Statistics 101A
Lab 4
Professor Esfandiari
Winter 2011
The objective of this lab is to show you:
•
Under what conditions use randomized block design
•
How to enter the data resulting from RBD into SPSS
•
How to analyze the data resulting from an RBD design
•
How to interpret the printouts resulting from an RBD design
I
Randomized block design can happen when:
I.1
Case I
o
You repeatedly measure the same subject more than twice.
o
Each subject is like a block.
o
Subjects serve as their own control.
o
Data resulting from this design are dependent and correlated.
Example:
A trainer wants to find out whether the body fat of his trainees has
changed over the course of six months. He has them follow a specific diet and
exercise program and measures their body fat at three time points: two months,
four months and six months after following the diet and exercise program.
I.2
Case II
•
You want to control for a confounding factor, such as prior knowledge of
math, prior weight, etc. by blocking.
•
You measure the subjects based on the confounding factor; prior weight,
prior knowledge of math, and then you rank them from the highest to
lowest on the confounder.
•
If you have three treatments, you randomly assign the subjects with ranks
13 to the three treatments, you do the same with the next three subjects
(46) and continue till you have made 30 blocks of three subjects.
•
This way you ascertain that the average of the confounder, weight or the
average knowledge of math, is similar for the subjects assigned to the
three treatments. Thus, blocking has helped to control the confounding
factor.
•
In case I.1 each block consisted of the same subject. In this case each
block consists of three different subjects that are highly similar with
respect to the confounder.
•
The treatments are randomly assigned to the subjects.
•
The objective is to create maximum homogeneity within blocks and
1
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Documentminimum homogeneity between different blocks.
1.3
Case III
o
You have a number of treatments (say 3) and you want every subject to
experience all of the three treatments.
o
Each subject will be one block
o
You have every subject experience all of the three treatments.
o
You randomly change the order of the treatment for each of the blocks.
o
The difference between this case and case I.1 is that in case I.1 each
subject is repeatedly measured on the same thing or repeatedly does the
same task. But, here each subject
Example:
A company has manufactured a soft drink wit three different flavors
(orange, grape, and blueberry). They want to see which drink the ten year olds
prefer. They randomly pick twenty ten year olds and have each of them try all of
the three different flavors. But, they randomize the order in which each subject is
given the drink. Then they ask each subject which drink they preferred.
II
This is the end of the preview.
Sign up
to
access the rest of the document.
 Winter '11
 MahtashEsfandiari
 Statistics, Spss, Standard Deviation, Kirk, ethnic background

Click to edit the document details