Term
Cross-sectional vs. longitudinal studies
|
|
Definition
- Cross-sectional studies
- Static snapshot
- Slice of population at one time
- Inherent limitations:
- Inability to capture change over time
- Making casual inferneces is dangerous
- Longitudinal designs
- Multiple observations across time
- Tracking changes over time
- Testing how people respond to stimulus that occurs between observations
- Pretest-experimental treatment-posttest
- Testing for changes resulting from some intervening factor or event
- e.g., survey before and after election campaigns
|
|
|
Term
Identify: three designs of longitudinal studies
|
|
Definition
- Trend studies
- Cohort studies
- Panel studies
|
|
|
Term
Longitudinal design type 1: Trend study Advantages and limitations |
|
Definition
- Advantages of trend studies
- Describing long-term changes in a population
- Establishing a pattern over time
- Can be based on a comparison of survey data originally constructed for other purposes
- Limitations of trend studies
- Ambiguous starting point
- Inability to capture individual change over time
- Changes in measurement schemes or the way questions were asked produce non-comparable results over time
|
|
|
Term
Longitudinal design type 2: Cohort study Advantages and limitations |
|
Definition
- Advantages of cohort studies
- Tracking changes in a group as they change
- Birth cohort is one example of "cohort"
- e.g., studying a sample of people born in 1940 every 10 years
- Attempt to identify a unique effect from age cohort (people born in 1940)
- Cannot be examined with a cross-sectional design
- Limitations of cohort studies
- Age effects/Period effects/Cohort effects
- These effects are often confouned
- Respondent mortality
|
|
|
Term
Longitudinal design type 3: Panel study Advantages and limitations |
|
Definition
- Advantages of panel studies
- Interview the same sample of respondents at different points in time
- Can examine changes in respondent' thoughts and behaviors over time
- e.g.,Studying the same sample of people during a campaign cycle
- Can reveal shifting attitude: Election candidate opinion
- Cannot be examined with a cross-sectional or cohort study design
- Can examine how people react to changes in the environment
- e.g., Examining effects of the anti-smoking health campaign --comparing respondents' knowledge/attitude before and after the campaign
- Limitations to panel studies
- Sensitization effects
- Panel members are often difficult to recruit
- The problem of respondent mortality
- Requires a large sample for the first wave data
- Possibility of bias in the panel sample -- Are those who drop out different?
- Expensive
|
|
|
Term
| Time dimension and causality in longitudinal research |
|
Definition
- Conditions of causality:
- Time order: cause must precede effect
- Correlation between cause and effect
- All other alternative causes must be ruled out: absence of third factors that make the relationship between cause and effect spurious
|
|
|
