Term
|
Definition
Overall effects of an IV, ignoring (collapsing across) the different levels of any other IVs
-Do the people who are associated to one condition differ overall from those who were not? |
|
|
Term
|
Definition
| Occurs when the influence of one IV on behavior differs (in direction and/or strength) DEPENDING ON the level of a 2nd IV |
|
|
Term
| A 2x4x3 factorial design has... |
|
Definition
| 3 IVs, with 2 levels on the first, 4 on the second, and 3 on the third |
|
|
Term
| A factorial with 3 IVs can have how many interactions? What is a 3-way interaction? |
|
Definition
| Four: AxB, AxC, BxC, AxBxC. The interaction of two independent variables depends on the level of a third independent variable not going to be asked to compute examples-you can have NO two way interactions but have a 3 way interaction |
|
|
Term
|
Definition
-Experimenter manipulates IV
-Measures DV
-Controls extraneous factors |
|
|
Term
|
Definition
-Experiment may not be manipulating the IV (may be a natural event ie government program)
-Experimenter may not have control over the measurement of the dependent variable (statistic…)
-Experimenter typically does not have as high a degree of control over extraneous variables as in a true experiment (possibly no control group, no random assignment etc.) |
|
|
Term
|
Definition
II. Types of QEDs
X=treatment/intervention/event
O=measurement
A. Post-test only Design
X → O (with the passage of time)
-Worst type of QED, no pretest for comparison
B. Pretest/Post-test Design
-O1 X O2
C. Time-Series Design
-Involves 2 or more measurements before vs. after the intervention
-O1 O2 O3 X O4 O5 O6
-Helps control for confounds the best, but history is still a big factor
These ALL are one group designs |
|
|
Term
|
Definition
| -Because of E’s bias, E influences the results of the experiment |
|
|
Term
|
Definition
| -Aspects of the experiment that tell the participant what they are testing for and this alters their behavior |
|
|
Term
|
Definition
| -People behave in unnatural ways because they know they’re in a study |
|
|
Term
|
Definition
-Something that varies with the levels of your IV, and potentially impacts the DV
-Example: Expectation of functionality of treatment vs. non-expectation |
|
|
Term
|
Definition
| -A feature of my treatment is causing a certain kind of subject to be more likely to drop out, and when that happens, it influences the results (Losing depressed people in a study because they don’t want to exercise) |
|
|
Term
| Uncontrolled Nuisance Factors |
|
Definition
| -Random annoying stuff that happens that affects the results |
|
|
Term
|
Definition
| -Anything that happens between pre and post that has nothing to do with the IV that affects the results |
|
|
Term
|
Definition
| -Changes in the capacity of the subject due to subject development (also includes factors like experience) |
|
|
Term
|
Definition
| If people get better at taking tests after they keep taking it |
|
|
Term
| QED Confound: Instrumentation |
|
Definition
-Changes in the instrument you’re using to measure your IV (scale, thermometers that gets knocked out of adjustment)
-Also includes changes that may occur in observers |
|
|
Term
| QED Confound: Statistical Regression to the Mean |
|
Definition
-Statistical phenomenon
-A tendency for extreme scores, when measured again, to return closer to the mean
-If you take measure of improbable scores, and measure again, it will be closer to the mean when measured again
-Happens when selection sample comes from some extreme part of the sample |
|
|
Term
| Designs with a non-equivalent control group (definition) |
|
Definition
| GET THE DEFINITION FROM THE BOOK |
|
|
Term
| Pretest-posttest design with a non-equivalent control group |
|
Definition
-O1 x O2 (with dashed line under to show they are non-equivalent)
-QED with 2 groups
(This changes the dynamic of the potential confounds)
-Now, a History confound with be a History x Selection confound
-Is there any differential history confound? (are the two groups’ history significantly different as to affect the experiment?)
-Maturation x Selection (is there differential maturation going on?)
-Testing x Selection
-Instrumentation x Selection
-Statistical Regression to the Mean x Selection
-Any general changes will affect both groups, but it’s always pssible that the two groups are facing different confounds |
|
|
Term
| Simple-Interrupted Time-Series Design w/ non-equivalent control group |
|
Definition
| -Like time series, but control group doesn’t receive treatment “interruption” in the middle |
|
|
Term
| F. Time-Series Design w/non-equivalent control group and switching replication (Crossover time series design) |
|
Definition
-Example: Treatment=Radio Soap Opera Series with AIDS education
-They instituted the treatment in different years; delayed it for the second group
-DV: Measured condom us, see if the trend happened in concurrence with institution of treatment |
|
|
Term
|
Definition
| (ABAB Designs, A=Initial observation, B=Treatment, Second A=Withdraw Treatment, Second B=Reinstate treatment) |
|
|
Term
|
Definition
TALK ABOUT GENERALIZATIONS, Skinner...
-Single Case designs can be applied to large groups (such as Nashville, Tennessee)
-With Reversal Designs…
-Is it ethical to withdraw treatment once it has started? (in the reversal stage)
-Irreversible behaviors (e.g. skills) |
|
|
Term
| Multiple Baseline Designs |
|
Definition
-A single case design that requires more than one person
-Example: Two or more separate participants with baseline periods (Mark’s is longer than Sara’s)
-You make the baseline periods different, an alternative explanation would HAVE to explain individually why the behaviors changed at different times, regardless of the treatment
Also: MBD Designs across settings(same thing, but takes place in different places), MBD Designs across behaviors (start giving treatment for the behaviors at different times) |
|
|
Term
|
Definition
To determine whether a program is operating as planned (feedback on first draft of paper)
-Program audit: How is the program actually being used?
-Pilot study
-Improvement while the study is going on, to continue to make it better/assess if it's going as planned |
|
|
Term
|
Definition
To assess the program’s overall effectiveness
-Tools:
-Quasi-experimental designs (most common)
-Small-N designs (e.g. police helicopter study)
-True experiments (i.e. field experiments) rare |
|
|
Term
| Cost-Effectiveness Analysis |
|
Definition
Assess program’s costs and judge in relation to effectiveness
-Can be performed during proposal phase, during formative evaluation, and during summative evaluation |
|
|
Term
|
Definition
| A variable that provides a causal link between an independent and dependent variable (distraction of attention for cell phone use/driving performance) |
|
|
Term
|
Definition
| A variable that alters the direction and/or strength of the relation between an independent and dependent variable (strength of traffic ->worse driving skills) |
|
|
