Basic HTML Version
Q
uantitative
R
easoning &
P
roblem
S
olving
150
© 2014 Pacific Crest
Scenario:
Scientific: Testing for drug efficacy for different drugs
Step
Watch it Work!
1.
Identify the hypothesis
or question you want to
test
What percentage of women with stage 4 breast cancer benefit from
an oral chemotherapy treatment?
2.
Determine the target
population
Stratified random sample; we want to see how age groups differ 20’s,
30’s, 40’s, and 50’s
3.
Determine an
appropriate sampling
strategy
Probabilistic
4.
Determine appropriate
sampling technique
Stratified random sample
5.
Determine appropriate
sample
Separate the selected women into two equal sized treatment and
control groups.
6.
Validate sampling
procedure
If all women have the same health status as a pre-participation
condition, the random sampling procedure should be unbiased.
7.
Test your hypothesis See Chapter 7
YOUR
TURN!
Scenario:
Social Science: Household Survey
O
ops
! A
voiding
C
ommon
E
rrors
●
Sample Selection Bias
Example
: A student wants to know what course the student body finds the most difficult at
the college. I determine it is Corporate Finance after sampling a class of business
students.
Why?
The answer
Corporate Finance
, represents the most difficult course for the sample
of business students; this sample does not necessarily represent the student body who
are not all business students.
●
Not selecting the appropriate sample size to estimate a sample statistic within an acceptable
margin of error
Example
: A company that produces 100,000 small wheels per day wants to ensure that the
average diameter of the wheels they produce is 5 inches. They measure 100 wheels
and compute the average diameter. The company wants the margin of error of the
estimate to be at most 5%.
Why?
If the company measures 100 wheels, they are only sampling less than one percent
of the population. This sample size will not produce an accurate estimate of the true
average diameter, that is, the margin of error will be very large. According to the
reading, if we have a population of 100,000, with a margin of error of 5 percent