3/3/26
Introduction to primary data
Chapter 7 – § What is Primary Data?
• Data collected directly by the researcher for a specific research
n Observation purpose.
Primary Data
• Unlike secondary data, primary data are original, tailored to your
Collection n Interviews & Diaries
research question, and require instrument design.
Questionnaires § Examples:
(Tool) n
üSurvey responses
üInterview transcripts
üObserved behavior
üDiary entries
1 2
1 2
1
3/3/26
When Do We Use Primary Data?
Types of Primary Data Collection Tools
q Use primary data when:
1. Observation
• No existing dataset answers your research question
• You need specific variables 2. Interviews
• You need attitudinal/behavioral data
3. Diaries
• You require up-to-date information
• You need experimental or field evidence 4. Questionnaires (Surveys)
3 4
3 4
2
3/3/26
OBSERVATION OBSERVATION
◦ Definition: Observation is a method of collecting data by
Type Description
systematically watching and recording behaviors, events, or
physical characteristics. Structured Predefined checklist
◦ When to Use Observation? Unstructured Open-ended recording
◦ Studying behavior (not attitudes) Participant Researcher participates
◦ When subjects cannot self-report accurately Non-participant Researcher observes only
◦ Field research
Overt Subjects know they are observed
◦ Retail studies
Covert Subjects do not know
◦ Classroom studies
5 6
5 6
3
3/3/26
OBSERVATION How to Conduct Observation – Example
v How to Conduct Observation RQ: “Does classroom layout affect student
participation?”
Step 1: Define behavior to observe
STEP 1: Define behavior: Number of times student raises hand
Step 2: Develop observation protocol (observation instrument)
STEP 2: Create checklist: | Student ID | Raised Hand (count) |
Step 3: Decide sampling (time, event, location)
STEP 3: Observe 3 classes with different layouts
Step 4: Record systematically
STEP 4: Record real-time
Step 5: Code data
STEP 5: Enter data → Run ANOVA
7 8
7 8
4
3/3/26
OBSERVATION INTERVIEWS
q Definition: An interview is a direct conversation between
q Advantages
ü Real behavior (not self-reported)
researcher and participant to collect in-depth information.
ü No recall bias q Types of Interviews:
ü Useful in natural settings
Type Description
q Disadvantages Structured Fixed questions
ü Time-consuming Semi-structured Guiding questions + flexibility
ü Observer bias Unstructured Open conversation
ü Ethical concerns (covert observation) In-depth Detailed exploration
ü Cannot measure internal attitudes Focus group Group discussion
9 10
9 10
5
3/3/26
INTERVIEWS How to conduct Interviews – Example
§ How to conduct Interviews? Research Topic: “Factors Affecting Student Satisfaction with
1. Define objectives
Online Learning”
2. Develop interview guide
3. Pilot test ../Thực hành/INTERVIEW STUDY_EXAMPLE.docx
4. Obtain consent
5. Record (audio/text)
6. Transcribe
7. Code responses
11 12
11 12
6
3/3/26
QUESTIONNAIRES (SURVEYS)
Advantages Disadvantages
INTERVIEWS
q Definition: A questionnaire is a structured instrument consisting
Rich data Time- q When to Use Interviews?
consuming of standardized questions used to collect quantitative data.
ü Exploratory research
Clarification Interviewer bias ü Understanding q Types of Questions:
possible motivations
Type Example
Flexible Small sample ü Complex issues Closed-ended Yes/No
size ü Qualitative studies Likert scale 1–5 agreement
Difficult to Multiple choice Select one
analyze Ranking Rank top 3
quantitatively Open-ended Explain opinion
13 14
13 14
7
3/3/26
QUESTIONNAIRES QUESTIONNAIRES (SURVEYS)
(SURVEYS) q How to design a Questionnaire
Advantages Disadvantages
Large data Low response rate q When to Use 1. Define research objective
Cost-effective Self-report bias Questionnaires? 2. Identify constructs (what variables)
(online) • Large samples
3. Operationalize variables (turn constructs into questions)
Statistical analysis Poorly designed • Quantitative analysis
possible questions cause
• Statistical modeling 4. Choose question types
error
• Hypothesis testing 5. Design structure
Standardized
6. Avoid bias / Avoid common question errors
7. Pilot test
15
8. Create a codebook 16
15 16
8
3/3/26
STEP 1: DEFINE THE RESEARCH OBJECTIVE STEP 2: IDENTIFY
Before writing any question, ask: What exactly do I want to
CONSTRUCTS (WHAT
measure? Objective Construct Type VARIABLES?)
Overall Dependent
Example: Satisfaction
satisfaction variable § Convert objective into
Research Topic: “Factors affecting students’ satisfaction with Teaching Instructor measurable constructs
Independent
online learning” quality support
Platform
Research Objective: Technology
quality
Independent
• Measure student satisfaction Control Gender, Year
Control
variable of study
• Identify influencing factors (technology quality, instructor support,
workload)
• Examine relationship between satisfaction and academic
performance 17 18
17 18
9
3/3/26
STEP 3: OPERATIONALIZE VARIABLES STEP 4: CHOOSE QUESTION TYPES
1. Closed-ended
Example Construct: Satisfaction
Operationalization: Use Likert scale (1–5) Gender: 1 = Male 2 = Female
2. Likert Scale: “The instructor provides clear explanations.”
Question items:
1–5 scale
1. I am satisfied with online learning.
3. Multiple Choice: “How many hours per week do you study
2. Online learning meets my expectations
online?”
3. I would recommend online courses to others
Less than 5 5–10 11–20 More than 20
Scale:
4. Open-ended : What improvements would you suggest?”
1 = Strongly Disagree 2 = Disagree
(Use carefully, avoid too many open-ended questions in
3 = Neutral 4 = Agree 5 = Strongly Agree
quantitative research)
=> This creates a measurable variable.
19 20
19 20
10
3/3/26
STEP 5: DESIGN STRUCTURE STEP 6: AVOID COMMON QUESTION ERRORS
q Questionnaire layout: §Double-barrel:
• Introduction
“Online learning is convenient and effective.”
• Consent statement
(What if convenient but not effective?)
• Screening question (if needed)
• Main questions §Leading:
• Demographics “Do you agree that online learning is superior?”
• Thank you note
§Ambiguous:
Example Introduction: “This survey aims to study student satisfaction with
online learning. Your responses are anonymous and will be used for “Do you often study online?” (Often = unclear)
academic purposes only.” 21 22
21 22
11
3/3/26
STEP 7: PILOT TEST STEP 8: CREATE A CODEBOOK
§ Test with 10–15 respondents. § Example:
§ Check: Variable Label Coding
ü Confusing wording? 1=Male,
gender Respondent gender
2=Female
ü Time needed?
sat1 Satisfaction item 1 1–5
ü Missing options? tech1 Platform quality 1–5
ü Revise before full data collection.
=> This will be essential for data processing.
23 24
23 24
12
3/3/26
QUESTIONNAIRES (SURVEYS) QUESTIONNAIRES - A full Example
q How to process survey data q Complete questionnaire example:
1. Data Entry ../Thực hành/QUESTIONNAIRE_EXAMPLE.docx
2. Data cleaning (Check for Missing values/Outliers/Impossible
values/Duplicate responses)
3. Code data properly (Convert text to numeric if necessary)
4. Check reliability (if multi-item scale: Cronbach’s alpha)
5. Create composite variables
6. Descriptive statistics
7. Hypothesis testing
8. Interpret results and Report results
25 26
25 26
13