RM 2 Final Exam
1.Power: the probability of finding a significant result
where you reject the null hypothesis and its false or avoiding a type II
error, B formula: power= 1 - b (type 2 error)
3 things go into it: sample size, p-level, effect size
sample size (n)-decrease sample decreases power, positive correlated
relationship p-level (or a=.05)-as a increases so does the power of the
test
effect size- increasing increases power, positive correlation
2.Effect Size: the magnitude of a relationship between two or more
variables, how big the significance is
for example—
cohens D small-0.20
medium-
0.50 large-
0.80 very
large-1.2
a .2 effect size means that mean difference is around .2 standard
deviations
3.Difference between statistical significance and effect size: while
statistical significance helps us understand if the findings are due to
chance (failing to reject null=chance, reject null=not chance), the
effect size helps us understand the mag- nitude of these findings or
differences.
4.between subjects pro's and con's: pros
-more generalizable
-since each person has their own score no time or order effects
-each score in independent
cons
-subject to individual differences, problems across groups
-requires more people large # participants
-more unexplained variability, reduces power
5.within subjects pro's and cons: pros
, RM 2 Final Exam
-less people
-more power
-prevents confounding variables from individual differences
-each subject is their own control
, RM 2 Final Exam
cons
-takes longer
-less generalizable
-subject to time/order effects
6.between subjects vs within subjects design: a between subjects design
as- signs participants only to one condition of the IV, so that there is
only one score per person. It needs more people but is more
generalizable, there is no issues with time/order effects but issues with
individual differences (so random assignment helps)
a within-subjects design has the participants undergo each of the IV
conditions/lev- els, so that each person serves as their own control. It
requires less people but that makes it less generalizable. It eliminates
issues with individual differences and the less variability increases its
power. However, it is subject to time and order effects so it requires
counterbalancing.
7.what kind of t-test is for between subjects? what kind of anova is for
between subjects?: you would use an independent-samples t-test and
either a one-way between subjects ANOVA, or a between-subjects
factorial ANOVA.
8.what kind of t-test is for within subjects? what kind of anova is for within
subjects?: you would use a paired-samples t-test and a within-subjects
repeated measures ANOVA or a one-way within-subjects ANOVA
9.what are the assumptions for a paired samples t test?: 1. DV is
normally distributed
2.random sampling and assignment
3.independence of observations
10.what are the assumptions for an independent samples t test?: 1. DV
is normally distributed
2.random sampling and assignment
3.homogeneity of variance
4.independence of observations
11.what are the assumptions for an factorial anova?: 1. data are
independent (between-subjects design)
2.DV is normally distributed
3.homogeneity of variance
4.random sampling
5.independence of observations
12.what are the assumptions for a one-way anova?: 1. random sampling
, RM 2 Final Exam
2.normally distributed
1.Power: the probability of finding a significant result
where you reject the null hypothesis and its false or avoiding a type II
error, B formula: power= 1 - b (type 2 error)
3 things go into it: sample size, p-level, effect size
sample size (n)-decrease sample decreases power, positive correlated
relationship p-level (or a=.05)-as a increases so does the power of the
test
effect size- increasing increases power, positive correlation
2.Effect Size: the magnitude of a relationship between two or more
variables, how big the significance is
for example—
cohens D small-0.20
medium-
0.50 large-
0.80 very
large-1.2
a .2 effect size means that mean difference is around .2 standard
deviations
3.Difference between statistical significance and effect size: while
statistical significance helps us understand if the findings are due to
chance (failing to reject null=chance, reject null=not chance), the
effect size helps us understand the mag- nitude of these findings or
differences.
4.between subjects pro's and con's: pros
-more generalizable
-since each person has their own score no time or order effects
-each score in independent
cons
-subject to individual differences, problems across groups
-requires more people large # participants
-more unexplained variability, reduces power
5.within subjects pro's and cons: pros
, RM 2 Final Exam
-less people
-more power
-prevents confounding variables from individual differences
-each subject is their own control
, RM 2 Final Exam
cons
-takes longer
-less generalizable
-subject to time/order effects
6.between subjects vs within subjects design: a between subjects design
as- signs participants only to one condition of the IV, so that there is
only one score per person. It needs more people but is more
generalizable, there is no issues with time/order effects but issues with
individual differences (so random assignment helps)
a within-subjects design has the participants undergo each of the IV
conditions/lev- els, so that each person serves as their own control. It
requires less people but that makes it less generalizable. It eliminates
issues with individual differences and the less variability increases its
power. However, it is subject to time and order effects so it requires
counterbalancing.
7.what kind of t-test is for between subjects? what kind of anova is for
between subjects?: you would use an independent-samples t-test and
either a one-way between subjects ANOVA, or a between-subjects
factorial ANOVA.
8.what kind of t-test is for within subjects? what kind of anova is for within
subjects?: you would use a paired-samples t-test and a within-subjects
repeated measures ANOVA or a one-way within-subjects ANOVA
9.what are the assumptions for a paired samples t test?: 1. DV is
normally distributed
2.random sampling and assignment
3.independence of observations
10.what are the assumptions for an independent samples t test?: 1. DV
is normally distributed
2.random sampling and assignment
3.homogeneity of variance
4.independence of observations
11.what are the assumptions for an factorial anova?: 1. data are
independent (between-subjects design)
2.DV is normally distributed
3.homogeneity of variance
4.random sampling
5.independence of observations
12.what are the assumptions for a one-way anova?: 1. random sampling
, RM 2 Final Exam
2.normally distributed
