Chapter 11

ANALYSIS OF VARIANCE (ANOVA)

11.1 An experiment has a single factor with five groups and seven values in each group.
a. How many degrees of freedom are there in determining the among-group variation?
b. How many degrees of freedom are there in determining the within-group variation?
c. How many degrees of freedom are there in determining the total variation?

11.2 You are working with the same experiment as in Problem 11.1.
a. If SSA = 60 and SST = 210, what is SSW?
b. What is MSA?
c. What is MSW?
d. What is the value of FSTAT?

11.3 You are working with the same experiment as in Problems 11.1 and 11.2.
a. Construct the ANOVA summary table and fill in all values in the table.
b. At the 0.05 level of significance, what is the upper-tail critical value from the F distribution?
c. State the decision rule for testing the null hypothesis that all five groups have equal population
means.
d. What is your statistical decision?

11.4 State the assumptions about the population to use a one-way ANOVA F test. What alternative
tests can be used if
a. only the assumption of normality is violated.
b. only the assumption of homogeneity-of-variance is violated.
c. both the assumptions of normality and homogeneity-of variance have been violated.

11.5 Consider an experiment with four groups, with eight values in each. For the ANOVA summary
table below, fill in all the missing results:

11.6 You are working with the same experiment as in Problem 11.5.
a. At the 0.05 level of significance, state the decision rule for testing the null hypothesis that all four
groups have equal population means.
b. What is your statistical decision?
c. At the 0.05 level of significance, what is the upper-tail critical value from the Studentized range
distribution?
d. To perform the Tukey-Kramer procedure, what is the critical range?

1
11.7 It is hypothesized that the environment in which students appear for a test might impact their
scores. An experiment was designed to test the effects of music and temperature in the classroom in
which a midterm test was being conducted. Twenty students with similar scores in a previous midterm
test in the same subject were randomly selected out of 40 students. They were then randomly divided
into four groups. Group A had music in the room and a temperature of 22 degrees Celsius during the
midterm test. Group B had no music and 22 degrees Celsius. Group C had music and 25 degrees
Celsius. Group D had no music and 25 degrees Celsius. The file Midterm contains the scores of the
students on the basis of their group (the maximum score was 50).
a. At the 0.05 level of significance, is there evidence of a difference in the mean midterm scores of the
different groups?
b. If appropriate, determine which group differs in mean midterm scores.
c. At the 0.05 level of significance, is there evidence of a difference in the variation in the mean midterm
scores of the different groups?

11.10 A manufacturer of pens has hired an advertising agency to develop an advertising campaign for
the upcoming holiday season. To prepare for this project, the research director decides to initiate a
study of the effect of advertising on product perception. An experiment is designed to compare five
different advertisements. Advertisement A greatly undersells the pen’s characteristics. Advertisement
B slightly undersells the pen’s chara Advertisement C slightly oversells the pen’s characteristics.
Advertisement D greatly oversells the pen’s characteristics. Advertisement E attempts to correctly
state the pen’s characteristics. A sample of 30 adult respondents, taken from a larger focus group, is
randomly assigned to the five advertisements (so that there are 6 respondents to each advertisement).
After reading the advertisement and developing a sense of “product expectation,” all respondents
unknowingly receive the same pen to evaluate. The respondents are permitted to test the pen and the
plausibility of the advertising copy. The respondents are then asked to rate the pen from 1 to 7 (lowest
to highest) on the product characteristic scales of appearance, durability, and writing performance.
The combined scores of three ratings (appearance, durability, and writing performance) for the 30
respondents, stored in Pen , are as follows:

a. At the 0.05 level of significance, is there evidence of a difference in the mean rating of the pens
following exposure to five advertisements?
b. If appropriate, determine which advertisements differ in mean ratings.
c. At the 0.05 level of significance, is there evidence of a difference in the variation in ratings among
the five advertisements?
d. Which advertisement(s) should you use, and which advertisement(s) should you avoid? Explain.

11.12 Brand valuations are critical to CEOs, financial and marketing executives, security analysts,
institutional investors, and others who depend on well-researched, reliable information needed for
assessments and comparisons in decision making. Millward Brown Optimor has developed the
BrandZ Top 100 Most Valuable Global Brands for WPP, the world’s largest communications services
group. Unlike other studies, the BrandZ Top 100 Most Valuable Global Brands fuses consumer
measures of brand equity with financial measures to place a financial value on brands. A research
assistant compared brand values for three sectors in the BrandZ Top 100 Most Valuable Global
Brands for 2016: the financial institution sector, the technology sector, and the telecom sector. The
research assistant findings were as follows:
2
a. Complete the ANOVA summary table.
b. At the 0.05 level of significance, is there evidence of a difference in mean brand value among the
sectors?
c. If the results in (b) indicate that it is appropriate, use the Tukey Kramer procedure to determine
which sectors differ in mean rating. Discuss your findings.

11.15 Consider a two-factor factorial design with three levels for factor A, three levels for factor B,
and four replicates in each of the nine cells.
a. How many degrees of freedom are there in determining the factor A variation and the factor B
variation?
b. How many degrees of freedom are there in determining the interaction variation?
c. How many degrees of freedom are there in determining the random variation?
d. How many degrees of freedom are there in determining the total variation?

11.16 Assume that you are working with the results from Problem 11.15, and SSA = 120, SSB =
110, SSE = 270, and SST = 540.
a. What is SSAB?
b. What are MSA and MSB?
c. What is MSAB?
d. What is MSE?

11.17 Assume that you are working with the results from Problems 11.15 and 11.16.
a. What is the value of the FSTAT test statistic for the interaction effect?
b. What is the value of the FSTAT test statistic for the factor A effect?
c. What is the value of the FSTAT test statistic for the factor B effect?
d. Form the ANOVA summary table and fill in all values in the body of the table.

11.18 Given the results from Problems 11.15 through 11.17,

a. at the 0.05 level of significance, is there an effect due to factor A?
b. at the 0.05 level of significance, is there an effect due to factor B?
c. at the 0.05 level of significance, is there an interaction effect?

11.20 Given a two-factor factorial experiment and the ANOVA summary table that follows, fill in all
the missing results:

3
11.21 Given the results from Problem 11.20,
a. at the 0.05 level of significance, is there an effect due to factor A?
b. at the 0.05 level of significance, is there an effect due to factor B?
c. at the 0.05 level of significance, is there an interaction effect?

11.22 Continuing from the data in Problem 11.14, the travel strategist now selects another sample of
flights traveling the same distance, between New York and Houston. In addition to the four seasons
of the year, he also records the time of the day in which each flight departed (morning or afternoon).
The data is stored in Delay2 .
At the 0.05 level of significance,
a. is there an interaction between season and departure time?
b. is there an effect due to season?
c. is there an effect due to departure period?
d. Plot the mean delay for each season for each departure period.
e. What can you conclude about the effect of season and departure period on mean delay?

11.23 Continuing from the data in Problem 11.11, the research group selects another random
sample for observation. This time the participants had to record if they had coffee, tea, or neither two
hours before they went to sleep and their mood upon waking the next morning. The file Sleep2
contains this data along with each participant’s sleep duration.
At the 0.05 level of significance,
a. is there an interaction between beverage and mood?
b. is there an effect due to beverage?
c. is there an effect due to mood?
d. Plot the mean delay for each beverage for both moods.
e. What can you conclude about the effect of beverage and mood on mean sleep duration?

11.25 A glass manufacturing company wanted to investigate the effect of breakoff pressure and
stopper height on the percentage of breaking off chips. The results, stored in Glass1 , were as
follows:

At the 0.05 level of significance,

a. is there an interaction between the breakoff pressure and the stopper height?
b. is there an effect due to the breakoff pressure?
c. is there an effect due to the stopper height?
d. Plot the percentage breakoff for each breakoff pressure for each stopper height.
e. Discuss the results of (a) through (d).