parameters. In other words, equals the tthat you
would get if the data exactly match what you think are
the values of the parameters.
8.13 This is a graphic.
8.15 He should use the dropout group. Assuming equal stan-
dard deviations, the H.S. dropout group of 25 would re-
sult in a higher value of and therefore higher power.
(You can let be any value as long as it is the same for
both calculations. Then calculate for each situation.)
8.17 Power 5 .49.
8.19 When 5 104.935, power will equal.
8.21 (a) I would not be assigning subjects to groups at ran-
dom, and there might be differences between labs
that would confound the results.
(b) I should pool my subjects and randomly assign
them to conditions.
(c) Sex differences, if they exist, would confound the re-
sults. We would need to use a procedure (see Chap-
ter 13) that separates any sex differences and looks
for different patterns of results in males and females.
8.23 Both of these questions point to the need to design
studies carefully so that the results are clear and
interpretable.
Chapter 9
9.1 This is a graphic.
9.3 r 5 .35
9.5 This is a graphic.
9.7 r 5 .99, .71, and 2 .99. Three arrangements will work:
2 8 6 4 or 6 4 2 8 or 6 2 8 4r 5 .14 for each
9.9 (a) d 5 .20, 5 .98, power 5 .17
(b) N 5197
9.11
9.13 If the high-risk fertility rate jumped to 70, we would
predict 8.36% of infants would be LBW.
sY.X=0.580dm bdsdd 9.15 The predicted value for ln(symptoms) 5 4.37.9.17You can calculate for several different values
of X, and then plot the results.
9.19 When the data are standardized, the slope equals r.
Therefore the slope will be less than 1 for all but the
most trivial problems, and predicted deviations from the
mean will be less than actual parental deviations.
9.21 For power 5 .80, 5 2.80. Therefore N 5 50.
9.23 (a) z 5 0.797. The correlations are not significantly
different.
(b) We do not have reason to argue that the relation-
ship between performance and prior test scores is
affected by whether or not the student read the
passage.
9.25 It is difficult to tell whether the significant difference is
to be attributable to the larger sample sizes or the higher
(and thus more different) values of. It is likely to be
the former.
9.27 (a) r 5 .224, p 5 .509. Do not reject H 0.
(b) The Irish are heavy smokers, but they certainly are
not heavy drinkers compared to other regions.
(c) The inclusion of Northern Ireland distorts the data.
If we leave them out, r 5 .784, p 5 .007, and there
is a strong relationship between smoking and
drinking.
9.29 (a) See table below.
(b) All of these correlations are significant, showing
that the symptoms are correlated with one another.
9.31 (a) This is a graphic.r¿dYN and sY¿.XCI(Y)=YN 6 (ta> 2 )sY¿.Xta> 2 =1.984s¿Y.X=0.1726
B
11
1
107
1(Xi 2 X)^2
(N 2 1)s^2 X742 Answers
SomT ObsessT SensitT DepressT AnxT HostT PhobT ParT PsyTObsessT 0.482
SensitT 0.377 0.539
DepressT 0.400 0.599 0.654
AnxT 0.569 0.621 0.550 0.590
HostT 0.420 0.470 0.451 0.508 0.475
PhobT 0.466 0.509 0.613 0.568 0.528 0.411
ParT 0.400 0.524 0.677 0.621 0.547 0.494 0.540
PsyT 0.334 0.503 0.625 0.725 0.509 0.404 0.529 0.651
GSIT 0.646 0.791 0.770 0.820 0.786 0.633 0.679 0.766 0.741