Study Design à review different study designs à chart
Panel
- one group studied
- exposed to intervention
- Group measured before / after treatment
Quasi Panel
- similar to panel design
- but two separate groups used
- one before treatment / other after treatment
- used to avoid difficulty of keeping track of participants over time
Steps in Survey Research à are there other lists?
- set specific measurable objectives
- select research design
- choose population and sample
- develop reliable / valid survey instruments
- administer survey
- analyse survey data
- interpret / report
Sampling Methods à review all
Systematic sampling
- probability sampling method
- every nth
- requires good sampling frame
- not suitable for f2f where respondents over large area
- systematic bias if periodicity in sampling frame exists
- simpler than other probability methods
Availability sampling
- non probability
- individuals who are easily available and willing to participate
- may not be rep
- large samples
- pilot test / exploratory
Sample Size / Statistical Accuracy
- what happens to statistical accuracy if small sample size used
o statistical accuracy ( + / - ) will be larger, meaning less accurate.
Potential Problems With Questions
- term not defined
- ambiguous question
Types of Questions à review list
- behavioural
o measure what people do / what they have done
- leading
o question structure / wording makes respondents provide answer they would not have given had question been asked in more neutral way.
§ Do you agree that
§ Does this seem like a good idea to you.
Exhaustiveness Principle à list other principles
- response alternatives provide sufficient range of responses to cover all responses
Advantages of face to face interviews over paper surveys à list other types
- questionnaires can be reasonably complex
- accurate recording of responses
- misunderstandings can be clarified
- body language etc can be used to encourage participation
- probe
- be certain right person has answered.
Disadvantages of face to face over paper surveys
- expense
- possible interviewer bias
- tiredness / impatience
- social desirability considerations
- interviewer contamination
Coding à can be more than 1 code per answer
List wise deletion
- any case that has missing data on any of the set of variables deleted
Pair Wise
- use only cases with complete responses for each calculation
Validity à look at associated definitions
- how well question / scale measures what it is intended to measure
alpha =
0.95 - very high internal consistency
0.55 - reliability is unacceptable
0.86 - high internal consistency
0.63 - low internal consistency
No comments:
Post a Comment