Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
42 Cards in this Set
- Front
- Back
Research design is intended to test if one variable:
|
1) causes another
2) causes change in another 3) requires 3 criteria |
|
The 3 criteria required for causality are:
|
1) covariation
2) time order 3) nonspuriousness |
|
---------- is evidence that two phenomena tend to occur at the same time or for the same cases.
|
COVARIATION
ex. everytime there is a crisis in foreign policy, the presidential popularity increases *all types of resarch deisgn are intended to determine whether causatin exists and are set up to measure the extent of covariation.* |
|
---------- must have evidence the assumed cause (the independent variable) happened before the assumed effect (the depenedent variable)
|
TIME ORDER
|
|
------------- makes sure that any covariation we observe between the independent and dependent variable is not caused by other factors.
|
NONSPURIOUSNESS
|
|
What are the four types of research design?
|
1) true experimental design
2) the quasi experiment (natural) 3) correlation design 4) designs lacking covariation |
|
How does the true experimental design start off?
|
start with two groups (unit of analysis), divided randomly into a treatment group and a control group which is given a stimulus. both groups given posttest, results compared to hypothesis.
|
|
What are some problems associated with the true experimental design?
|
Impossible to manipulate many independent variables (age, race, gender or social phenomena like war or elections) Lack of representative samples (conducted in artificial settings, no control on outside influences like news, conversations, or experience)
|
|
This is the comparison of the dependant variable before and after the independent variable has been applied. The subjects in this case are not assigned to groups.
|
QUASI "NATURAL" EXPERIMENT
|
|
In this experiment, researchers observe which subjects have something happen to them and then they put them into experimental and control groups. This requires a pretest of dependent variables so the amount of change is measured.
|
THE QUASI "NATURAL" EXPERIMENT
|
|
This is the collecting of data on an independant and dependant variable, and determining rather or not there is a pattern or relationship.
|
CORRELATION DESIGN
|
|
In a correlation deisgn, it is also advisable to what?
|
Collect data from other potentially relevant variables and control for them
This is the most common approach in politcial science. |
|
This is descriptive work and it may serve to increase our knowledge but it can't "prove" anything in a scientific sense.
|
DESIGNS THAT LACK COVARIATION
|
|
What are the two basic types of survey items?
|
Closed ended and open ended.
|
|
This is a survey when respondents are given all possible answers.
|
CLOSE ENDED
|
|
This is a survey when respondents are given a general question and asked to provide their own answer.
|
OPEN ENDED
|
|
Are most surveys open or close ended and why?
|
CLOSED
EASIER AND CHEAPER |
|
What are some advantages to open ended survey items?
|
Better measurements of opinion, attitudes, and concerns
People may choose just for the heck of it on close ended... On OPEN ended, they can express their real feelings. |
|
What are some disadvantages to close ended survey items?
|
More difficult to record responses
More difficult to classify responses |
|
What is a problem associated with survey item construction?
|
The respondents must be competent enough to answer a question.
-You can present a summary of difficult info. before asking the questions -Use a filter question to determimne if the respondent is familiar |
|
In survey construction, what are you to avoid?
|
Using biased or emotional language.
--Do not allow phrasing of the question to sway respondents. Avoid leading questions --Alwyas ask "do you agree or disagree..." Stating the question in a negative way |
|
In survey construction, what is best?
|
Short and simple questions
--Longer questions are harder to understand --If it's a complicated issue, provide background then ask a simple question |
|
Problem in survey construction:
Never use big words that are unfamiliar to the average person. If you must... |
If you must use them, explain them first.
|
|
Problem in survey construction:
Avoid ambigous questions. Be what... |
Be as clear as possible.
|
|
Problem in survey construction:
Avoid or minimize threats. Threats may what... |
If threats are personal or societal, they may lead to less than honest questions, if any at all.
ex. do you use illegal substances? do you want to keep homosexuals out of ur neighborhood? |
|
How do you minimize the threats in survey construction?
|
Stress confidentiality and anonymity
Change your language --"would a less traditional family living next door bother you?" |
|
Problem in survey construction:
To avoid double barrelled questions, |
Ask only one question at a time.
|
|
Published data is usually organized by what?
|
UNITS OF ANALYSIS
--Be sure that published info. is for the unit of analysis needed |
|
Most political science data is ----.
|
AGGREGRATE
--summary figures on population of geographic political units |
|
When analyzing units of analysis, what is something to avoid?
|
ECOLOGICAL FALLACY
--do not draw conclusions about individuals from aggregarate info ex. study says all ISU students are alcoholics. im in rantoul, person assumes im an alcoholic b/c i have an ISU shirt on. |
|
When dealing with political science data, you should ---- the information.
|
STANDARDIZE
--ex. convert into percentages of total |
|
In units of analysis, "individual" information is sometimes available. This means..
|
Not usually for "ordinary" people, rather senators, etc.
May include gov't agencies, political parties, etc. |
|
This is any technique for making inferences by objectively and systematically identifying specified characteristics of messages.
|
CONTENT ANALYSIS
--this is a data collection method, not a research design |
|
Content analysis can be used to analyze...
|
Published verbal texts
--is news coverage of a political campaign biased |
|
Content analysis also provides...
|
Answers to open ended survey questions
Historical perspectives |
|
One issue of content analysis is ----.
|
OBJECTIVITY
Subjective placement of data into categories |
|
To reducse objectivity and subjective placement, ------.
|
Have more than one person do the coding and compare results.
--intercoder reliability is the extent of similarity between difference coders |
|
Contigency tables are often presented as ------.
|
PERCENTAGES
|
|
What are the conventions of contigency tables?
Columns, rows.. |
Columns: INDEPENDENT VARIABLE
Rows: DEPENDENT VARIABLES --Calculate column percentages by dividing the frequency of each cell by the total for that column ---percentages for each column will = 100 |
|
The missing zero point is one way not to lie with graphics, what is this?
|
Exaggerates the difference between indepenedent variables categories.
|
|
Why is their a need for standardization of graphics?
|
-Aggregarate data only useful if standardizerd
-May otherwise be misleading |
|
Name some principles of good graphics.
|
-Keep it simple
-Clearly label and provide units -Provide title -Provide source -Describe the graphic in your text |