Final Flashcards
Matched groups
Between subjects design using manipulated independent variable ; having 2 groups that are matched on similar variable before random assignment groups
Independent groups
Between subject design using manipulated independent variable ; having 2 groups of randomly assigned subjects
Ex post facto
Between subject design with 2 groups either use:
Subject variable or non- equal groups
Also called “non-equalivent groups”
Repeated measures
Another name for “ within subjects design”
- participants tested in each experiment condition
Identify the 4 single varieties of single factor designs:
Independent groups
Matched groups
Ex post facto
Repeated groups
Single factor design
Basic research design with independent variable
Inter observer reliability
Agreement between 2 or more observers of same event
Homogeneity of variance
Variability of each set of scores being compared should be similar
Independent sample “t” test
Inferential stats analysis; comparing 2 samples of data in either independent groups or non equal vent groups design.
Dependent samples “ t” test
Inferential stats analysis; comparing 2 samples of data in either matched groups or repeated measures design
Using an independent t test which design is best?
Independent groups
Ex post facto
Using an dependent t test which design is best?
Matched groups
Repeated groups
Continuous variable
The variable exist on a continuum
Discrete variable
Represent distinct category ; no intermediate points occur
Single factor multi- level design
A single independent variable and more than 2 independent variable levels
Discrete variable
Each level represents a distinct category ; no intermediate points
One way ANOVA
One independent variable; overall presence of an significant effect existing between independent variables
ANOVA source table
It’s a method to portray inferential stats of the F ratio.
One way ANOVA for independent groups is best for which designs?
Multilevel independent groups design
Multilevel ex post design
One way ANOVA for repeated measures is best for what type of designs?
Multilevel matched groups design
Multilevel repeated measures design
Placebo
A pill form suggesting pharmacological effects, when it’s actually inactive
Placebo control group
Participant are lead to believe “ their treatment” is real when it’s not.
Method to observe behavior
Wait list control groups
Control group in which participants aren’t yet receiving treatment but will eventually
Yoked control group
A member of the control group is matched exactly with the treatment given a member of the experiment group
Factorial design
Study with more than one independent variable
Factorial matrix
Row/ column arrangement that characterizes independent variable and the total number of conditions ( cells)
Mixed factorial design
Mixture of within and between subjects facto sexist in same experiment
PxE factorial design
P= person, E= enviroment
Design with one subject factor and one manipulated factor
Mixed PxE factorial
See PxE factorial design
ATI design
It’s a form of PxE design; to examine possible interactions between aptitude variable ( p) and treatment variable ( e)
Simple effects analysis
Follow up test to significant interaction; comparing individual cells.
Main effect
Stats significance difference between the levels of an independent variable in a factorial design
Interaction
The effect when one independent variable depends on the other independent variable. (Factorial design)
Positive correlation
A relationship between variables “x” and “y”. “X” has high scores, so does “y” . If “x” has low scores , so does “y” .
Negative correlation
A relationship between variables “x” and “y”. A relationship between variables “x” and “y”. “X” has high scores, “y” has low scores. If “x” has low scores , “y” has high scores.
Pearson’s r
Measure of the size of the correlation between two variables; ranges from a perfect positive correlation of +1 to a perfect negative correlation of -1 . If “0”= no relationship.
Restricting the range
In correlate real studies; when a limited range of scores for one or both of the variables is used. Ranger restrictions tend to lower correlations
Coefficient of determination
For two correlated factors, where the variance in one factor is proportional to the other second factor, found by squaring persons r
Regression analysis
Occurs when knowing the size of the correlation and value of “ x” in order to predict the value of “y”.
Criterion variable
The variable being predicted from the predictor variable. ( I.e. College grades from sat scores )
Predictor variable
The variable used to predict the criterion variable. ( I.e. Sat. Scores are used to predict College grades )
Third variable problem
The problem of drawing a conclusion from variables “ a” and “b” due to interference from a third variable ( “c”) that is an uncontrolled factor that could. Underline the correlation between “a” and”b”
Partical correlation
A multivariate stats procedure for evaluating the effects of third variables; if the correlation between “ x” and “y” remains high, even after the third factor was partialed out. The third variable can be eliminated.
What process to determine if a correlation study truly has a third variable problem is called?
Partial correlation
Split -half reliability
A form of reliability which half the items on the test are correlated with remaining items. (I.e. Odd numbered questions)
Test re-test reliability
A form of reliability in which a test is administered on two occasions and the correlation between them is calculated.
Intra class correlation
A form of correlation used when pairs of scores don’t come form the same indivdual, when correlations are calculated for pairs of twins.
Bi variate
Relationships among 2 variables
Multivariate
Relationship among more than two variables
Multiple regression
A multivariate analysis investigating the relationship among more than two variables
Quasi experimental design
Occurs when casual conclusions about the effect of an independent variable can’t be drawn due to:
Lack of random assignment
Different levels of independent variable.
Archival research
Existing records are used to test an hypothesis
No equalivant control groups
Quasi- experimental design. Control and experiment groups can’t be randomly assigned