Term
| 5 threats of external validity |
|
Definition
1. Subjects not representative of Gen Pop
2. Artificial X
3. Limited Realization of X
4. Artificial circumstance
5. Testing X interaction |
|
|
Term
|
Definition
| the extent to which you can generalize findings from a study to the real world. |
|
|
Term
| 4 types experiment types (location/type) |
|
Definition
1. Lab
2. Field
3. Survey
4. Natural |
|
|
Term
| Vulnerability of Lab. experiment |
|
Definition
Artificial x (circumstances are fake)
External valid lower
subjects self select
treatment condition nonexistent in world |
|
|
Term
| Field Experiment Strength |
|
Definition
increased external val. over lab
natural setting
experimenter has control over who does/doesn't get treatment
Usually random assignment |
|
|
Term
| Strength/weakness of field study |
|
Definition
-fewer vulnerability in respect to artificiality and w/ respect to subject.
-Greater in terms of Generalization
-vulnerable to threats of internal valid |
|
|
Term
| 2 survey types & definition |
|
Definition
1. Survey Method: Goal improve survey method/figure how to access pub. opinion
2. Substantive: used to address substantive political science q's
2. Substantive |
|
|
Term
|
Definition
| Low external Validity. Artificial manipulation of X and artificial circumstance |
|
|
Term
|
Definition
| Random assignment comes from nature, or some actor, not the researcher (TV reception in mountain city), Military draft (failure of random assignment, lots of ways to get out of draft). |
|
|
Term
| Threats to Internal Validity/Problems of X (list) |
|
Definition
1. Complex X
2. Incomplete administration of X
3. Induced State Failure
4. Diffusion of X
5. Intra session history confound |
|
|
Term
|
Definition
Threat to internal validity. Different parts to the X, how do we tell which part of x was cause? Ex. Placibo effect
(R) X O drug, and pill (two things!)
(R) O |
|
|
Term
|
Definition
| the difference between the control and treatment |
|
|
Term
|
Definition
| Subjects do not know if they are treatment or control. Allows us to determent actual treatment effect and keep purpose of exper. hidden or misrepresented. |
|
|
Term
| Incomplete Administration of X (treatment) |
|
Definition
| Threat to internal validity. Attrition of subjects in the study. Groups no longer equivalent, selection threat and z confound is possible. Occurs mostly in field or longer term lab exper. |
|
|
Term
|
Definition
| a particular psychological state was necessary in the subject for study (anxiety on test performance), but the state is not achieved. |
|
|
Term
|
Definition
| way of accessing whether or not induced state was achieved (Fear ups eye blinking) |
|
|
Term
|
Definition
| threat to internal validity mostly in field exp. Control group members who weren't supposed to get x, got it (ex contact between C and T group) |
|
|
Term
| Intra session history confounds |
|
Definition
| IV prob of X. Something unintended occurring during experiment which differentiates groups or effects/absorbs results of exp. Ex: Studying effect of anxiety on test performance, but a/c breaks and rooms are hot, how do we know it was anxiety or the heat?) |
|
|
Term
| basic features of a good theory |
|
Definition
1. generality
2. breadth
3. simplicity
4. Accuracy |
|
|
Term
|
Definition
| Concerns specificity, the more general the theory, the more phenominon explains. ex, human behavor (all)-human voting behavior in 2000. |
|
|
Term
|
Definition
| has to do with the richness of details/facets explored/explained in relation to the theory. Ex. Voting behavior: high breadth would explore everything that relates to voting, candidate appearance, media coverage,campaign money, ect. |
|
|
Term
|
Definition
| occam's razor, if two theories are equal in every other respect, the simpler explanation is preferred. |
|
|
Term
|
Definition
| research based on objective observation of phenomena to achieve knowledge |
|
|
Term
|
Definition
| research designed to produce knowledge useful in real world situations. |
|
|
Term
|
Definition
| to satisfy intellectual curiosity and gain better understanding of the universe. |
|
|
Term
| deterministic causal reasoning |
|
Definition
| No EXCEPTION, even if 99% cases ok, the 1% is disconfirming |
|
|
Term
|
Definition
| almost always required, some exceptions ok |
|
|
Term
|
Definition
x is independent variable
y is dependent variable |
|
|
Term
|
Definition
| counterfactually, if there is a change in x, then there will be a change in y |
|
|
Term
| 4 criteria for establishing causal effect |
|
Definition
1. demonstrate x and y are associated empirically as explained by hypothesis,
2. Cause proceeded effect in time
3. the association is not a result of chance or a fluke
4. The association is not spurious |
|
|
Term
|
Definition
| simplified view of the world that allows us to describe, explain, and make predictions |
|
|
Term
| necessary and sufficient example |
|
Definition
anti-doping
absence of a (+)test you're not kicked out
if a + test is present, you will |
|
|
Term
|
Definition
measures single independent variable over time.
ex. x pre or post 9/11 y: pres. approv
cases
june, july, aug, sept, |
|
|
Term
|
Definition
data with variations across cases a given moment in time.
ex. x; educ y; income
ex. x: income y: % given to charity |
|
|
Term
| Fluke (how to determine if association is) |
|
Definition
1. Data dredging
2. no logical explanation/theory to demonstrate association
3. not replicable
4.does it pass test of statistical significance? |
|
|
Term
|
Definition
x: knee ache y: rain
z: humidity |
|
|
Term
|
Definition
No evidence that z causes x, but they are associated. z causes y.
x age, y: inclination to vote
z1. residential mobility
z2. political interest
z3. stake in election |
|
|
Term
| cross sectional analysis examp |
|
Definition
fair weather fan
n=schools x: wins
berk
stanf
usc
mich |
|
|
Term
| internal validity definition |
|
Definition
| extent to which the design enables the researcher to reach sound and valid causal conclusions about the effect of x on y |
|
|
Term
| why is a rce with pre and post test stronger in terms of internal validity than rce post test only |
|
Definition
1. allows us to access equiv. with respect to the DV prior to treatment , allowing us to reach more valid conclusions because we know whether or not differences were pre-existing.
2. allows us to demonstrate change over time/that the cause proceeded the effect. |
|
|
