Ch. 10 Exam (Exam 4)
What are the assumptions of a dependent-group samples t-test?
1. IV (or predictor) is dichotomous (nominal scale w/ 2 groups) 2. Groups are dependent via matching or repeated measures 3. n of 2 groups is equal 4. DV (or outcome) is interval or ratio 5. DV (or outcome) is normally distributed
Concerns for Repeated measures design (7) (COMP FC)
1. Order effect 2. Carryover 3. Practice 4. Fatigue 5. Maturation/history 6. Takes time and effort 7. Can't change participant in a permanent way
Fatigue effect
A confound that occurs when changes in the DV occur because of participants becoming tired
Carryover effect
A confound that occurs when the effect of one condition of the treatment continues (or carries over) into the next condition.
Order effect
A confound that occurs when the order of each treatment condition cannot be separated from the condition
Dependent-groups design
A design where participants in different conditions are related or are the same person - More sensitive & More powerful than independent-groups design - Decreases random error from individuals -Increases power, which is the ability to detect the difference
Cohen's d
A measurement of effect size; describes the magnitude of the effect of our IV (or predictor) on the DV (or outcome) in standard deviation units.
Counterbalancing
A procedure to remove order effects in a repeated measures experiment or quasi-experiment Participant's experience the different levels of the IV in different orders randomly assigned to the different orders even numbers of participants assigned to the different orders
Latin square balancing
A type of partial counterbalancing where each condition appears once in each sequence; participants are then randomly assigned to different sequences sudoku
Repeated measures design is also called
Between-subject design
Which of the following is NOT an assumption of the dependent-samples t test? A. Predictor is dichotomous B. The n of the groups is equal C. The groups are independent D. The outcome is interval or ratio
C. The groups are independent
____________ is an appropriate measure of effect size for dependent-samples groups
Cohen's d
Practice effect
Confound that occurs when the participants' score changes due to repeating a task rather than because of the level of the IV
Power of a repeated-measures dependent design comes from
Controlling all potential confounding participant variables
Where does the power of a repeated measures design come from?
Controlling all potential confounding participant variables
Dependent-groups design decreases _______, which is created from __________
Decreases "random error", which is created from participant characteristics This will decrease the chances of confounds due to participant variables and provides assurance of homogeneity of variance This is because the participants show similar characteristics in each group
Repeated measures design (within-subjects design)
Design in which participants experience every condition in a study; in an experiment, they also randomly assigned to the order of conditions.
Matched-pairs design
Design in which participants in each group are matched on a characteristic relevant to the variable that is being measured; in an experimental design, a member of each matched pair is randomly assigned to each IV condition.
Matched groups two-group design and repeated measures two-group designs are analyzes with different types of t tests. True or false?
False
Standard error of the mean difference (SD D)
Standard deviation of the difference for a sampling distribution of mean differences; estimated from the standard deviation of the difference scores in a dependent-samples study
Dependent-samples t-test (or paired-samples t-test/within-subjects t-test)
Statistical test used to analyze results from a dependent two-groups design
Appropriate statistic for a simple experiment
independent samples t test
Appropriate statistic for a multiple-groups design
one-way ANOVA
Appropriate statistic for a matched-pairs design
paired-samples t test
Appropriate statistic for a repeated measured design
paired-samples t test
Appropriate statistic for a repeated measured design dependent multiple-groups design
repeated measures ANOVA
Partial Eta Squared
the effect size for a dependent multiple-group design that removes the variability unique to individual participants (SS s) from the error term
Overall dependent-groups designs _______power of a study, because you ___________ the difference between groups and ________the error of variability in scores
- Increase - Maximize - Decrease
Why is it necessary in a repeated measures design?
. Counterbalancing is necessary in a repeated measures design, because it avoids confounding orders and condition
Concerns for Matched-Pairs Design (4) (CAAM)
1. Can't get perfect match 2. Attrition double (lose one person, lose pair) 3. Additional screening / testing of participant (e.g., participants know purpose study; SENTIZATION, fatigue, demanding more from participants) 4. More time and effort
Assumptions for one-way within-subjects ANOVA
1. DV is interval or ratio 2. DV is normally distributed 3. Sphericity in variances of differences between pairs of groups 4. IV has 3+ levels 5. Groups are dependent
Partial counterbalancing
Randomly assigning participants to different sequences of conditions so that each condition is represented in each order an equal number of times, not all sequences are represented when participants are randomly assigned to different sequences of conditions so that each condition is represented in each order an equal number of times but not all sequences are represented
Overall dependent groups designs
Increase power of study Maximize difference between groups Minimize error variability in scores
What is the major advantage of dependent designs?
Increases power - ability to detect the difference Decreases the random error from individual participants More sensitive
Which of the following is NOT a characteristic of a dependent-groups design when compared to an independent-groups design: A. Assurance of homogeneity of variance B. Decreased chance of confounds C. Greater sensitivity to changes in the measured variable D. Larger sample size
Larger sample size
Advantage of matched pairs design
Minimizes error due to individual differences Matching ensures groups are equivalent Reduce number of participants needed Increases number of observations
Randomized partial counterbalancing
Randomly assigning each participant to one of the possible sequences of conditions w/out concern about order or sequence; used when you have a larger number of sequences than participants
Complete counterbalancing
Randomly assigning participants to all the possible sequences of conditions in an experiment
Advantage of repeated measures design
Reduces error variance Increased power Decrease error variance
Sphericity
The assumption that the variances of the differences between all the combinations of pairs of groups are equal
Mean Difference (M D)
The average difference between the scores of the matched pairs or the scores for the same participants across two conditions
Standard error of the difference between means
The average variability in a sampling distribution of differences between means.
Where does the power of the dependent groups design come from?
The decrease in random error that is created by participant characteristics
Where does the power of the dependent-groups design come from
The decrease in random error that is created by participant characteristics
Within-subjects ANOVA (or repeated measures ANOVA/dependent-groups, one-way ANOVA)
The statistical test used to analyze dependent multiple-groups design
Matched pair design is also called
Within-subject design