Design and Analysis

of Experiments

Chapter 7

Blocking and Confounding

(Đóng khối và gây nhiễu)

in the 2k Factorial Design

Dr. Tran Thanh Hung
Department of Automation Technology,
College of Engineering, Can Tho University
Email: tthung@ctu.edu.vn
 Chapter objectives

At the end of this chapter, students can

understand:
• Concept of Confounding
• Blocking of replicated 2k factorial designs
• Confounding high order interaction effects of
the 2k factorial design in 2p blocks
• How to choose the effects to be confounded
with blocks

2
 Blocking a Replicated Design

• Consider 2k  factorial design, n replications. In

many situations , we can not runs all experiments
under homogeneous conditions; or we want to vary
the experimental conditions to test the robustness.
• We can use these n replicates as blocks, and
assign the 2k treatments to the experimental units
within each of the n blocks
• Each replicate is run in one of the blocks (time
periods, batches of raw material, etc.)
• Runs within the block are randomized
3
 Blocking a Replicated Design

Consider again the

Chemical Process
example in chapter
6:
k = 2 factors,
n = 3 replicates Block sum of squares
3
Bi2 y...2
Each batch of raw SS Blocks  
material corresponds i 1 4 12
to a block.  6.50
Other sums of squares are calculated
exactly as in a standard, unblocked 2k
design 4
 ANOVA for the Blocked Design

Compare with the result in Chapter 6:

5
 Confounding (gây nhiễu) in Blocks

• In many problems it is impossible to perform a

complete replicate of a factorial design in one block 
incomplete block designs

• Confounding is a design technique for arranging a

complete factorial experiment in blocks, where the
block size is smaller than the number of treatment
combinations in replicate case.

• Confounding  information about certain treatment

effects (usually high order interactions) to be
indistinguishable from, or confounded with, blocks
6
 Confounding the 2k Factorial
Design in Two Blocks
• Suppose that we wish to run a single replicate of the 22
design  4 treatment combinations  need raw material.
• However, each batch of raw material is only large enough
for two treatment combinations to be tested  need 2
batches of raw material.
• If batches of raw material are considered as blocks, then
we must assign two of the four treatment combinations to
each block.

7
 Confounding the 2k Factorial
Design in Two Blocks

Suppose we estimate the main effect

of A, B just as if no blocking:
1 - Both A and B are unaffected by
A  [ab  a  b  (1)] blocking because in each estimate
2
there is one plus and one minus
1
B  [ab  b  a  (1)] treatment combination from each block.
2  Any difference between block 1 and
block 2 will cancel out. 8
 Confounding the 2k Factorial
Design in Two Blocks
Interaction effect:
1
AB  [ab  (1)  a  b]
2
Block effect:
Block = y Block1  y Block 2

- Two treatment combinations with the plus sign [ab and

(1)] are in block 1
- The two with the minus sign (a and b) are in block 2,
 The block effect and the AB interaction are identical.
That is, AB is confounded with blocks.
9
 Confounding the 2k Factorial
Design in Two Blocks
• This scheme can be used to confound any 2k design in
two blocks.
• Consider a 23 design run in two blocks. Suppose we wish
to confound the 3 factor interaction ABC with blocks.

10
Confounding the 2k Factorial
Design in Two Blocks

11
 Experiment from Example 6-2:
filtration rate of a resin
This is a 24, n = 1 replicate. 4 factors: temperature (A), pressure (B),
concentration of formaldehyde (C), and stirring rate (D)
Suppose only 8 runs can be made from one batch of raw material

12
The Table of + & - Signs

13
 ABCD is Confounded with
Blocks

Suppose that when we select the two batches of raw material required to
run the experiment, one of them is of much poorer quality and, as a result,
all responses will be 20 units lower in this material batch (block 1) than in
the other.
14
This is the simulated “block effect”
 Effect Estimates

Block effect = y Block1  y Block 2 15

The ANOVA

16
 Another Illustration of the Importance of
Blocking

If the experimenter doesn’t block when he or she should have

Now the first

eight runs
(in run
order, that is
Block 1, if
blocking)
have
filtration rate
reduced by
20 units

17
The interpretation is
harder; not as easy to
identify the large
effects
One important
interaction is not
identified (AD)
Failing to block when
we should have causes
problems in
interpretation the result
of an experiment and
can mask the presence
of real factor effects18
 Confounding in Blocks

• More than two blocks (page 313)

– The 2k factorial design can be confounded
in 2, 4, 8, … (2p, 1 < p < k) blocks
– For four blocks, select two effects to
confound, automatically confounding a
third effect
– See example, page 313
– Choice of confounding schemes non-trivial;
see Table 7-9, page 315
• Partial confounding (page 316)
19
20
 General Advice About Blocking

• When in doubt, block

• Block out the nuisance variables you know about,
randomize as much as possible and rely on
randomization to help balance out unknown nuisance
effects
• Measure the nuisance factors you know about but
can’t control
• It may be a good idea to conduct the experiment in
blocks even if there isn't an obvious nuisance factor,
just to protect against the loss of data or situations
where the complete experiment can’t be finished
21
 Thực hành chương 7

• Bài 1: Dùng Minitab phân tích kết quả thí

nghiệm sau. Tìm các yếu tố quan trọng.

22
 Thực hành chương 7

• Bài 2: Thiết kế thí nghiệm xếp máy bay giấy

với 2 yếu tố hình dạng mũi và độ rộng cánh,
mỗi yếu tố 2 mức, một lần lập (xếp). Giả sử
không đủ giấy để xếp, phải dùng 2 loại giấy
khác nhau, nên cần đóng khối theo loại giấy.
• Nếu phóng thử mỗi máy bay 5 lần thì phân tích
kết quả thế nào?

23