In: Math
The following table summarizes the results of a two-factor ANOVA evaluating an independent-measures experiment with 2 levels of factor A, 3 levels of factor B, and n = 6 participants in each treatment condition.
A. Fill in all missing values in the table. Show your work (i.e., all computational steps for finding the missing values). Hint: start with the df values.
B. Do these data indicate any significant effects (assume p < .05 for hypothesis testing of all three effects)?
Between treatments: SS=75 df=? MS=?
Factor A: SS=? df=? MS=? Fa=
Factor B: SS=? df= ? MS= 15 Fb=
AXB: SS=? df=? MS=? Faxb=6.00
within treatments: SS=? df=? MS=?
total: SS=165 df=?
Solution
Part (A)
ANOVA Table: Significance Level 0.05
[Given figures are in bold font. Other figures are derived. Explanations follow the table.
Source |
SS |
df |
MS |
F |
Fcrit |
Factor A |
9 |
1 |
9 |
3 |
4.17 |
Factor B |
30 |
2 |
15 |
5 |
3.22 |
A x B Interaction |
36 |
2 |
18 |
6 |
3.22 |
Between Treatments |
75 |
5 |
|||
Within Treatments |
90 |
30 |
3 |
||
Total |
165 |
35 |
Explanations
Additional information given
Factor A – 2 levels
Factor B – 3 levels
Replication - 6
The above imply:
Total number of observations N = 2 x 3 x 6 = 36
Number of cells C = 2 x 3 = 6.
df
Factor A |
Number of levels – 1 = 2 – 1 = 1 |
Factor B |
Number of levels – 1 = 3 – 1 = 2 |
A x B Interaction |
Between – A – B = 5 – 1 – 2 = 2 |
Between Treatments |
Number of cells – 1 = 6 – 1 = 5 |
Within Treatments |
Total – Between = 35 – 5 = 30 |
Total |
N – 1 = 36 – 1 = 35 |
For filling other figures, the following relationships are employed:
MS = SS/df, which also implies, SS = MS x df
F = MS[A/B/A x B}÷ MSW
Fcrit = Fυ1, υ2, where υ1, υ2 are the df of the MS in the numerator and denominator of the respective F.
ANOVA Table with details of computations [with order of computations in ()]
Source |
SS |
df |
MS |
F |
Fcrit |
A |
9 = 75 – 36 – 30 (6) |
1 |
9 = 9/1 (7) |
3 = 9/3 (8) |
4.17 |
B |
30 = 15 x 2 (1) |
2 |
15 |
5 = 15/3 (9) |
3.22 |
A x B |
36 = 18 x 2 (5) |
2 |
18 = 6 x 3 (4) |
6 |
3.22 |
Between |
75 |
5 |
|||
Within |
90 = 165 – 75 (2) |
30 |
3 = 90/30 (3) |
||
Total |
165 |
35 |
Fcrit is obtained from standard F-Table [can also be obtained using Excel Function: Statistical FINV ]
Answer 1 DONE
Part (B)
Hypotheses
Null H01: Mean effect of Factor A is the same for both levels Vs Alternative HA1: Mean effect of Factor A differs between the two levels
Null H02: Mean effect of Factor B is the same for all three levels Vs Alternative HA2: Mean effect of Factor B differs for at least one level
Null H03: Mean interaction AxB effect is zero Vs Alternative HA3: Mean interaction AxB effect is not zero
Answer 2
Fcrit values are given in the above ANOVA Table. Answer 3
Decision
Since F > Fcrit for Factor B and AxB Interaction, these two effects are significant.
Since F < Fcrit for Factor A, its effect is not significant. Answer 4
Conclusion:
There is sufficient evidence to suggest that Factor B and AxB Interaction effects exist, but evidence is not sufficient to support the effect of Factor A.
Hence we conclude that Factor B impacts the response variable, but not Factor A. However, their interaction does affect the response variable. Answer 5
DONE