You are on page 1of 3

Stat 1001

Winter 1998
Geyer

Homework 3
Problem 6.1
False. This was widely believed into the 1700's. The main point of this chapter
is that there is always measurement error, even under the best of circumstances.
Some of this is always chance error, varying from measurement to measurement.

Problem 6.3
(a) False. Bias is systematic error, the same in every measurement.
(b)False. Bias is the same in every measurement. Chance error varies from
measurement to measurement.
(c) True. That's what the pseudo-equation on p. 97 says.

Problem 6.4
The average of the three measurements is 36.00 inches and the SD is .029 inches.
You would expect another measurement to be o by about one SD or about .03
inches.

Problem 6-S.1
False. Not everything follows the normal curve. This is a problem only if you
want to use the normal curve to analyze the test scores. It doesn't mean there
is anything \wrong" with the test from any other perspective.

Problem 6-S.2
(a)(ii) all the numbers on the list are the same. An SD of zero means no spread
among the numbers: all equal to the average.
(b) (iii) all the numbers on the list are zero. (ii) and (iv) are also true, but they
don't give as much information about the list as (iii).

1
Problem 6-S.14
The improper randomization does indeed bias the study.
Assigning high risk women to the screening group increases the number
of cancers in the screening group and decreases the number of cancers in the
control group. Since the study looked at death rates, not detection rates, this
will make the screening group look worse, even though many of the cancers in
the screening group will be treated and cured. The biased assignment means
that there will be fewer deaths in the control group than there would have been
if the randomization had been done correctly and that there will be more deaths
in the screening group.

Problem 8.1
(d). The point of averages for (a) and (c) do not seem to be (100,100). Moreover
(c) seems to have a higher correlation than 0.6, more like 0.9, and bigger SDs
than 15, more like 30. The SDs for (b) are too small; the cloud doesn't go 2
SDs out from the mean.

Problem 8.3
This would be perfect positive linear association, = 1 00.
r :

Problem 8.4
There is some positive association, but the correlation is nowhere near 0 9. So
:

r= 0 3 is the right answer.


:

Problem 8.7
The ones with negative association are upper right, middle left, and lower right.
Matching the available negative numbers with these scatterplots in order of
decreasing association gives ,1 0 for upper right, ,0 85 for middle left, and
: :

,0 38 for lower right. The others must be, again going in order of decreasing
:

association 0 97 for middle right, 0 62 for upper left, and 0 06 for lower left.
: : :

Problem 8.8
(a) 42 inches.
(b) 2.5 inches.
(c) 0.80.
(d) solid.

2
Problem 8.11
There is a perfect linear association wrong = total , right. The association is
negative, so = ,1 00.
r :

You might also like