22.10 Survey Research Designs

Question 1

2. Steps in Survey Research?

Answer 1

Determine study purpose (RQ and HY)
Select research design
Choose sampling design (probability or non-probability) and determine sample size
Decide how to collect data (in person, by telephone, by mail, by internet)
Develop survey instrument
Collect data, analyze, report, interpret

Question 2

3. Research Type/Independent Variable/Purpose/ Approach

Answer 2

Research	Independent 	
Type		Variable	 Purpose	Approach
Survey
(Non-		Measured	Description	Survey
Intervention)			Explanation	Existing data

Evaluation
(Intervention) Manipulated Explanation Experimentation

Question 3

Uses of Non-Intervention Research

Answer 3

-Descriptive research
What way are things?
Identifies characteristics of the cases under study.
Can describe associations among variables.
-Explanatory research
Why are things the way they are?
Describes associations among variables and attempts to infer causal relationships.

Question 4

5. Evidence for Inferring Causality?

Answer 4

Association
Temporality
Non-spuriousness

Question 5

6. Survey Research Designs?

Answer 5

Cross-sectional: data collected at one time

Longitudinal: data collected at two or more times

Question 6

7. Cross-Sectional Research Designs?

Answer 6

Cross sectional

Data on a sample of respondents gathered at one time (short period of time)

Question 7

8. Cross-Sectional Design?

Answer 7

Common non-intervention research design
Data gathered from sample (‘cross-section’) of respondents chosen to represent target population
Gives a single ‘snap shot’ in time
Epidemiology uses cross-sectional design to estimate disease prevalence (prevalence studies)
Two variations: - contextual designs, - social network designs

Question 8

9. Contextual Designs?

Answer 8

Sample enough cases within particular groups or contexts to describe accurately certain characteristics of those contexts
-If the gender gap in math achievement favoring boys over girls might be explained by the contextual resources of their schools and their neighborhoods

Question 9

10. Social Network Designs?

Answer 9

Focus on the relationships among social actors and the transaction flows occurring along the connection links

• Interview of every person in the group under study
• Delineate networks of relationships
• Ex, women had to form costly, differentiated interactional networks by choosing (a) other women as friends and sources of support and (b) men mostly as instrumental resources. The men, in contrast, chose other men for both expressive and instrumental purpose

Question 10

11. Cross-Sectional Design – Causality?

Answer 10

Association
-Need enough cases to represent all levels of variables being measured
Temporality
-Difficult because only have information at one time ( investigators should make inferences about the logical relations among variables; respondents can be asked about both past and present events)
Non-spuriousness
-Must measure and control potential confounders, but but can never be sure all are accounted for

Question 11

12. Longitudinal Research Designs?

Answer 11

Trend
Data gathered from different samples of the population at two or more times (i.e., series of cross-sectional surveys)
Panel (epidemiology: prospective)
Data gathered from the same respondents at two or more times and the different sets of data are linked by respondent

Question 12

13. Trend Design?

Answer 12

Independent samples give responses at 2 or more data collection times
Variables stay the same each time
Provides multiple ‘snap-shots’ across time
Used to study change in variables over time
Useful for studying social changes - allows for study of change in the population as a whole over time

Question 13

14. Trend Design – Causality?

Answer 13

-Association
Same as cross-sectional design
-Temporality
Maybe for the whole population, but not for individuals
-Non-spuriousness
Same difficulties as cross-sectional

Question 14

15. Panel Design?

Answer 14

Same sample gives responses at each data collection time
Information is linked to the respondent over time
Allows for study of change in individuals over time
The cohort study in Epidemiology is more like panel design, not trend study suggested by the book

Question 15

16. Panel Design Causality?

Answer 15

-Association
Same as other designs
-Temporality
Because of following individuals over time, can refer to earlier data points to determine time sequence of events
-Non-spuriousness
Same difficulties as other designs
Be aware of attrition effects

Question 16

17. Major Disadvantages of Survey Research?

Answer 16

-Hard for inferring causality
-More rely on reports of behavior rather than observations of behavior
Measurement error may be produced by respondents’ lack of truthfulness, misunderstanding of questions, and inability to recall past event
-Susceptible to reactivity, which introduces systematic measurement error
Social desirable effects