ANOVA Assumptions

Question 1

What is the main assumptions of an independent ANOVA? - (3)

Answer 1

• normal distribution
• Independence of scores
• homogenity of variance?

Question 2

What is main assumptions for repeated ANOVA? - (3)

Answer 2

• normal distribution
• repeated measure design (same participants)
• sphereicity (Maulchy’s test)

Question 3

What is main assumptions of mixed ANOVA?- (3)

Answer 3

• normal distribution
• independent and repeated factors
• homogenity of variance for the independent factor + sphereicity for repeated factor

Question 4

What test is used for homogenity of variance?

Answer 4

Levene’s test

Question 5

Levens’ tests if the

Answer 5

variances in independent groups are similar

Question 6

Spherecity is an assumption of both

Answer 6

repeated measures and mixed models

Question 7

Spherecity is the assumption that

Answer 7

variances of differences between all combination of related groups (levels are equal)

Question 8

The violation of spherecity is when the

Answer 8

variances of the differences between all combination of related grps (levels) are not equal

Question 9

The violation of spherecity is serious for repeated measures ANOVA as

Answer 9

it causes test to be too liberal (i.e., increase in Type 1 error rate)

Question 10

The spherecity assumption applies when there are more than

Answer 10

2 data points from eac pp (3 or more levels in IV)

Question 11

With spherecity you need to check in SPSS using

Answer 11

Mauchly’s test

Question 12

For assumption of spherecity to be satisfied we need the Maulchly’s test in SPSS to be

Answer 12

non-significant (p > 0.05)

Question 13

If Mauchly’s Test of Spherecity is significant ( p < 0.05) then reject H0 that variances of differences are equal and

Answer 13

aceept H1 that the variances of the differences are not equal (i.e., spherecity assumption has been violated)

Question 14

If assumption of spherecity is violated then correct for lack of spherecity using

Answer 14

Greenhouse Geisser, Huynh-Feldt or Lower-bound estimates

Question 15

Greenhouse-Geisser, Huynh-Feldt or Lower-Bound estimates rely on

Answer 15

estimating spherecity

Question 16

You need to use adjusted DF for

Answer 16

greenhouse-geisser, Huynh-Feldt or Lower-Bound estimates

Question 17

If GG statistic less than 0.75 if spherecity is violated then

Answer 17

use GG

Question 18

If GG statistic more than 0.75 if spherecity is violated then

Answer 18

Huynh-Felft (HF) statistic

Question 19

Report the following F statistic for this output - (3)

Answer 19

• Spherecity is violated since p = 0.000
• GG is less than 0.75 so use GG
• F(1.24, 21.00)= 212.32, p < .001

Question 20

The researcher hypothesized that there would be an interaction between dog breed (Collie or German Shepherd) and week of obedience school training (all dogs measured at 1 week and 5 weeks) as they relate to the number of times the dog growls per week. Specifically, it was hypothesized that Collies would show no difference in growls between 1 week and 5 weeks, but German Shepherds would growl less at 5 weeks than at 1 week

What is IV, more than one IV, levels of IV, DV, BS or WS , what type of design? - (6)

Answer 20

1. Dog breed and measurement time
2. Yes
3. Collie-German Sehpard/Week 1 to Week 5
4. Number of growls
5. Dog breed, BS, measurement time - WS
6. 2-way mixed ANOVA

Question 21

The researcher hypothesized that there would be an interaction between dog breed (Collie or German Shepherd) and week of obedience school training (all dogs measured at 1 week and 5 weeks) as they relate to the number of times the dog growls per week. Specifically, it was hypothesized that Collies would show no difference in growls between 1 week and 5 weeks, but German Shepherds would growl less at 5 weeks than at 1 week

1. IVs: Dog breed and measurement time
2. more than one IV? : Yes
3. Levels of IVs: Collie-German Sehpard/Week 1 to Week 5
4. DV: Number of growls
5. BS or WS? : Dog breed, BS, measurement time - WS
6. Type of design2-way mixed ANOVA

What does SPSS output show? - (3)

Answer 21

• Is there an effect of time overall? = Yes as p = 0.00
• Is there an effect of breed ? Yes as p = 0.000
• Is there an interaction effect? Yes as p = 0.000

Question 22

The F-ratio tells us only whether the model fifted to the data accounts for more variation than extraneous factors but does not tell us

Answer 22

where the differences lie

Question 23

Two ways to know where differences lie between the groups in ANOVA is - (2)

Answer 23

1. Post-hoc
2. Planned contrasts