Research Methods and Techniques - 04

Introduction to Sampling

Prof. Dr. Irena Ograjenšek

[email protected]

Academic Year 2020/2021

Agenda
• Clarifying terminology.

• Types of errors in sampling.

• Types of samples.

• Overview of probability samples.

• Overview of non-probability samples.

• New developments in sampling.

1
 Clarifying
Terminology

Basic Terminology (1)

• Population.

• Sample.

• Unit of observation.

• Population parameter.

• Sample statistics.

• Variable.

2
Basic Terminology (2)
Statistical population is always described in terms of:

– time
– place
– contents

Target Population
Its definition dependent on research goals.

Illustrative examples?

3
Basic Terminology (3)
• Population-base research.
→ census

• Sample-based research.
→ statistical inference

Pros of Sample-Based Research

• Temporal, financial and practical limitations imposed on
researcher.

• Quality samples yield good research results.

• Simpler and more efficient practical implementation.

• Census often not feasible.

4
Which information …

… is key in sample-based research?

Stages in Sample Selection

for Known Population
• Define the target population.

• Select a sampling frame.

• Determine a sampling method.

• Plan procedure for selecting sampling units.

• Determine sample size.

• Select actual sampling units.

• Conduct fieldwork.

5
Sample Frame
• An approximation of target population.

• A list of target population units from which we select sample

units.

Important issues:

• Classic sample frame?

• The importance of statistical and administrative registers for

sampling?

Typology of Sampling Units

• Primary

• Secondary

• Tertiary

• …

Illustrative examples?

6
 Types of Errors
in Sampling

Errors Connected to Sampling

Total Error

Random Systematic
Sampling Error
Error (Bias)

Respondent Administrative
Error Error

Sample Selection
Error

7
Random Sampling Error

• The difference between the sample result and the

result of a census conducted using identical
procedures.

• A statistical fluctuation which occurs because of

chance variation in the elements selected from a
sample.

• It is a function of sample size  as the sample

size increases, the error decreases.

Sample Selection Error

• An administrative error caused by improper

selection of sample due to either errors in sample
design or execution of the sampling procedure.

• Results in an unrepresentative sample.

8
 Types of Samples

Basic Typology

• Probability sample: every population unit has a known

non-zero probability of selection.

• Non-probability sample:
o sample units are selected on the basis of personal
judgement or convenience

o orthodox view: sample statistics obtained from non-

probability samples do not allow for statistical
hypothesis testing

o unorthodox view ?

9
 Probability
Samples

Simple Systematic Stratified Cluster Multistage

Random Sample Sample Sample Area
Sample Sample

Non-Probability
Samples

Convenience Judgement Quota Snowball

Sample (Purposive) Sample Sample
Sample

Factors Determining the Choice of

the Relevant Probability Sample
• Availability of sampling frame.

• Required sample size.

• Geographical dispersion of the population.

• Cost.

10
Applicability and Limitations
of Non-Probability Samples
• Applicable in early stages of research.

• Cost-effective and quick without demand for

representativeness.

• Serious limitation (orthodox view!): not possible to infer

from sample to population.

Factors Determining the Choice of

the Relevant Non-Probability Sample

• Nature of research.

• Temporal and financial limitations.

Discussion:
• Are all probability samples representative?
• Are all non-probability samples non-representative?

11
 Overview of
Probability Samples

Simple Random Sample

• Researcher assigns each population unit a number, than
selects sample units by random.

• Pro: A simple and well understandable procedure

• Cons:
– complicated for large samples
– inefficient use of information available for the population

12
Systematic Sample
• Researcher uses natural ordering or order of sampling
frame, selects an arbitrary starting point, then selects items
at a preselected interval.

• Might end up in a biased sample due to the preselected

interval.

Stratified Sample
• Researcher divides the population into groups (known
as strata – singular form = stratum) and randomly
selects subsamples from each group.

13
Cluster Sample
• Researcher selects sampling (primary-level) units at
random, then performs complete observation of all
secondary-level units (units in the selected primary
level unit).

• Extremely useful for geographically dispersed

populations.

Multistage Sample
• Researcher selects progressively smaller areas in each
stage.

14
 Overview of
Non-Probability Samples

Quota Sample
• Researcher classifies population by known
characteristics and determines desired proportion of
sample from each subgroup → fixes quotas.

• Very useful when we do not have access to sampling

frame but know characteristics of the target population
well.

15
Judgement (Purposive) Sample
An expert or experienced researcher selects the sample to fulfil
a purpose such as ensuring that all members have the desired
characteristics we might aim for:

– Heterogeneous units
– Homogeneous units
– Extreme units
– Typical units
– Critical units

Convenience Sample
Researcher uses the most convenient or the most economical
sample (units which are easiest to reach).

Illustrative examples?

16
Snowball Sample
• Initial respondents are selected by probability samples, additional
respondents are obtained by referral from initial respondents.

• Suitable for hidden (unknown) populations

Illustrative examples?

New Developments
in Sampling

17
Internet and Sampling
• Advantages: large, geographically dispersed samples
achievable quickly

• Bias stemming from self-selection of units into the sample.

Self-Selection of Units
Unites decide for themselves whether they want to be
part of the sample or not.

Why problematic?

18
Typology of Panels
• Classic panels.

• On-line opt-in panels.

Repetition

19
Mini Case Study
1st year students enrolled into the graduate study
programmes at SEB UL in the academic year 2019/2020
are a known population.

School management would like to know the share of this

population which has, in their circle of family and friends, at
least one person who has already successfully coped with
the COVID-19 infection.

You and three of your classmates were asked to help in the

survey data collection process. How would you tackle the
sampling challenge in order to obtain a sample that is
representative for the population as a whole?

Corona Challenge
The period of pandemic-related lockdown has many
negative effects on our daily lives. At the same time it
shows us more than ever how persistently optimistic the
human spirit is. The lockdown spawned many new jokes
and a series of personal reflections on lockdown and its
consequences for private and business life as well as
study processes.

Would you care to share you personal reflection on

lockdown? Or your favourite corona joke?

If yes, send it to the e-mail address

[email protected]
until Tuesday, November 3rd, at midnight.

20