2 ANOVA

Analysis of Variance is used as a test of means for two or more populations

H0 is that all means are equal:
Мю1 = Мю2 = ... = Мюk

DV is metric (interval or ratio scaled)

IVs are called factors. Each combination of factors is called a treatment

One IV

Multiple IVs

N-Way ANOVA

One-Way or Simple ANOVA

Relationships among techniques

Metric DV

IV

IVs

Binary

t-test

Categorical & Interval

Interval

Categorical: Factorial

ANCOVA

Regression

ANOVA

One Factor

More than one factor

One way ANOVA

N-Way ANOVA

has only one categorical IV. Hence, a treatment is the same as a factor level.

Examples in Marketing

Are customers equally satisfied with brand A, B, and C?

Do store 1, 2, and 3 differ in their performances?

Do segments differ in terms of the amount of their coffee consumption?

Do the evaluations of different ads vary?

Steps

2) Decompose the total variation

3) Measure the effects

1) Identify DV and IVs

4) Test the significance

5) Interpret the results

Example from lecture:

Do the store sales depend on the level of promotion in each store?

A department store chain wants to determine the effect of in- store promotion (IV) on sales (DV).

Results: (for high, medium and low in-store promotion)
Column mean: 8.3, 6.2, 3.7.
Grand mean: 6.067

Suppose you had to guess the sales level of an arbitrary store. What would be your best guess?

What if you would know that the store has a high level of in-store promotion?

6.067

8.3

A certain store is likely to differ from the group mean. In other words, there is a prediction error for each specific store.

concerned with the effect of more than one factor simultaneously

Examples

How does satisfaction with brand A, B, and C vary with sex?

Do ad evaluations depend on brand familiarity (high, medium, and low)?

How does the effect of in-store promotion on sales depend on the size of the store (small, medium, large)?

Analysis of Covariance