e.g., Donators to charity; the charity sector
• Sample
- your strategically selected subset of the population
Definitions

• Element: a single member of the population

• Sampling frame: a list/ directory of all the
elements in the population
- More later…
The essential question

“Why are the people in my

sample best placed to help me
answer my research questions?”
In reality…
• Sampling is usually a trade-off between what is
desirable & practical
• Best fit with research question BUT:
- Valid/ rich data vs. access to subjects
- Confidence & margin of error tolerated (later on)
- Time, cost, skills
Theoretical assumptions for
research projects
• Theoretical assumptions:
- Generalisation
- Specialisation
Theoretical assumptions (1)
• Generalisation
- If we ask enough people, selected using
objective statistical techniques we can more or
less assume their responses will match those of
the population
- Usually quantitative research (standardised,
measurable) and not usually interpretivist aim
- Probability sampling needed for statistical
generalisation
Types of probability
sampling
• Simple random sampling
• Stratified random sampling
• Systematic sampling
• Cluster sampling
Simple Random Sampling

• In simple random sampling, each

element in the population has a
known and equal probability of
selection.
• Each possible sample of a given size
has a known and equal probability of
being the sample actually selected.
• Size of sample needed: better with
over a few hundred
Systematic Sampling (1)
• Systematic sampling: two steps
- Step 1: A sampling interval (i) is determined
by dividing the population size (N) by the
sample size (n) and rounding to the nearest
integer.
- Step 2: The sample is chosen by selecting
a random starting point, and then picking
every ith element in succession from the
sampling frame.
• Easier to do manually with larger sample
size.
Systematic Sampling (2)
• E.g.,
- There are 100,000 elements in the
population and a sample of 1,000 is
desired. In this case, the sampling
interval (i) is 100.
- Then a random number between 1 and
100 is selected. If, for example, this
number is 23, then the sample consists of
elements 23, 123, 223, 323, 423, 523…
Systematic Sampling (3)

• However, if the target population covers a

large geographical area, the random
selection means that the sample cases are
likely to be dispersed throughout the area.
 is suitable for geographically dispersed
cases only if you do not require face-to-
face contact when collecting your data.
Stratified Sampling (1)
• Stratified sampling is a also two-step
process
• Step 1:
- the population is partitioned into
subpopulations, or strata.
- The strata should be mutually exclusive
and collectively exhaustive, i.e., every
population element should be assigned to
one and only one stratum, and no
population elements should be omitted.
Stratified Sampling (2)
- The elements within a stratum should be as
homogeneous as possible, but the
elements in different strata should be as
heterogeneous as possible.
• Step 2:
- Elements are selected from each stratum
by a random procedure, usually Simple
Random Sampling.
• Better comparison and representation
across strata; difficult to select relevant
stratification variables.
 Mini case
Dilek worked for a major supplier of office supplies to public
and private organisations. As part of her research into her
organisation’s customers, she needed to ensure that both
public- and private sector organisations were represented
correctly. An important stratum was, therefore, the sector of
the organisation. Her sampling frame was therefore divided
into two discrete strata: public sector and private sector.
Within each stratum, the individual cases were numbered.
She decided to select a systematic random sample. A
sampling interval of 4 meant that she needed to select every
fourth customer on the list. As indicated by the ticks ( ✓),
random numbers were used to select the first case in the
public sector (001) and private sector (003) strata.
Subsequently, every fourth customer in each stratum was
selected.
Mini case
Cluster sampling (1)
• In cluster sampling, the target population
is first divided into mutually exclusive
and collectively exhaustive
subpopulations, or clusters.
• Elements within a cluster should be as
heterogeneous as possible, but clusters
themselves should be as homogeneous
as possible.
Ideally, each cluster should be a small-
scale representation of the population.
Cluster sampling (2)
• Then: a random sample of clusters is
selected, based on a probability
sampling technique such as SRS.
Easy to implement; reduced precision.
 Mini case
Ceri needed to select a sample of firms to undertake an
interview-based survey about the use of large multi-
purpose digital printer copiers. As she had limited
resources with which to pay for travel and other associated
data collection costs, she decided to interview firms in four
geographical areas selected from a cluster grouping of
local administrative areas. A list of local administrative
areas formed her sampling frame. Each of the local
administrative areas (clusters) was given a unique number,
the first being 0, the second 1 and so on. The four sample
clusters were selected from this sampling frame of local
administrative areas using simple random sampling.
Ceri’s sample was all firms within the selected clusters.
She decided that the appropriate directories could probably
provide a suitable list of all firms in each cluster.
Cluster sampling vs.
Stratified sampling
Calculating the minimum
sample size
Based on:
• how confident you need to be that the
estimate is accurate (the level of
confidence in the estimate);
• how accurate the estimate needs to be
(the margin of error that can be
tolerated);
• the proportion of responses you expect
to have some particular attribute.
Calculating the minimum
sample size
Once you have all the information you substitute it into
the formula:

Where:
• n is the minimum sample size required
• r% is the percentage belonging to the specified category
• q% is the percentage not belonging to the specified
category
• z is the z value corresponding to the level of confidence
required (see the table on the next slide)
• e% is the margin of error required
 Calculating the minimum
sample size
• Levels of confidence & associated z values

• To estimate the proportion of responses you

expect to have a particular attribute:
- collect a pilot sample of about 30 observations
- from this to infer the likely proportion for your
main survey
Calculating the minimum
sample size

• Where your population is less than

10,000, a smaller sample size can be
used without affecting the accuracy.
• This is called the adjusted minimum
sample size.
Calculating the minimum
sample size
Adjusted minimum sample size is
calculated using the following formula:

Where:
• n' is the adjusted minimum sample size
• n is the minimum sample size (as
calculated previously)
• N is the total population
To answer a research question Jon needed to estimate the
proportion of a total population of 4000 restaurant
customers who had visited that restaurant at least five times
in the past year. Based on his reading of the research
methods literature he decided that he needed to be 95%
certain that his ‘estimate’ was accurate (the level of
confidence in the estimate); this corresponded to a z score
of 1.96. Based on his reading he also decided that his
‘estimate’ needed to be accurate to within plus or minus 5%
of the true percentage (the margin of error that can be
tolerated).
In order to calculate the minimum sample size, Jon still
needed to estimate the proportion of respondents who had
visited the restaurant at least five times in the past year.
From his pilot survey he discovered that 12 out of the 30
restaurant customers did so – in other words, that 40%
belonged to this specified category. This meant that 60% did
Calculating the minimum
sample size
Jon substituted these figures into the formula:

His minimum sample size, therefore, was 370.

As the total population of restaurant customers

was 4000 (less than 10,000), Jon could now
calculate the adjusted minimum sample size:
Calculating the minimum
sample size

Because of the small total population, Jon’s

adjusted minimum sample size is only 339.

However, this assumed he had a response rate of

100%!
Estimating response rates &
Actual sample size

Where:
• is the actual sample size required
• n is the minimum (or adjusted
minimum) sample size
• re% is the estimated response rate
expressed as a percentage
Theoretical assumptions (2)

• Specialisation
- We need to find out as much as possible about a
specific topic, so we must choose a sample who
will be able to tell us the most about it
- Statistical generalisation is not an aim
- Usually qualitative research (in-depth)
Sampling for non-generalisable
research projects
• Purposive, judgmental process
- Choose that from which we can learn the most
• Common sampling methods
- Snowball
- Case-study
- Convenience
Sampling for non-generalisable
research projects
Snowball Sampling
- In snowball sampling, an initial group of
respondents is selected, usually at random.
- After being interviewed, these respondents are
asked to identify others who belong to the target
population of interest.
- Subsequent respondents are selected based on
the referrals.
- Useful if hard to reach population, e.g., drug users.
- Can estimate rare characteristics; time consuming.
Sampling for non-generalisable
research projects
Case-study
- Using a set of criteria to identify the critical case
for in-depth explorations
- Sampling criteria: serve the aim of effectively
answer your research questions
Sampling for non-generalisable
research projects
Convenience sampling
- attempts to obtain a sample of convenient
elements.
- Often, respondents are selected because they
happen to be in the right place at the right time.
- E.g., our samples today; “people on the street”
interviews
Generalisation from qualitative
research
• We do it all the time (most, few, some, many)
• Not statistically generalisable
• Taps into beliefs about nature of knowledge
• Recognise limits of knowledge
- ‘Moderatum generalisation’ (Williams 2002)
Analogy: photography vs. art?
Now you know…
• …why researchers sample, and what this means
for the data they produce
• …sampling methods can be served for
generalisable & non-generalisable purposes
• …how to choose your sampling strategy
• …what generalisation means for qualitative
research
OK – So where are you going
to find your sample?
Research sites
• You define this
- Depends on research objectives
- Logical and justified
• Employees
- Organization
- Professional body
- Trade unions
• Types of consumer
- Where are they most likely to be found?
Research sites
Your population is: ‘users of premium price
beauty products’ – where’s the best place to
find them?
1. Upmarket gym/ spa
2. Student union
3. In the supermarket
4. At the beauty counter in a dept. store
5. All of these…
Getting Access - problems
• Orgs have many requests for research access
• They are busy
• Commercial sensitivity
• Student status can be low
• May not see the value of research findings
- Why a combination of academic & applied rationale
needed
Strategies to gain access
• Be realistic
- But choose the very best site you can
e.g., Street surveys are weak…!
• Gatekeepers
- Personal contacts
- Find out who to approach first!
• Supervisor letter
• Good clear research plan & objectives
Once you’re in…
• Physical access
- Permission to carry out the research
• Cognitive access
- Trust and co-operation of respondents
- Response rates (survey)
- Rapport – rich data (interviews)
- Acceptance (participant observation)
- ‘Truthful’ data
• Ongoing process – leads into research ethics
(1st session)
Quiz time
________ is a probability sampling
technique in which each element in the
population has a known and equal
probability of selection.
A) Simple random sampling
B) Snowball sampling
C) Cluster sampling
D) Systematic sampling
Quiz time
A probability sampling technique in which the
sample is chosen by selecting a random
starting point and then picking every i th
element in succession from the sampling frame
is called ________.
A) simple random sampling
B) snowball sampling
C) cluster sampling
D) systematic sampling
E) succession sampling
Quiz time
________ is a nonprobability sampling
technique that attempts to obtain a sample of
convenient elements and leaves the selection of
sampling units primarily to the interviewer.
A) Convenience sampling
B) Snowball sampling
C) Simple random sampling
D) Systematic sampling
E) Window sampling
Quiz time
A nonprobability sampling technique in which an
initial group of respondents is selected randomly
and subsequent respondents are selected based
on the referrals or information provided by the
initial respondents is called ________.
A) respondent sampling
B) snowball sampling
C) stratified sampling
D) cluster sampling
Quiz time
A(n) ________ is the total of all the elements
that share some common set of characteristics.
A) sample
B) population
C) inference
D) hypothesis
E) elements
Quiz time
A(n) ________ is a subgroup of the elements of
the population selected for participation in the
study.
A) census
B) element
C) incident
D) sample
E) segment
Quiz time
Probability sampling relies on the personal
judgment of the researcher, rather than chance, in
selecting sampling elements.

false
Group Presentation 2
Mid-term test…
• When: 9th week (after the immersion week)
- Check BB announcement for the exact date,
time & venue!
• Duration: Two hours
• Form: multiple choice, T/F, short essay questions
• Worth: 25% of the final grade
• N.B. Adjustment in assessment ！
- 10%: class attendance
- 10%: class interaction participation
Next…
• Enjoy the immersion week & Good
luck with your mid-term test!
• Read Chapter 4 before coming to
the next session (on April 30th)
• 3rd group presentation in the next
session
- Case & questions will be available on
BB one week in advance