Descriptive & Correlational Methods 3: Survey Research
A common method, but…
1. Begin with clear purpose
2. Question format (open or closed)
3. Writing good questions (10 guidelines)
4. Writing response options (4 guidelines)
Based on simple idea:
If you want to learn about people’s thoughts, feelings, and behaviour, then ask.
A common method (based on a simple idea), so it must be a simple method? No!
Part 1: Begin with Clear Purpose
Who do we want to study (also see next week: sampling)
What do we want to study? Constructs, operational definitions to measure variables?
Why conduct the study? Description? Correlations? Theory and hypothesis testing?
Next Step: construct the survey questionnaire (parts 2, 3, & 4)
Part 2: Decide on Question Format
A. Open-ended questions (answer in own words)
Ask them: “what is the most important problem facing the country today?”
Allow maximum freedom in how people respond; no constraints
Useful in preliminary stages of research
Difficult to analyze responses
Responses must be content analyzed (W&M 207-208)
Questions can be difficult to answer
B. Closed-ended questions
a. Check-lists (e.g., select one problem from list provided)
b. Rankings (e.g., rank list of problems)
c. Rating Scales (more quantitative info). E.g.,
Attitude Scales (evaluative responses):
Likert Scale : Tuition should be Frozen (agree/disagree scale)
Disagree 1—2—3—4—5—6—7 Agree
Semantic Differential scale: A tuition freeze is:
Use bipolar adjective scales
Bad 1—2—3—4—5—6 Good 2. Other decisions (with rating scales)
a. How many response options should we use (2, 20, 2000)?
Difficult to distinguish between more than “7 + or – 2” options (so use 5-9 options)
b. Should options be labeled? E.g.
Disagree -2 -1 0 +1 +2 Agree
strongly slightly neither slightly strongly
Labels clarify the meaning of scale values;
But might be confusing if scale can’t be seen (e.g., phone survey)
Simplify this with a branching format (phone survey):
Ask: Do you agree, disagree, or neither?
If agree (or disagree), branch to more detailed questions
Do you agree strongly, moderately, or slightly?
See page 223 for another example of “branching”
c. Conduct a “pilot test”: try different formats and ask which is easiest to use
Important in all research (identify and solve problems before we conduct the main
C. Question format can influence responses
Participants were asked: What is the most important problem facing the country today?
Some had to answer closed ended:
Quality of public schools 32%
Legalized abortion 8%
The energy crisis 6%
Other participants were given open ended response format:
Public Schools: 1%
Energy crisis 0%
Do pilot with open-ended format; then use closed-ended format for main survey
Part 3: Writing the Questions
1. Use appropriate complex (not too complex)
When making for adults, aim for language that average 14 year old would understand
2. Avoid technical jargon (conversion disorder?) – most people don’t know what that is
3. Provide clear definitions of terms, even if common (e.g.,