Math 1070-Chapter 27 Notes One-Way Analysis of Variance (ANOVA) : Comparing Several Means by Using ANOVA F Test

Math 1070-Chapter 27 Notes
One-Way Analysis of Variance(ANOVA):
Comparing Several Means by using ANOVA F test
1 Comparing Ch 21. with Ch 27.
Ch 21 Ch 27
The number of comparing populations(or groups) 2 3 (or more)
H0 µ1 = µ2 µ1 = µ2 = µ3
µ1 6= µ2 H0 is not true.
Ha µ1 < µ2 =
µ1 > µ2 Not all of µ1 , µ2 , and µ3 are equal.
Significance Test t test ANOVA F test
Test statistic t = rxs12−x2s2 F = MSG
MSE
1+ 2
n1 n2
2 Example
Table 27.1 gives length measurements (in millimeters) for samples of three va-
rieties of Heliconia, each fertilized by a different species of hummingbird. Do
the three varieties display distinct distributions of length? In particular, are the
mean lengths of their flowers different at α = 0.05 ?
1
2.1 Preliminary works
2.1.1 The summary of Table 27.1
2
2.1.2 Conditions for using ANOVA inference
In the second condition, ANOVA becomes safe if the sample sizes is large
even though there is not normality assumption , because of the CLT. The third
condition is annoying: It is not easy to check the condition that the populations
have equal standard deviations. Here is a rule of thumb that is safe in almost
all situations.
largest sample sd
smallest sample sd <2
=⇒ We can safely use ANOVA F test.
Let’s check if standard deviations above satisfy rule of thumb. If yes then we
can safely use ANOVA.
3
2.1.3 F distribution and F Table
Our brief notation will be F (df1, df2) for the F distribution with df1 degrees
of freedom in the numerator and df2 in the denominator.
4
2.2 The ANOVA F test
2.2.1 Step 1) Make hypotheses
H0 : µ1 = µ2 = µ3 vs. Ha : Not all of µ1 , µ2 , and µ3 are equal
2.2.2 Step2) Critical value of F
• degree of freedoms in F (df1, df2)
– df1: degrees of freedom in the numerator = I − 1,
– df2: degrees of freedom in the denominator = N − I
– ,where I is the number of populations and N is the total number of
observations.
– In the example above, I = 3, N = 16 + 23 + 15 = 54.
Therefore, df1 = 3-1 = 2, and df2 = 54-3 = 51 .
• Since F Table does not have df2 = 51, we take df2 = 50 close to 51 and
α = 0.05, F ∗ = 3.18. The right area of 3.18 is the rejection region.
5
2.2.3 Step 3) ANOVA F statistic
Overall mean x is the mean of all N observations together. You can find x
from the I sample means by
• The basic ANOVA table
6
• How to calculate
– The overall mean x

n1 x1 +n2 x2 +n3 x3 (16)(47.60)+(23)(39.71)+(15)(36.18) 2217.63
x= N = 54 = 54 = 41.067
– The mean squre for group (MSG)

n1 (x1 −x)2 +n2 (x2 −x)2 +n3 (x3 −x)2
MSG = I−1
(16)(47.60−41.067) +(23)(39.71−41.067)2 +(15)(36.18−41.067)2
2
= 3−1
1083.476286
= 2 = 541.74
– The mean square for error (MSE)

(n1 −1)s21 +(n2 −1)s22 +(n3 −1)s23
MSE = N −I
(16−1)(1.213 )+(23−1)(1.7992 )+(15−1)(0.9752 )
2
= 54−3
106.580
= 51 = 2.09
– The ANOVA F statistic
MSG 541.74
F = MSE = 2.09 = 259.20
7
• In Excel
Our work differs slightly from the output in Excel because of roundoff error.
2.2.4 Step 4) Conclusion
• Method 1)
F ∗ < F implies that F is in the rejection region. Hence, we reject H0 .
• Method 2)
Since P −value = P (F (2, 51) > 259.20) ≈ 0, which means that α > P −value,
we reject H0 .
• Not all of µ1 , µ2 , and µ3 are equal.
8
3 Exercise
Do high school students from 3 different racial/ethnic groups have different
attitudes toward mathematics? Measure the level of interest in mathematics on
a 5-point scale for a national random sample of students. Here are summaries
for students who were taking math at the time of the survey assuming that all
three samples are from normal distribution populations respectively.
Racial/ethnic group n x s
African American 10 2.57 1.40
White 14 2.32 1.36
Asian/Pacific Islander 6 2.63 1.32
(a) Can we safely use ANOVA F test to compare the mean for the three popu-
lations?
(b) Conduct the ANOVA F test at α = 0.05.
9
10

Math 1070-Chapter 27 Notes One-Way Analysis of Variance (ANOVA) : Comparing Several Means by Using ANOVA F Test

Uploaded by

Copyright:

Available Formats

Math 1070-Chapter 27 Notes One-Way Analysis of Variance (ANOVA) : Comparing Several Means by Using ANOVA F Test

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Math 1070-Chapter 27 Notes One-Way Analysis of Variance (ANOVA) : Comparing Several Means by Using ANOVA F Test

Uploaded by

Copyright:

Available Formats

Math 1070-Chapter 27 Notes

One-Way Analysis of Variance(ANOVA):

Comparing Several Means by using ANOVA F test

1 Comparing Ch 21. with Ch 27.

rieties of Heliconia, each fertilized by a different species of hummingbird. Do

mean lengths of their flowers different at α = 0.05 ?

2.1.1 The summary of Table 27.1

=⇒ We can safely use ANOVA F test.

can safely use ANOVA.

of freedom in the numerator and df2 in the denominator.

2.2.1 Step 1) Make hypotheses

H0 : µ1 = µ2 = µ3 vs. Ha : Not all of µ1 , µ2 , and µ3 are equal

2.2.2 Step2) Critical value of F

• degree of freedoms in F (df1, df2)

– df1: degrees of freedom in the numerator = I − 1,

– df2: degrees of freedom in the denominator = N − I

– ,where I is the number of populations and N is the total number of

– In the example above, I = 3, N = 16 + 23 + 15 = 54.

Therefore, df1 = 3-1 = 2, and df2 = 54-3 = 51 .

α = 0.05, F ∗ = 3.18. The right area of 3.18 is the rejection region.

from the I sample means by

• The basic ANOVA table

– The overall mean x

– The mean squre for group (MSG)

– The mean square for error (MSE)

– The ANOVA F statistic

2.2.4 Step 4) Conclusion

F ∗ < F implies that F is in the rejection region. Hence, we reject H0 .

• Not all of µ1 , µ2 , and µ3 are equal.

Do high school students from 3 different racial/ethnic groups have different

attitudes toward mathematics? Measure the level of interest in mathematics on

three samples are from normal distribution populations respectively.

(b) Conduct the ANOVA F test at α = 0.05.

You might also like