MRM2 – PC Lab 2 – Open book assignment
MRM2 – PC Lab 2 – Open book assignment
This data set refers to a study exploring differences in consumer attitudes towards new
products that brands introduce. It distinguishes 3 strategies to introduce new products
under existing brand names.
1. Brands can introduce products in a new category. We call these category extensions, e.g.
M&M's would introduce chocolate milk.
2. Brands can introduce products in categories in which they are already present. We call
these line extensions, e.g. M&M's introduces a dark chocolate version of the product
already on the market.
3. Brands can introduce new products together with another brand. We call these brand
alliances, e.g. M&M's would introduce ice cream together with Ben & Jerry's.
Brand equity is a construct that summarizes the strength of a brand (in terms of familiarity,
image, perceived quality, loyalty, etc.). We are interested in whether the evaluation of these
branding strategies would be different for strong brands (brands with high brand equity) vs.
weak brands (brands with low brand equity). To analyze this, we use the following variables
from a larger data set:
• BrandStrat, labeled Branding Strategy. This is a categorical variable, which acts as a PV
in your analyses. It has 3 groups: 1 = Category Extension; 2 = Line Extension; 3 = Brand
Alliance.
• BrandEquity, labeled Brand Equity. This is also a categorical variable, which acts as a
second PV in your analyses. It has 2 groups: 0 = Low (i.e. a weak brand); 1 = High (i.e. a
strong brand).
• Attitude, labeled Attitude towards Branded Product. This variable measures consumer
reactions to the product a brand introduces (via the strategies outlined above) on a 7-
point scale, from 1 = Very unfavorable attitude to 7 = Very favorable attitude.
You expect that Attitude towards Branded Products will be higher for Brand Alliances than
for Line Extensions, while at the same time, attitudes will be higher for Line Extensions than
for Category Extensions.
Furthermore, you expect that this effect depends on the Brand Equity of the brand; that is
for weak brands, Category extensions will have lower Attitude towards Branded Products
than Line Extensions and Brand Alliances separately, while for strong brands, Attitude
towards Branded Products will be similar across the three branding strategies.
You want to run a 2-way independent ANOVA to analyze the impact of the two PVs and
their interaction on the OV.
Please answer the questions on the following pages:
MRM2 – PC Lab 2 – Open book assignment

2

1. A) Draw the conceptual model based on the expectations given, ensure to identify which
PV is the moderator.
B) Formulate the hypothesis for the expected main effect of Branding strategy on
Attitude.
C) Formulate the hypothesis for the expected moderating effect of Brand Equity on the
relationship between Branding strategy on Attitude.

A.




B.
H0: Category Extensions, Line Extensions, and Brand Alliances will lead to similar attitudes.
H1: Line Extensions lead to higher attitudes than Category extensions, but to lower attitudes
than Brand Alliances.
C.
H0: The effect of brand strategy on attitude is not moderated by brand equity.
H1: The effect of brand strategy on attitude is moderated by brand equity, such that:
- for weak brands, Category extensions will have lower Attitude towards Branded Products
than Line Extensions or Brand Alliances
- for strong brands Attitude towards Branded Products will not differ across the three
branding strategies

2. Run the 2-way independent ANOVA indicated above.
A) Manually check the F-statistic for the entire model with the information from the
ANOVA table and report your calculation.
B) Report the p-value for this F-statistic and draw a conclusion: what can we say about
the model?
C) Calculate the proportion of variance in the OV explained by the model. Report your
calculation.

Branding
Strategy
Attitude towards
Branded Product
Brand
Equity
MRM2 – PC Lab 2 – Open book assignment

3


A.
F = (SSM / dfM) / (SSR / dfR) = (198.67/5) / (260.767 / 193) = 29.408. For 2-way ANOVA, you
use the rows containing information about the Corrected Model and the Error.

B.
p<0,05, thus at least one of the PVs, or their interaction (i.e. the third PV) has a significant
impact on the OV. The model significantly explains differences in the outcome variable and
further analysis is meaningful.

C.
R2 = SSM / SST = 198.67 / 459.44 = 0,432. This is the percentage of the total
variance in the OV (Attitude) that is explained by the model (i.e. by the 2 PVs and their
interaction).




3. Regarding the main effect of branding strategy, say you are mainly interested in whether
a Line extension scores different than either a Category Extension or a Brand Alliance.
A) Is there a significant main effect of branding strategy on attitude towards the
product?
B) Report the lines of code you add to the syntax to conduct the follow-up tests you are
interested in.
C) Execute the post-hoc test and report your results: are Line Extensions evaluated
differently than Category Extensions or Brand Alliances? Is your expectation met?

A.
From the test of Between-Subjects Effects table, we observe a significant main effect of
branding strategy on attitude towards the product (F[2, 193] = 38.88, p < .05).

MRM2 – PC Lab 2 – Open book assignment

4

B.
/EMMEANS=TABLES(BrandStrat) COMPARE(BrandStrat)

Alternatively, in SPSS 28, you can tick
Analyze → General Linear Model → Univariate → Post Hoc, and select BrandStrat; since the
variable has at least 3 categories. Note, this procedure will produce slightly different
estimated means and mean differences.

C.
From the mean estimates table, preceding the Pairwise comparisons table, we find that the
mean score of Line Extensions on Attitude is equal to 4.853, this is…
(1)… significantly higher than the mean Attitude of Category Extensions (3.821): the
difference in means = 1.031 (CI [0.635 to 1.428], and this difference is significantly different
from zero (p < .05);

(2)… significantly lower than the mean Attitude of Brand Alliances (5.613): the difference in
means = -.761 (CI [-1.157 to -.364], and this difference is significantly different from zero (p <
.05).

So yes, the expectation regarding the main effect of Brand Strategy on Attitude toward the
Brand Product is met.



4. Besides the main effects and differences between groups, you are obviously interested
in the interaction effect between Branding Strategy and Brand Equity.
A) Is there significant interaction in this model? Report the relevant F-statistic and p-
value, and draw the conclusion.
MRM2 – PC Lab 2 – Open book assignment

5

B) Which indication of effect size for the interaction does SPSS report? Please check the
effect size by manually calculating it and report your calculation below.

A. There is a significant interaction effect of Branding Strategy and Brand Equity on attitude
towards the product (F[2,193] = 7.15, p < .05).

B. Partial η2 = SStreatment / (SStreatment + SSR) = 19,318 / (19,318 + 260,767) = 0,069.
According to the rule of thumb, this is a medium effect size.

5. You now want to interpret this interaction effect. You have SPSS plot a graph depicting
the interaction between the PVs.
A) Based on the information available, which PV would you preferably place on the
horizontal axis, and which do you plot as separate lines?
B) Briefly describe the interaction pattern for this model (you can also copy/paste the
graph).

A. The conceptual model shows that we expect the effect of Branding Strategy on attitude
to be moderated by Brand Equity, so we put Branding Strategy on the horizontal axis
and Brand Equity as separate lines.
B.


Firstly, for completeness we review the main effects.
You can see an indication of a main effect of both Branding Strategy and Brand Equity.
For Brand Equity, we see that High Equity consistently scores higher than Low Equity,
regardless of which type of brand strategy we look at.
MRM2 – PC Lab 2 – Open book assignment

6

For Branding Strategy we see a consistent upward trend from Category extension to Line
Extension, to Brand Alliance.

Moving on to the interaction effect.
For both High and Low Brand Equity, Line Extension is evaluated more favorably than a
Category Extension, while a Line Extension is evaluated less favorably than a Brand Alliance.
However, the increase in evaluation from Line Extension to a Brand Alliance is much larger
for Low Equity brands than for High Equity brands.
In other words, a low equity brand benefits much more from a brand partner (brand
alliance) than a high equity brand, when comparing to other types of branding strategies.

In that sense, we see how the effect of PV1 Branding Strategy (‘effect’ in the sense of the
differences in scores on attitudes between the branding strategies)
is not the same
for different levels of PV2 (Brand Equity, where the levels are high vs. low).

PV2, Brand Equity, moderates the effect of PV1 (Branding Strategy) on the OV (Attitudes
towards Branded Products).

6. Finally, you want to do a follow-up test to further qualify the interaction effect you
expected.
A) report the lines of code you add to the syntax to conduct the correct follow-up test
for the moderation effect that you expect (think about which variable is the moderator).
B) Based on the output, between which groups do we see significant differences? Are
the expectations met?
A.
/EMMEANS=TABLES(BrandStrat*BrandEquity) compare (BrandStrat)

We are thus comparing the effect of Branding Strategy on Attitudes, but separately for each
level of Brand Equity (high and low), as Brand Equity is our moderator.
MRM2 – PC Lab 2 – Open book assignment

7


B.
• For low equity brands,
Brand Alliances (mean = 5.469) lead to higher attitudes than Line Extensions (mean =
4.029), mean difference = 1.1440, p < .05, CI (.754 to 2.127)
Brand Alliances (mean = 5.469) lead to higher attitudes than Category Extensions (mean
= 3.029), mean difference = 2.439, p < .05, CI (1.748 to 3.131)

• For high equity brands,
Line Extensions (mean = 5.676) lead to higher attitudes than Category Extensions (mean
= 4.613, mean difference = 1.064, p < .05, CI (.366 tot 1.761)
Brand Alliances (mean = 5.758) lead to higher attitudes than Category Extensions (mean
= 4.613, mean difference = 1.145, p < .05, CI (.443 to 1.847)
There is no significant difference between Line Extensions (mean = 5.676) and Brand
Alliances (mean = 5.758), mean difference = -.81, p = .776, CI (-.767 to .605).

The expectation was:
H1: The effect of brand strategy on attitude is moderated by brand equity, such that:
- for weak brands (low equity brands), Category extensions will have lower Attitude towards
Branded Products than Line Extensions or Brand Alliances
MRM2 – PC Lab 2 – Open book assignment

8

→ this expectation is met as for low equity brands:
µcategory extension < µline extension < µbrand alliance



- for strong brands (high equity brands), Attitude towards Branded Products will not differ
across the three branding strategies
→ this expectation is partially met as for high equity brands:
mcategory extension < µline extension
µcategory extension < µbrand alliance
µline extension = µbrand alliance
So only for line extensions and brand alliances, the difference in attitude means do no differ.
Full syntax:

UNIANOVA Attitude BY BrandStrat BrandEquity
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/PLOT=PROFILE(BrandStrat*BrandEquity) TYPE=LINE ERRORBAR=NO MEANREFERENCE=NO
/EMMEANS=TABLES(BrandStrat) compare (BrandStrat)
/EMMEANS=TABLES(BrandEquity) compare (BrandEquity)
/EMMEANS=TABLES(BrandStrat*BrandEquity) compare (BrandStrat)
/PRINT ETASQ DESCRIPTIVE HOMOGENEITY
/CRITERIA=ALPHA(.05)
/DESIGN=BrandStrat BrandEquity BrandStrat*BrandEquity.