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;
51 Cards in this Set
- Front
- Back
Cohort Effects
|
Differences between subjects of cross-sectional research that have to do with experience rather than with age. Cohort effects threaten the conclusion that observed differences between subjects are related to age.
|
|
Case Study
|
The in-depth study of a single unit of investigation (e.g., individual, community, institution). From an experimental viewpoint, case studies are primarily useful to identify variables for future research.
|
|
Control Group
|
In a research study, the comparison group that is not exposed to the "active" level of the independent variable; i.e., a "no-treatment" or placebo group.
|
|
Cross-Sectional Research
|
Research design in which a sample of individuals representative of several dimensions of the population is assessed at the same time. For instance, in studying the effects of age, the design might include young, middle-aged, and older subjects.
|
|
Cause-and-Effect Sequence
|
the framework for explanation in scientific research
Policy sciences are a special case because you are evaluating a policy or program intervention to see if it work and if it makes a difference |
|
Proof
|
not the outcome of social science research
the most we can do is gradually strengthen our confidence in the validity of a causal sequence by eliminating possible alternative explanations |
|
Repeated testing
|
Through repeated testing we build confidence in a causal sequence
|
|
Controls
|
In testing controls must be introduced for alternative explanations that are suggested by the theory, intuition, and observation
|
|
Science is empirical
|
Measurement in the defining feature of science
Precision is achieved though quantitative measurement “When you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind…” But quantification isn’t enough (“When you can measure it, when you can express it in numbers, your knowledge is still of a meager and unsatisfactory kind…”) |
|
“When you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind…”
|
science is empirical but quantification isn't enough (“When you can measure it, when you can express it in numbers, your knowledge is still of a meager and unsatisfactory kind…”)
|
|
Unit of Analysis
|
The major entity that you are analyzing in your study
U/A determined by the actual analysis done in the study All of the study factors are things that describe the u/a |
|
Data sets
|
Collections of measurements for specified units of analysis
|
|
“clean” data
|
Not collected but developed
Developed by filtering out unwanted influences by Statistical analysis (most common) Research design Better data = more reliable study = great validity to relationships |
|
Research Design
|
the study of techniques for organizing a research process to achieve variance control prior to and during data development
compare to after-the-fact cleaning and controlling achieved through statistics and econometrics |
|
Bias
|
Errors that seep into a structured inquiry at one or more junctures
|
|
Five Types of Bias
|
Sampling
Selection Measurement Model Specification Statistical |
|
Sampling Bias
|
From not selecting a "representative” group to analyze
“Is the sample representative of the larger population?” |
|
Model Specification Bias
|
In Model Specificatin you define basic concepts in a model
Model Conceptualization and Omitted Variable Bias: -Omitted variable bias (forgotten “z” has an effect) -Functional form -“Is the model complete?” -“Are all of the major influences present?” |
|
Measurement Bias
|
In measurement you assign values to the u/a atrributes
Measurement Bias: -Response and/or non-response, coding errors, etc. -Do we have sloppy measurement? -“Do variables vary sufficiently?” (If variables don’t vary than they can’t explain) -“Is indicator measurement reliable (accurate) and valid?” |
|
Statistical Bias
|
Select appropriate data analysis model
Statistical Bias -Violation of the assumptions associated with data analytical techniques (biased estimators) -“Are the assumptions underlying the analysis model met?” |
|
Selection Bias
|
Estimation and inference
Selection bias -When background characteristics other than the main IV can explain an outcome -Means a better model needs built -Draws attention to the u/a -“How adequate are the controls for eliminating alternative explanations?” |
|
4 Elements of Experimental Design
|
IV & DV
Randomization & Ceteris Paribus Comparison Group Quasi-Experiments |
|
Quasi-experiments
|
introduced in 1950’s by Campbell
Do not use randomization to create comparison groups Confidence in causal process can be developed across a sequence of studies linked to relatively weak non-experimental designs Depends on design structure Uses either time or comparison group or both to diminish/eliminate specific threat to internal validity |
|
Randomization
|
Assigning randomly U/A’s to differing levels of IVs
|
|
Control
|
Rule out threats to causal inference
|
|
5 Schools of Thought
|
Positivism (Hume)
Essentialism (probabilistic versus deterministic) Concomitant Variation (Mill) Falsification (Popper) Activity (“cause” as manipulation) |
|
Main philosopher for Positivism
|
Hume
|
|
Positivism (Hume)
|
Only recognized observable phenomena
Causal relationship based only on highly correlated variables |
|
Hume's 3 Conditions for Inferring Cause
|
Contiguity
Temporal precedence Constant conjunction – connection must be apparent constantly, when one variable changes so does the other |
|
Philosopher who Challenged Positivism
|
B. Russel
|
|
Russel's arguments against positivism
|
challenged positivism & the need for causal relationships (not needed in hard science)
causal chains or intervening variables establish unacceptability of positivism |
|
Essentialism
|
Associated with reductionism and determinism
“cause” suggests variables both necessary and sufficient for effect to occur Factor must work all of the time to be “cause”, but causal relationships more likely to be probabilistic than deterministic |
|
Philosopher for Concomitant Variation
|
Mill
|
|
Concomitant Variation (Mill)
|
Key to design is establishing “counterfactual”
Implied comparison, sets up elementary comparison with and without treatment/IV Great Contribution: realize comparison of situations where particular threat to valid inference was/wasn’t operating provided key for assessing whether the threat may occur for any observed relationship |
|
Mill's Criterian for Cause-Effect Relationships
|
1. Variable must co-relate
2. Time Order – cause precedes effect 3. Eliminate alternative explanations – the reason why “control” is critical |
|
Philosopher for Falsification
|
Popper
|
|
Falsification (Popper)
|
Base of knowledge is ruling out alternative explanations
Can never prove theory to be true, but failures to prove true can prove false The more causal variables ruled out, the more confident we become with the relationship between IV and DV |
|
Three Types of Relationships
|
1. Symmetrical
2. Reciprocal 3. Asymmetrical |
|
Symmetrical Relationship
|
2 variables co-vary, but neither responds to the other
• Alternative indicators • Effects of a common cause • Functional interdependence • Parts of a system w/o interdependence • chance |
|
Reciprocal Relationships
|
Two variables co-vary, but each may influence the other
• Difficult to trace • Non-recursive systems: allows for influence feedback, analytic techniques include OLS w/ lagged variables or simultaneous equations |
|
Asymmetrical Relationships
|
one variable influences the other, influence flows in one direction only
• “Heart” of Causal logic • Rational for independent and dependent variables • Recursive systems: ideal for analytic techniques e.g. OLS |
|
Recursive Systems
|
influence arrows only run in one directions i.e. asymmetrical relationships
|
|
Variation
|
Greater variation in IV is good when trying to examine variation in DV
|
|
“sine qua non”
|
without which there is nothing
The essential element in causal inference is variation across U/As |
|
Types of Data Structures
|
Panel
Cohort Cross Sectional |
|
Panel
|
• Internal focus
• Use U/As as their own control • Controls for unwanted variance associated with individual differences |
|
Cohort
|
• External focus
• Controls for sources of unwanted influence associated with period influences • Controls for temporally defined influences (birth, graduation) |
|
Cross-Sectional Data Structure
|
•Controls specified variables
•Not located in time and space |
|
Types of Design Structures
|
Experiments
Quasi-Experiments Cross Sectional |
|
Experimental Design Structure
|
• BEST
• Controls for all extraneous influences known & unkown • Effectiveness may diminish over time |
|
Cross-Sectional Design Structure
|
• Depends on theory guided model specification
• Weakest and least effective control strategy |