SMPA1050 Exam2 questions with
answers passed
External vs internal validity for studies - correct answer ✔✔External:
- Does the same thing happen in other settings, like other labs and everyday settings?
- Generalizability: Is the study generalizable to the broader population you're interested in?
(doing a study on voting with high school students). Might sacrifice this to establish causality
Internal:
- Was the research done "right"?
Survey Research - correct answer ✔✔Example: Who are you voting for in the presidential
election?
Advantages: Generalizable assuming you've done a proper random sample (although there's a
margin of error), help us understand and describe
Disadvantages: Can't make strong causal claims b/c evidence about relationships tends to be
correlational not causal
Issues: Sampling has to be random (if it's a survey of GW students, every GW student has to
have an equal chance of being picked), sample size, matching sample to group you want to
speak about
Types of surveys used and survey questions - correct answer ✔✔Types: phone, mail, online, in-
person interview
Questions: open or closed-ended, but minor wording variation can matter (campaigns often test
messaging this way by asking one person a question a different way than another)
Question order important (context impacts responses)
Issues w/social desirability and survey setting influence responses
, Non-attitudes/memory - correct answer ✔✔People who have no idea what you're talking about
in a survey, but don't want to sound stupid so they answer anyway
Content Analysis - correct answer ✔✔Used to look into content to prove a point.
Example: Is there a liberal bias in the media?
Issues: What you analyze (newspaper, broadcast news), sample (just a paper like NYT or all
news?), quantitative vs qualitative analysis
Codebook and intercoder reliability - correct answer ✔✔Codebook: Variables of study laid out
and given to researchers
Intercoder reliability: Have multiple coders for your study. Have a good statistical test of
reliability to show how much coders agree and what the standard is to make sure the results
you get from coders are valid. Two people see the same story the same way during your study.
Not all the time, just usually. Are they coding enough of the studies the same way that we can
make claims and conclusions?
Experiments - correct answer ✔✔Used to see effects of a "dose" of something on people; if
there's a difference in two groups, it's b/c of a dose
Example: Strategy vs issue coverage on healthcare
Challenges: external validity-biggest (want to make sure it's not just one group of people, want
to make it as close to real life as possible. Example: SMPA students are going to give a different
answer than the rest of the country), random sampling
Advantages: Can establish causality b/c of random assignment and control over variables (good
internal validity)
Design of Experiments - correct answer ✔✔Experimental group- pretest, treatment, posttest
Control group-pretest, posttest
Posttest only- making sure the results aren't just because of treatment
answers passed
External vs internal validity for studies - correct answer ✔✔External:
- Does the same thing happen in other settings, like other labs and everyday settings?
- Generalizability: Is the study generalizable to the broader population you're interested in?
(doing a study on voting with high school students). Might sacrifice this to establish causality
Internal:
- Was the research done "right"?
Survey Research - correct answer ✔✔Example: Who are you voting for in the presidential
election?
Advantages: Generalizable assuming you've done a proper random sample (although there's a
margin of error), help us understand and describe
Disadvantages: Can't make strong causal claims b/c evidence about relationships tends to be
correlational not causal
Issues: Sampling has to be random (if it's a survey of GW students, every GW student has to
have an equal chance of being picked), sample size, matching sample to group you want to
speak about
Types of surveys used and survey questions - correct answer ✔✔Types: phone, mail, online, in-
person interview
Questions: open or closed-ended, but minor wording variation can matter (campaigns often test
messaging this way by asking one person a question a different way than another)
Question order important (context impacts responses)
Issues w/social desirability and survey setting influence responses
, Non-attitudes/memory - correct answer ✔✔People who have no idea what you're talking about
in a survey, but don't want to sound stupid so they answer anyway
Content Analysis - correct answer ✔✔Used to look into content to prove a point.
Example: Is there a liberal bias in the media?
Issues: What you analyze (newspaper, broadcast news), sample (just a paper like NYT or all
news?), quantitative vs qualitative analysis
Codebook and intercoder reliability - correct answer ✔✔Codebook: Variables of study laid out
and given to researchers
Intercoder reliability: Have multiple coders for your study. Have a good statistical test of
reliability to show how much coders agree and what the standard is to make sure the results
you get from coders are valid. Two people see the same story the same way during your study.
Not all the time, just usually. Are they coding enough of the studies the same way that we can
make claims and conclusions?
Experiments - correct answer ✔✔Used to see effects of a "dose" of something on people; if
there's a difference in two groups, it's b/c of a dose
Example: Strategy vs issue coverage on healthcare
Challenges: external validity-biggest (want to make sure it's not just one group of people, want
to make it as close to real life as possible. Example: SMPA students are going to give a different
answer than the rest of the country), random sampling
Advantages: Can establish causality b/c of random assignment and control over variables (good
internal validity)
Design of Experiments - correct answer ✔✔Experimental group- pretest, treatment, posttest
Control group-pretest, posttest
Posttest only- making sure the results aren't just because of treatment
