2x2x2 Analysis of Variance for Independent Samples

This page will perform an analysis of variance for the situation where there are three independent variables, A, B, and C, each with two levels. As illustrated in the following table, this situation yields 2x2x2=8 unique treatment combinations— a1b1c1, a1b1c2, and so forth— one for each of 8 independent samples of subjects.

  A1
  A2
  C1
C2
C1
C2
B1
measures
for sample
a1b1c1

measures
for sample
a1b1c2

B1
measures
for sample
a2b1c1

measures
for sample
a2b1c2

B2
measures
for sample
a1b2c1

measures
for sample
a1b2c2

B2
measures
for sample
a2b2c1

measures
for sample
a2b2c2


Please note that this simple version of a three-way ANOVA is not recommended for data sets with an unbalanced design. A balanced design is one in which all eight cells in the analysis have equal values of n. The design is unbalanced in the degree that the eight cell values of n are heterogeneous.

Procedure:



Data Entry
  A1
  A2
  C1
C2
C1
C2
B1


B1


B2


B2


               

Values of N
ABC
A1 A2
C1 C2 C1 C2
B1
B1
B2 B2

    A1 
    A2 
    B1 
    B2 
    C1 
    C2 
 Total 


Mean Values
ABC
A1 A2
C1 C2 C1 C2
B1
B1
B2 B2
AB
  A1
A2
B1



B2






AC
  A1
A2
C1



C2






BC
  B1
B2
C1



C2








ANOVA Summary
Source SS df MS F P
 A
 B
 C
 AB
 AC
 BC
 ABC
 Error
 Total


Home Click this link only if you did not arrive here via the VassarStats main page.


©Richard Lowry 2001-
All rights reserved.