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;
35 Cards in this Set
- Front
- Back
Benefits and Limitations of > two group designs
|
We gain more information for a more complex design but add works and increases the likelihood of extraneous factors
|
|
Analysis of and Interpretation > two group design
|
ANOVA f-test
F= variability due to IV + error variability/ error variability |
|
One way ANOVA for independent sample
|
"one way" means one IV
"independent" means different participants in each condition |
|
Planned comparisons (a priori test)
|
To compare specific groups, test specific hypotheses
|
|
Post Hoc Test
|
Evaluate apparent difference among group means - design to control risk of type 1 error
|
|
Repeated Measures design- Analysis test
|
One way ANOVA for correlated samples
|
|
Factorial Design
|
Studies effect of 2 or more independent variables
|
|
Interaction effects:
|
occur when the effect on one OV differs depending on the level of the second IV
|
|
Main effect:
|
effects of each IV alone
|
|
synergetic interaction
|
Combination results of an effect beyond expected
|
|
Identifying an interaction (2x2)
|
Row 1 subtract the slower score
Row 2 subtract the slower score If these to outcomes differ it likely indicates an interaction |
|
2x2 Factorial 3 hypothese
|
A main effect of factor A
A main effect of factor B An interaction between factor A and B |
|
Analysis - Two ANOVA for Independent Samples (Factorials)
|
Graph analysis
Non parallel means an interaction Parallel means no interaction |
|
Repeated Measure Factorial - Pros & Cons
|
Each participants under each condition
Pros - control individual difference, more sensitivity Cons- Carryover effects, influence of treatment order |
|
Mixed Factorial Designs (3)
|
-One independent group factor + One repeated measures factor
- One manipulated factor + one non-manipulated factor -Repeated measures factor + independent group factor |
|
One independent group factor + One repeated measures factor
|
I'm not sure you will fine :)
|
|
One manipulated factor + one non-manipulated factor
|
Treatment and analysis as if an independent group
Allows controls of confounds Difference matter of interpretation Does not cause but may associated with.. |
|
Repeated measures factor + independent group factor
|
Effects analysis of variance
Not causal just correlational |
|
Floor effects
|
Participants scores at the minimum on a s scale/measure - due to limited response options
|
|
Ceiling effects
|
Participants score at the maximum on a scale/measure - due to the limited response options
|
|
Alternative Research Design - threats to internal validity (8)
|
Maturation
History Testing Instrumentation Regression to Means Selection Mortality Diffusion of treament |
|
Post-test Only Control Group design
|
Randomize Experimental group - treatment - post test
Randomize Control group - no treatment - post test Compare post test results Randomization - controls regression, mortality, selection Control group - controls instrumentation, history, maturation |
|
Pretest - post test control group design
|
Randomize Experimental group - pretest - treatment - post test
Randomize Control group -pretest - no treatment - post test compare post test results Pretest ensure groups' equivalence, controls testing effects Randomization - controls regression, mortality, selection Control group - controls instrumentation, history, maturation |
|
Solomon's Four Group Design
|
R Group A - Pretest - Treatment - Post test
R Group B - Pretest - - Post test R Group C - - - Post test R Group D - - - Post test Complex but controls for most effects (Regression to means, instrumentation, history, maturation) -Ideally B = D and A = C |
|
Single Case Designs - N = 1 designs
|
Involve manipulated variables, controls for confounds
To evaluate change in behaviour of individual Gain information lost by average scores Useful to assess individual performance Weak external validity |
|
Single Case Designs - N = 1 designs (3 types)
|
Reversal Design
Multiple Baseline Design Randomized Time-series Design |
|
Reversal Design (ABA design)
|
A - Baseline period - observe dependent behaviour
B - Treatment period - manipulate IV A - Baseline returned Dramatic change means relationship If a treatment IV has a beneficial effect on a DV, we have an ethical obligation to return effect to the participant - ABAB |
|
Multiple Baseline Design
|
Sometimes reversals aren't ethical or feasible
The effects of treatment are added one at time on each other We look for effects only on the behaviours being targeted in each phase, coinciding with each change in the intervention |
|
Randomized Time-Series Design
|
Establish a baseline, randomly begin treatment - change in behaviour most dramatic with onset of the manipulation
Difficult to draw firm conclusions when data is messy Don't need statistical analysis |
|
Quasi-experiment designs
|
-Employed in field research
- Applied when true experiments are impossible -Requires cautious interpretation State one or more hypotheses with causal effects Include at least two levels of the IV Typically, do not assign participants to groups, but use already existing groups Include procedures Include some controls for threats to validity |
|
Quasi-experiment designs (3 types)
|
Nonequivalent Control Group Design
Interrupted Time Series Design Time Series with non equivalent control group design |
|
Nonequivalent Control Group Design
|
Ex Group - Pretest - treatment - post test
Co Group - Pretest - no treatment - post test Compare scores of Post test scores - Pretest scores of each group -Groups differ on DV from beginning -Other difference may affect DV -Comparison group controls history, maturation, testing Check out graphs |
|
Interrupted Time Series Design
|
1. Establish baseline
2. Apply treatment 3. Measure behaviour over several more periods Often rely on data collected over time used when hypothesized IV affects entire population or when we ca't use a control group |
|
Time Series with non equivalent control group design
|
1. Establish baseline
2. Apply treatment 3. Measure behaviour over several more periods 4. Add similar existing group as a control group Often rely on data collected over time used when hypothesized IV affects entire population or when we ca't use a control group |
|
Program Evaluation
|
Social services, governmental services/interventions, Programs within companies or schools
Goal is to provide evidence that a program is accomplishing intended goals Realities - practical consideration, - Ethical constraints - can't limit access to social programs, free informed consent, evaluators becomes diplomats - staff and clients |