ANOVAs

Question 1

ANOVA

Answer 1

-compares 3+ means
-when more than 1 IV
-tests for difference but not what differs
-F statistic

Question 2

One-Way Repeated Research Design

Answer 2

-one group doing the same thing over titme

Question 3

Posttest-Only Randomized Group

Answer 3

-randomized
-2+ groups that are only measured after the intervention

Question 4

Posttest-Only Non-Randomized Research Design

Answer 4

-not randomized
-2+ groups that are only measured after the intervention

Question 5

Factorial-Fully Independent Research Design

Answer 5

-with or without randomization
-each intervention has subsets
-2+ interventions at the same time
-each person stays in the same group the whole time
-look for interactions, main effect of each IV
-if interaction is significant (crossing lines), dont worry about main effects (differing)= run simple effects

Question 6

Factorial-Fully Repeated Research Design

Answer 6

-each intervention has subsets
-2+ interventions at the same time
-all participants switch groups throughout the study

Question 7

Factorial Mixed Research Design

Answer 7

-with or without randomization
-Participants in designated groups move through several parts
-most common in PT

Question 8

One Way independent ANOVA

Answer 8

-one IV
-is there a difference somewhere?
-3 kinds of viagra and their effects

-non parametric: Kruskal-Wallis

Question 9

One way Repeated Measures ANOVA

Answer 9

-one repeated IV
-travel through time

-Sphericity: variannces of the differences are equal
-tested using Mauchley’s test, dont want sig

Non parametric: friedman’s ANOVA

Question 10

2x2 independent ANOVA

Answer 10

-2 or + IVs
-look for significant differences NOT interactions (factorial)

Question 11

2x2 Repeated Measures ANOVA

Answer 11

-2 or more repeated IVs

Question 12

2x2 Mixed or Split plot ANOVA

Answer 12

-2 or more IV
-at least 1 is independent and one is repeated

Ex: 3 groups being tested 5 times

Question 13

OMNIBUS Test

Answer 13

-F(x,y)=#,
x= groups
y=people
df= degrees of freedom
f= bigger=more significant

does not alter alpha level, post tests do

Question 14

Planned Contrasts

Answer 14

-additional test to determine which groups differ from each other
-more statistical advantage
-must be done prior
-a prori

Question 15

Post-Hoc Comparisons

Answer 15

-additional test to determine which groups differ from each other after ANOVA
-not planned
-control the alpha level
-test all pairwise comparisons

Bonferroni: controlls for type 1, only small variances, normality, better if small

Turkey: controlls for type 1, only small variances, normality, better if large

Games-Howell: liberal when smaple is small, if variances arent homogenous

Question 16

Pairwise

Answer 16

-only 2 means
-no interactions
-report main effects only (dont divide alpha for main effects)

total= (tests)^2

Question 17

Simple Effects

Answer 17

-run after ANOVA
-only if significant or interactions are found
-divide by tests run

-interactions within groups (3 DV = 9 tests)
-number of one way comparisons (not individual tests, just directions (3 dv and 3 iv= 6 one way)

Question 18

Kruskal-Wallis Test

Answer 18

-non parametric equivalannt to independent 1 way anova
-doesnt meet asumptions
-compares means

Question 19

Friedman’s ANOVA

Answer 19

-non parametric equivalent to RM one way ANOVA

Question 20

Covariates

Answer 20

-test ofr differences between means when you have outliers
-used to control extraneous variables

Question 21

ANCOVA

Answer 21

-ANOVA with covariates

Question 22

MANOVA/MANCOVA

Answer 22

-multiple dependent variables measured in a ANOVA or ANCOVA
-better than multiple ANOVAs
-DFA (discriminant function analysis): used to determine which of 4 DVs is best, run after MANOVA

Question 23

Sphericity

Answer 23

-variannces of the differences are equal
-tested using Mauchley’s test, dont want sig
-only when interactions occur (no independent ANOVAs)

Question 24

Mann Whitney U

Answer 24

Non parametric for independent t test
-ranks means