18-Oct-24

Chapter 3: Sampling design 3.1. Sampling procedures

and data collection
 Planning the sample
 Probability and non-probability
Sampling procedures
sampling procedures
Data collection
146 147
18-Oct-24 18-Oct-24

146 147

3.1.1. Planning the sample 3.1.1. Planning the sample

Population: The identifiable set of elements
of interest to the researcher and pertinent to
Objectives of sampling
the information problem. Estimation of information about a
Definition of sampling population based on a sample from
 Selection of a small number of elements from the population.
a larger defined target group of elements and
expecting that the information gathered from Hypothesis testing to compare the
the small group will allow judgments to be relationships between data items for
made about the larger group. selected population groups.
 Sampling is often used when it is impossible or
unreasonable to conduct a census.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
148 18-Oct-24
149

148 149

3.1.1. Planning the sample Sample planning process

 Define the population: A population consists of the
Steps in sample planning
total set of individuals, households, or businesses you
want to include in a study. Location and time are
Define Census vs. Sample important factors in defining a population.
population sample design
 Census vs. sample: A census is a complete survey
of an entire population, while a sample gets
information from just a small, but hopefully
Execute Estimate
sampling costs of Sample size representative, fraction of the population. A census
process sampling has many advantages if the population is small and
within a workable location. The two major advantages
of using a sample: speed and timeliness.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
150 18-Oct-24
151

150 151

1
 18-Oct-24

Sample planning process Sample planning process

 Sample design: Specifying a sample  Sample size: The number of elements to be

design involves both theoretical and included in a study.
practical considerations, such as: cost, Important qualitative factors to be considered
time, labor involved, and organization. in determining the sample size include: (1) the
Researcher should consider: Survey importance of the decision; (2) the nature of
objectives; eligibility criteria for survey the research; (3) the number of variables; (4)
respondents; exclusion criteria rule out;
the nature of the analysis; (5) sample sizes
sampling method; type of sample;
sampling unit; sampling frame; handling used in similar studies; (6) incidence rates; (7)
refusals and nonresponse. completion rates; and (8) resource constraints.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
152 18-Oct-24
153

152 153

Sample planning process Sample planning process

Costs of sampling: The plan must  Execution of the sampling process:
account for the estimated costs of Sample needs to be representative and to
sampling. There are two traditional adequately mirror the various patterns and
cost categories: subclasses of the population. The sample size
should be large enough to provide confidence
 Overhead costs: relatively fixed for a
in the stability of its characteristics.
sampling procedure.
There two different sampling procedures:
 Variable costs: depend on the number of
 Non-probability sampling procedures
respondents contacted and interviewed in
the study.  Probability sampling procedures
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
154 18-Oct-24
155

154 155

3.1.2. Probability and non- 3.1.2. Probability and non-

probability sampling procedures probability sampling procedures
Probability sampling Probability sampling methods
 Definition of probability sampling: Each  The simple random sample
sampling unit in the defined target
population has a known probability of  The systematic sample
being selected for the sample.  The stratified sample
 The cluster sample
 The area sample: single stage and
multistage

18-Oct-24
156 18-Oct-24
157

156 157

2
 18-Oct-24

3.1.2. Probability and non- 3.1.2. Probability and non-

probability sampling procedures probability sampling procedures

Non-probability sampling Non-probability sampling methods

 Definition of non-probability sampling:  The quota sample
Sampling designs in which the probability  The judgment sample
of selection of each sampling unit is not
known. The selection of sampling units is  The convenience sample
based on the judgment of the researcher  Snowball sampling
and may or may not be representative of
the target population.

18-Oct-24
158 18-Oct-24
159

158 159

Q&A

3.2. Data collection

 Questionnaire design
 Measurement and scaling

NGUYEN HOANG GIANG – Principle of Marketing Department 161

18-Oct-24
160 18-Oct-24

160 161

3.2.1. Questionnaire design 3.2.1. Questionnaire design

Questionnaire definition
Questionnaire definition
Questionnaire is a tool for collecting
Objectives of a questionnaire information to describe, compare, or
Questionnaire design process explain an event or situation, as well as,
Confidentiality and ethics knowledge, attitudes, behaviours, and/or
socio-demographic characteristics on a
particular target group.

NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
162 18-Oct-24
163

162 163

3
 18-Oct-24

3.2.1. Questionnaire design 3.2.1. Questionnaire design

Objectives of a questionnaire Questionnaire design process
 It must translate the information needed into
questions that the respondent can and will
answer.
 It must motivate and encourage the respondent
to co-operate and complete the interview.
 It must minimise error.
 Questionnaire design is a major source of non-
sampling error.
 Respondent error:
 Non-response errors
 Response errors

NGUYEN HOANG GIANG – Principle of Marketing Department

18-Oct-24
164 18-Oct-24
(Source: Carl McDaniel, Jr. and Roger Gates, 2012) 165

164 165

Questionnaire design process Step 1: Preliminary considerations

1. Preliminary considerations The researcher must make decisions regarding:

2. Question content  What information is required?
 Who are the target respondents?
3. Response format
 What method of communication will be used to
4. Question wording reach these respondents?
 How will the resulting data be processed and
5. Question sequence
analysed?
6. Physical characteristics

7. Pre-test
NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
166 167

166 167

 What information is required?  Who are the target respondents?

o There is need for good problem definition and o The characteristics of the respondent groups
clear objectives for the study to deal with the have a great influence on questionnaire design.
problem at hand. o Questions appropriate for specialists will not be
o Research objectives (listing of information appropriate for a layman.
needs) should be spelt out as clearly and o A questionnaire must speak to the respondents
precisely as possible. in an understandable language and at the
o The questionnaire is the link between the appropriate intellectual level.
information needed and the data to be collected. o The more heterogeneous the groups, the more
o difficult it is to construct a single questionnaire
that is appropriate to everyone.

168 169

168 169

4
 18-Oct-24

 What method of communication  How will the data be processed

will be used? and analysed?
o The method of administering a questionnaire will o Decisions regarding editing, coding,
have a bearing on the structure and content. processing and statistical analysis of the data
o In person administered surveys the respondents need to be factored in to the design of the
will see the structure and interact with the
interviewer. questionnaire.
o Therefore more complex and difficult ?ns can be o E.g. where statistical analysis is to be
used. employed, highly structured questions are
o In telephone interviews the respondent does not needed.
see the questionnaire.
o Shorter questionnaires are more ideal in such cases.
o Postal questionnaires are self completing.

170 171

170 171

Step 2: Question content Step 3: Response format

When evaluating possible questions for inclusion in Once the content of individual questions has
the data collection instrument the following issues
must be addressed; been determined, the researcher needs to
 Is the question really necessary? decide the type of question to use.
 Does the respondent have the information Three basic types of questions may be used:
requested?
 The respondent may not know the answer.  Open ended
 Can the respondent remember/recall the information:  Multi choice ( multichotomous)
omission, telescoping, creation.
 Dichotomous
 Willingness to respond accurately (non-response
error).  Each poses a degree of structure on the
 Response errors: faulty recall due to selective person’ s responses and has its own particular
perception, time compression, averaging. uses, pros and cons.
172 173

172 173

 Open ended questions  Multi choice questions

o The respondent can reply in his or her own words. o Presents the respondent with a set of mutually
o The researcher does not limit the respondent’s exclusive and collectively exhaustive
choices, alternatives.
o They are useful at the beginning of a questionnaire o Mutually exclusive (no overlap occurs, only one
or alternative can be selected).
o When there are many possible responses to be • Less than 15 ----
listed or • 15 to 20 ----
o When the researcher is not sufficiently aware of all • 20 to 25 ----
possible responses.
o Collectively exhaustive (alternatives include all
o E.g. why do you shop here?
possible responses).

174 175

174 175

5
 18-Oct-24

 Dichotomous questions Class activity

o An extreme form of the multi choice question. Merits and demerits of multi choice
o Allows for only two response categories. questions also apply to dichotomous
• e.g. yes and no, like and dislike, agree and
questions.
disagree.
o However the two alternatives of interest are often Highlight the merits and demerits of the
supplemented by a neutral alternative. following types of questions:
• e.g. both, neither, don’t know Open ended
o Note: Most questionnaires mix open-ended, Multi choice
multiple choice and dichotomous question Dichotomous
formats.
176 177

176 177

Step 4: Wording the question Step 4: Wording the question

It is critical that the researcher and

respondent assign the same meaning to
the questions asked, otherwise the results
will be seriously biased.
Poor phrasing can cause respondents to
refuse to answer (item non-response),
Or answer incorrectly either on purpose or
due to misunderstanding (measurement
error).
178 179

178 179

Designing the wording of a question Designing the wording of a question

1. The question should clearly define the issue 5. Avoid double barrelled questions (wording calls
being addressed. for two responses)e.g. do you like the taste
2. Use simple, clear, unambiguous words. and aroma of the coffee?
3. Avoid leading questions ( suggest or imply 6. Consider using both positive and negative
certain answers) e.g. statements e.g. Do you think TV advertising
Do you think that the food in the hotel made you sick? should be increased or decreased during
Did the hotel staff seem unhygienic to you? Christmas.
Do you agree that the hospital staff were close to 7. Avoid generalisations and estimates. E.g. How
exhaustion? many litres of milk do you purchase in a year.
4. Avoid loaded questions (suggest social Rather ask how many pints of milk did you
desirability). purchase this last week?
180 181

180 181

6
 18-Oct-24

Step 5: Question sequence Step 5: Question sequence

(flow and layout) guidelines (flow and layout) guidelines
 Always start with a preamble The next questions should be simple and
 The first few questions are normally interesting.
screening/filter questions. They are referred to as warm up
 These are used to determine whether the questions.
respondent fits the predetermined selection They are designed to put the respondent
criteria to participate in the survey. at ease and to encourage co-operation.
 E.g. have you ever stayed at this hotel They may not be of primary concern to
before? the researcher.

182 183

182 183

Step 5: Question sequence Step 5: Question sequence

(flow and layout) guidelines (flow and layout) guidelines
Within the main body of the questionnaire,  Group related questions/topics together e.g.
the researcher should: section A, section B etc.
 Within a section, move from general to specific.
Use transition statements to inform the
respondent of what is ahead (i.e. moving  Ask questions about present behaviour before
examining past behaviour.
on to a new section/topic).
 If possible, try to place together questions that
E.g. Next I will ask you questions are similar in format.
regarding your leisure activities.  Use skip questions (branching questions) where
This helps respondents switch their train appropriate e.g. If you answered NO in Q5,
of thought. please go to Q7.

184 185

184 185

Step 5: Question sequence Step 6: Physical characteristics

(flow and layout) guidelines (Physical appearance of the questionnaire)

Complicated and difficult to answer questions  “Looks are everything”

are best placed deep in the questionnaire.  The format, spacing and positioning of
Sensitive questions are best placed near the questions can have a significant effect on the
end of the questionnaire. results.
Place questions of a classificatory nature in  Appearance and layout of the questionnaire is
the final section at the end of the a major determinant of the response rate,
questionnaire e.g. demographic information especially in self administered questionnaires.
such as age, gender, income, occupation,
education level etc.
Thank the respondent for their co-operation.
186 187

186 187

7
 18-Oct-24

 Design of questionnaire  Considerations of layout and design

Masern-Ausbruch Nordfriesland Juni 2001
Questionnaire should be as short as possible.
Vom Befrager einzutragen: Datum der Befragung: __.__._____ Initialen des Befragers: ____
Name des Arztes: _____________________
Dr. Gerard Krause Befragung vollständig, keine weiteren Anrufe oder Nachfragen notwendig!
Abt. Infektionsepidemiologie
Robert Koch-Institut Guten Tag, mein Name ist _______________: Ich bin Mitarbeiter im Kreisgesundheitsamt in Husum /

It should look as professional as possible

Seestrasse 10 Als Mitarbeiter des Robert Koch-Institutes unterstütze ich das Kreisgesundheitsamt Husum. Wir haben
13353 Berlin
derzeit im Landkreis Nordfriesland eine ungewöhnliche Häufung von Masernerkrankungen. Wir möchten
Vom Befrager einzutragen: uns zunächst herzlich dafür bedanken, dass Sie uns Ihren Patienten
____________________________(Name des Falles) geboren am ______ als Masernfall gemeldet haben.

(printed on good quality paper).

- Datum der Befragung: ___.___._____
- Initialen des Befragers (Vor-+ Nachname): ___ ___ Wir führen derzeit eine epidemiologische Untersuchung durch um diese Masernhäufung eindämmen zu
- Name des Arztes: ________________________________ können. Das Erfassen von Masernkomplikationen ist hier besonders wichtig. Vor kurzem müssten Sie
- Befragung vollständig? (1)  Ja (2)  Nein, Nachfrage notwendig.
von uns ein Fax mit weiteren Erläuterungen zu dieser Häufung erhalten haben, in dem wir auch auf diese
Nachrecherchen hingewiesen haben. Um den von Ihnen gemeldeten Fall korrekt einordnen zu können,

A crowded questionnaire appears complex

möchten wir noch kurz einzelne Informationen abklären.
Guten Tag, mein Name ist ___ (bitte Namen nennen)___.
Ich bin Mitarbeiter im oder Als Mitarbeiter des Robert Koch-Institutes Fragenkomplex Symptome:

and difficult to read & can lead to errors in

Kreisgesundheitsamt in Husum unterstütze ich das Kreisgesundheitsamt Husum.
Hatte Ihr Patient eines der mehrere der folgenden Symptome?
Wir verzeichnen derzeit im Landkreis Nordfriesland seit Anfang September 2005 eine ungewöhnliche
Häufung von Masernerkrankungen. Wir möchten uns zunächst herzlich dafür bedanken, dass Sie uns Generalisierter Hautausschlag > 3 Tage: Ja  Nein  Unbekannt 

data.
Ihren Patienten ___(Name des Falles wie in Kopfzeile)___, geboren am __(Datum wie in Kopfzeile)__, Wenn Beginn des Hautauschlags exakt bekannt ist, Datum genau angeben: __.__.____
als Masernfall gemeldet haben. Wir führen derzeit eine epidemiologische Untersuchung durch um diese Wenn Beginn des Ausschlages NICHT exakt bekannt ist, dann Zeitpunkt eingrenzen, nämlich:
Masernhäufung eindämmen zu können. Das Erfassen von Masernkomplikationen ist hier besonders Beginn Ausschlag irgendwann zwischen dem __.__.____ und dem __.__.____
wichtig. Um den von Ihnen gemeldeten Fall korrekt einordnen zu können, möchten wir noch kurz einzelne
Fragenkomplex Labor:

Directions and instructions for questions

Informationen abklären.
Liegt Ihnen zu diesem Patienten einer oder mehrere der folgenden Labornachweise vor?
1. Hatte Ihr Patient eines der mehrere der folgenden Symptome?
Generalisierter Hautausschlag > 3 Tage: (1)  Virusisolierung (2)  Nukleinsäurenachweis (PCR)

should be placed as close as possible to the

(1)  Ja* (3)  IgM (4)  IgG mit 4-fachem Titeranstieg

(2)  Nein (5)  Antikörpernachweis mit 4-fachem Titeranstieg

(3)  Unbekannt
War Ihr Patient gegen Masern geimpft?

particular question.
Erste Masernimpfung:
* Wenn ja: Wann begann der Hautauschlag?
 Ja (Impfausweis)  Ja (aber nicht dokumentiert) wenn ja, wann (z.B. laut Impfausweis): __.__._____
(4)  Datum genau bekannt: ___.___.2005  Nein  Unbekannt
(5)  Datum nicht genau bekannt: Auschlag begann zwischen ___.___.2005 und ___.___.2005.
Zweite Masernimpfung:
 Ja (Impfausweis)  Ja (aber nicht dokumentiert) wenn ja, wann (z.B. laut Impfausweis): __.__._____

Instructions should be distinguished from the

2. Welche der folgenden Labornachweise liegen Ihnen zu diesem Patienten vor?
(Mehrfachnennungen möglich)  Nein  Unbekannt

(1)  Virusisolierung Dritte Masernimpfung:

(2)  Nukleinsäurenachweis (PCR)  Ja (Impfausweis)  Ja (aber nicht dokumentiert) wenn ja, wann (z.B. laut Impfausweis): __.__._____

questionnaire ( i.e. instructions can be in

(3)  IgM
 Nein  Unbekannt
(4)  IgG mit 4-fachem Titeranstieg
Arzt hat in der Praxis noch andere Masernfälle gehabt, die noch nicht gemeldet wurden:
(5)  Antikörpernachweis mit 4-fachem Titeranstieg
 Ja  Nein

capital letters or bold)

Wenn ja, dürfen wir Sie bitte uns diese Fälle schnellst möglich alle nachzumelden, denn davon hängt ab,
3. War Ihr Patient gegen Masern geimpft? (bitte eine Angabe pro empfohlene Impfung)
wie realistisch wir diese Häufung einschätzen können.
Erste Masernimpfung: Zweite Masernimpfung: Dritte Masernimpfung: Haben Sie bei Ihrem Patientengut subjektiv den Eindruck, dass im Vergleich zum entsprechenden
Zeiträumen der Vorjahre dieses Jahr deutlich mehr Masernfälle aufgetreten sind?  Ja 
(1)  Nein (1)  Nein (1)  Nein Nein  unbekannt
(2)  Ja* (nicht dokumentiert) (2)  Ja* (nicht dokumentiert) (2)  Ja* (nicht dokumentiert) Sie werden in wenigen Tagen auf dem Postweg weitere Informationen zu der Masernhäufung von uns
(3)  Ja* (Impfausweis) (3)  Ja* (Impfausweis) (3)  Ja* (Impfausweis) erhalten, Haben Sie hierzu Fragen?
Vielen Dank für ihre Mithilfe!

188 189

188 189

 Considerations of layout and design Step 7: Pre-testing the questionnaire

Numbering of questions is important if  No survey should be taken without a pre-test.

responses are to be coded e.g.  The trial run of the questionnaire should
 Male (1) ----- simulate as closely as possible the actual
 Female (2) ______ research conditions under which it will be
Open- ended questions require a reasonable administered.
amount of space to ensure eligible writing.
 Pre-test respondents should be taken from the
It is a good idea to spend time phrasing the title population of interest.
of the questionnaire, because a well worded
title can arouse interest amongst respondents.  A pre-test should be conducted in the same
survey mode as the actual survey.

190 191

190 191

Confidentiality and ethics 3.2.2. Measurement and scaling

 Allow for privacy and do not ask questions Overview of measurement process
which may offend, or ask for data that is not
essential. Apart from anything else, the Construct development
response rate will suffer. Scale measurement
 In surveys promises of confidentiality are Evaluating measurement scales
often made to the respondents to reassure
and encourage replies. Developing scale measurements
 Aside from the legal issues, it is unethical to Scales to measure attitudes and
obtain data from respondents by misleading behaviors
them about the survey purpose and the Other scale measurement issues
method of analysis.
NGUYEN HOANG GIANG – Principle of Marketing Department
192 18-Oct-24
193

192 193

8
 18-Oct-24

Overview of measurement
process Construct development
Measurement is the process of developing Construct: A hypothetical variable made up of
methods to systematically characterize or a set of component responses or behaviors
quantify information about persons, events, that are thought to be related.
ideas, or objects of interest. Marketing constructs must be clearly defined.
Measurement is essential to effective decision Construct development is the process in which
making so the popular scales that measure researchers identify characteristics that define
attitudes and behavior should be consider. the concept being studied by the researcher.
Once the characteristics are identified, the
The measurement process consists of two tasks:
researcher must then develop a method of
construct selection/development and scale
indirectly measuring the concept.
measurement.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
194 18-Oct-24
195

194 195

Scale measurement Scale measurement

Scale measurement: The process of assigning  Nominal scale: The type of scale in which the
descriptors to represent the range of possible responses questions require respondents to provide only some
to a question about a particular object or construct. type of descriptor as the raw response.
The scale descriptors are a combination of labels, such  Ordinal scale: A scale that allows a respondent to
express relative magnitude between the answers to a
as “Strongly Agree” or “Strongly Disagree” and question.
numbers,...
 Interval scale: A scale that demonstrates absolute
Scale points: Designated degrees of intensity assigned differences between each scale point.
to the responses in a given questioning or observation  Ratio scale: A scale that allows the researcher not only
method. to identify the absolute differences between each scale
All scale measurements can be classified as one of four point but also to make comparisons between the
basic scale levels: nominal, ordinal, interval, and ratio. responses.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
196 18-Oct-24
197

196 197

Evaluating measurement Construct/Scale development

scales process
Scale reliability refers to the extent to which a scale
can reproduce the same or similar measurement
results in repeated trials. Thus, reliability is a
measure of consistency in measurement. Random
error produces inconsistency in scale
measurements that leads to lower scale reliability.
Two of the techniques that help researchers
assess the reliability of scales are:
 Testretest
 Equivalent form.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
198 18-Oct-24
199

198 199

9
 18-Oct-24

Developing scale Scales to measure attitudes

measurements and behaviors
 Criteria for scale development: Questions Likert scale: An ordinal scale format that asks
must be phrased carefully to produce accurate respondents to indicate the extent to which they
agree or disagree with a series of mental belief or
data. To do so, the researcher must develop behavioral belief statements about a given object.
appropriate scale descriptors to be used as the Semantic differential scale: A unique bipolar
scale points. ordinal scale format that captures a person’s
Designing measurement scales requires attitudes or feelings about a given object.
 Understanding the research problem. Behavioral intention scale: A special type of
rating scale designed to capture the likelihood that
 Establishing detailed data requirements. people will demonstrate some type of predictable
 Identifying and developing constructs. behavior intent toward purchasing an object or
 Selecting the appropriate measurement scale. service in a future time frame.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
200 18-Oct-24
201

200 201

Scales to measure attitudes Scales to measure attitudes

and behaviors and behaviors
Noncomparative rating scale: A scale format Rank-order scales: These allow respondents to
that requires a judgment without reference to compare their own responses by indicating their
another object, person, or concept. first, second, third, and fourth preferences, and so
Comparative rating scale: A scale format that forth.
requires a judgment comparing one object, person,
or concept against another on the scale. Constant sum scales: Require the respondent to
allocate a given number of points, usually 100,
Graphic rating scales: A scale measure that among several attributes or features based on their
uses a scale point format that presents the
importance to the individual; this format requires a
respondent with some type of graphic continuum
as the set of possible raw responses to a given person to evaluate each separate attribute or
question. feature relative to all the other listed ones.
NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
202 18-Oct-24
203

202 203

 Other scale measurement

issues Q&A
Single-item scale: A scale format that
collects data about only one attribute of an
object or construct.
Multiple-item scale: A scale format that
simultaneously collects data on several
attributes of an object or construct.

NGUYEN HOANG GIANG – Principle of Marketing Department NGUYEN HOANG GIANG – Principle of Marketing Department
18-Oct-24
204 18-Oct-24
205

204 205

10