Please enable JavaScript.
Coggle requires JavaScript to display documents.
Research Methodology 2 - Coggle Diagram
Research Methodology 2
Sampling
Probabilistic (probability of drawing units known in advance, not necessarily equal) - Usually quantitative, to generalise
-
Multistage Samples (when we do not have a full list of the universe): First, divides into clusters, which are further divided into subclusters, in which units are randomly chosen.
- Draw Primary Sampling Units (PSU) similar to in stratified samples, whose selection probability is known in advance. Might have multiple stages
- Desired number of cases is drawn within each PSU.
Simple random and systematic: Every unit exact same chance, when they cannot be drawn by a machine -> Systematic. (known & equal)
-
Coverage Error: Main error in sampling, whenever we have a sample that does not accurately represent the universe.
-
-
-
-
Questionnaires
Types of questions
-
Closed responses
Scaled: ordinal, interval or ratio -
Likert scales: Ordinal - 5-7 options with a neutral value, no true zero point.
-
Ratio: Ratio - With a zero point, meaning zero quantity of what is being measured
Categorical: Nominal - Yes/No, Male/Female etc.
-
Questionnaires P.2
Administering them
Interviews
Telephone Surveys: More time and cost efficient, for short-medium length interviews. High coverage error.
Face-to-face: Complete, longer questionnaires. Perfect with a probabilistic sampling method. Requires big staff.
Semi-structured: Medium of all. → Suitable for particular populations (elderly, or children)
Unstructured:Topics are specified in detail, it contains the same topics that change according to the situation, the interviewee answers with her/his own words (no predefined answer).
Structured: Structured: Topics specified in a general way, always same questions with predefined answers. Usually proposed to a large sample as a survey.
-
Self-administered
-
Paper & Pencil: Small staff, more willingness to answer controversial topics. Non-representative
Pilot testing
- Examine the completed questionnaire and make amendments
- Administer the questionnaire
- Ask people from the target pop. to examine it
The role of time
-
Cohort studies: Same population over time, different units
-
Cross-sectional surveys: One-time, particular population and sample
-