Transcript Introduction to Social Survey Methodology

Introduction to Social Survey
Methodology
Map Your Hazards!
Combining Natural Hazards with Societal Issues
Rules of Practice for Research
•
•
•
•
•
•
•
Appropriate Questions
Ethics, Credibility, Reliability, Validity
Technique and Design
Proper Sampling
Pilot Testing
Analysis
Specific Purpose for Study
– What are you trying to describe, explain or
explore?
2
Appropriate Questions & Means
• Collecting information about individuals:
– Attitudes, ideas, beliefs, opinions, feelings,
background, behavior, orientations or plans for the
future
• Questionnaires (Surveys) vs. Interviews
• Quantitative vs. Qualitative
– Systematic and structured questions (closed)
– Open-ended, more descriptive
3
Ethics, Credibility, Validity, Reliability
• Ethics – Protect privacy of individuals,
confidentiality, willingness to participate
• Credibility – Conveys purpose, how
worthwhile it is, and who is responsible
• Validity – Data test hypothesis, measure
what it claims to measure
• Reliability – Consistent measure and
representative
4
Population/Sample
• Who is your target population?
• Identify representative sample and make
generalizations
– Sample size: The larger the better. Good size is
between 100–250. Need at least 30 to be reliable
(including interviews)
– Random: known chance of inclusion
– Non-random measures:
• Snowball
• Quota (street survey)
• “Accidental” (student projects)
5
Research Design
• Relevance – What are you trying to discover (do
questions get at this)?
• Comprehensiveness – What are the independent
& dependent variables? How do you classify? Is
it clear?
• Aptness – Data that are readily coded and
analyzed
• Feasibility – Not too long or complicated
• Unambiguous – Are the questions and
categories exhaustive and mutually exclusive?
6
Question Formats
• Scales of measure (nominal, ordinal and
interval)
– Likert measures strength of motivations, attitudes,
opinions (e.g. Strongly Agree – Strongly Disagree)
– Checklists, rankings, ratings and degrees of
importance (e.g. scale of 1–5)
– Recurrent behavior (e.g. How often do you . . .)
– Coded categories (e.g. yes/no, male/female)
• Including “DK” or “Other” as optional
categories
• Open-ended (categories must be mutually
exclusive)
7
Pilot Study
• Must conduct a trial run with sample survey
to verify the design and fine-tune the
questions
– Looking for bias, clarity and spuriousness
• Everyone (even seasoned social scientists)
goes through this process
– Otherwise you end up with a lousy survey that will
not get a good response
8
Data Analysis
• What will the questions produce?
• Discursive (summarize in words)
• Graphs (bar charts of nominal and ordinal
variables)
• Histograms (continuous and interval data)
• Crosstabs (chi-square)
• Multivariate analysis (optional)
9