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;
57 Cards in this Set
- Front
- Back
- 3rd side (hint)
How can you conclude that the independent variable was the cause for the change? |
If groups differ on dependent variable at the end of experiments but starts the same. |
|
|
What experimental design do you chose if there is only 2 levels (groups/categories) of the independent variable? But only one independent variable |
Independent group t-test |
|
|
How many levels does a curvilinear relationship need? |
At least 3 levels (groups) of the 1 independent variable |
|
|
If your studying 2 or more independent variables in the same experiment what experimental design is used? |
Factorial design |
|
|
2 reasons for a factorial design: |
1) 2 (or more) qualitative (categorical data) IV are used 2) DV is quantitative (measures on a scale) |
|
|
Simplest kind of factorial design? |
Two by two |
|
|
What is examined in a simple main effect analysis? |
The differences at each level of the IV, If the differences themselves are different, as in not equal, then there is an interaction. |
|
|
Factorial designs with both manipulated and non-manipulated IV's are called: |
IV x PV design |
Causal claims cannot be made with complet confidence |
|
What is a participant variable (PV)? |
A variable that cannot be manipulated by the researcher. Ex: sex, age, etc. And is not a true experiment but a "quasi experiment" |
|
|
IV x PV studies are common because everyone: |
1) has personal characteristics 2) interacts with the environment |
Becomes a "higher-order factorial" |
|
Interactions often discussed in terms of a ? |
Moderator variable: affects the relationship between 2other variables. |
|
|
3 main types of assignment procedures in factorial design: |
Independent groups Repeated measures Mixed factorial |
|
|
If you want to increase the complexity of the factorial design you can: |
1) increase the number of levels in each IV 2) Increase the numbers of IV's 3) or both |
|
|
What was more common in the early days of psychology? |
Single case design or single participant (B.F skinner -> behaviourist) |
|
|
Béhavioriste critic to the statistical approach? |
It loses info about the individual |
|
|
Basic goal of a signal case study? |
To observe cause-effect in an individual directly rather than infer cause statistically. |
|
|
Single case design |
Used when individual is the focus of attention. -effectiveness of clinical treatment -medical research -behaviour modification |
|
|
Basic method of single case design: |
-establish baseline -administer treatment/therapy (IV) *chAnges from baseline indicate IV was effective. |
|
|
When is reversal designs (ABA designs) used? |
To show that an effect of the treatment can be undone |
Also known as withdrawal design |
|
Reasons for ABAB design? |
-Better evidence (a single reversal could be due to other causes) -ethics reasons (would be unethical to remove an effective treatment) |
|
|
3 basic types of multiple baseline design |
1) across subjects 2) across behaviours 3) across situations |
|
|
Across subjects is to show: |
That you can rule out alternative explanation for results of multiple baseline design |
|
|
Across situations is used to: |
Attempt to replicate across different people to show generalizability of treatment. |
|
|
What do ABAB replications often used in place of statistics? |
Graphical analysis |
|
|
Why use graphical analysis over statistics in ABAB ? |
On a Line graph, decide in advance: -minimum range for stable baseline (A) -minimum cut off for treatment (B) |
|
|
5 types of program evoluation: |
Need assessment Program theory assessment Process evaluation Outcome evaluation Efficiency assessment
|
|
|
When is the quasi-experimental design used in program evaluation? |
During the outcome stage |
|
|
What does a quasi experiment design(QE) study? |
The effect of an IV when true experiments are not possible because: A) lack of control group B) lack of random assignment *they have less internal validity than true experiments |
|
|
Types of QE designs: |
-one group, post test only -one group, pretest-posttest -non equivalent control group -non equivalent control group pre-test posttest -interrupted time series -control series |
|
|
Threats to internal validity: |
-History: any event that happens between the pretest and posttest. -Maturation: any systematic changes in people that occur over time. -Testing: pretest can change behaviour. -Instrument decay Regression towards the mean |
|
|
Overcoming threats to internal validity can be addressed using: |
-non equivalent control groups -non equivalent control group pretest-posttest |
|
|
Major difference of non equivalent control groups from experimental design? |
Control group is not randomly assigned |
|
|
Interrupted Time series design |
Multiple pre- and post measures interrupted by an event. |
|
|
Solution to the interrupting event |
Add a control series- a data that would not be affected by the archival records you are studying. |
|
|
3 basic methods of developmental research design |
-Longitudinal: same group measured as the age. -Cross-sectional: compare same age groups at one point in time. -Sequential: a mix of cross sectional and longitudinal. |
|
|
Why are cross sectional more common than longitudinal? |
-Take less time -Less expensive -Results are immediately available. |
|
|
Disadvantage to cross sectional? |
Cannot conclude that changes are caused by aging, can only infer. |
|
|
Reason why one can only infer age affects data in a cross sectional? |
Cohort effect (the era that the person was born in affects experiences throughout life) |
|
|
2 reasons for using statistics: |
-to describe data from the sample -to make inferences from the sample to the population |
|
|
4 types of scale measurement |
Nominal Ordinal Interval Ratio |
|
|
What do interval and ration scales share in common? |
Statistical analysis is the same even though conceptually they are different. Beachside they both have a meaningful average value (mean) |
|
|
3 basic ways of describing results of scale measurements: |
Comparing group percentages Correlating scores Comparing group means(averages) |
|
|
When data is nominal (categorical) use : |
Comparing group percentages |
|
|
Use correlating scores for: |
To see the linear association between two variables |
|
|
When data is nominal (categorical) use : |
Comparing group percentages |
|
|
Use correlating scores for: |
To see the linear association between two variables |
|
|
Use comparing group means when: |
IV is nominal/categorical DV is continuous |
|
|
Frequency distribution is used for: |
-To look at shape of distribution -To identify outliers (unusual extreme scores) -For number of responses in each category - for percentages in each category |
|
|
Bar graph are used when: |
Data is discrete categories |
|
|
Histograms are used : |
When data are continuous (or have underlying continuity) |
|
|
2 types of descriptive statistics: |
1) measure of central tendency 2) measures of variability |
|
|
Measures or central tendency are: |
Mean (the average) Mode (the most frequently occurring score) Median (the middle score in a range of scores) |
|
|
What is sometimes better at describing the central tendency? |
Median or mode |
|
|
Median is used when: |
Scores are rank-ordered can also be used with continuously scaled data) |
|
|
Mode is used when: |
Data are nominal categories |
|
|
Measures or variability (spread) |
Standard deviation Range |
|
|
What makes the shape of data? |
Mean and SD together |
|