Assigned 4/7; due 4/14; returned 4/21. Point items (random sample) shown in bold. There are 2 free points.
(17.2) Data do not support the
hypothesis that cell phone use causes intracranial tumors.
Explanation: The two primary odds ratios from the studies are 1.0. The
other reported odds ratios are either below 1 or not significantly above
1.
(17.4) BD2
(A) Cross-tabulated data in SPSS -- distributed via WebCT
(B) Calculate the odds ratio and its 95% confidence interval.
y^ = 1.650
lny^ = 0.5011;
selny^ = 0.05587;
95% CI for lny = 0.5011 � (1.96)(0.05587) =
(0.3916, 0.6106)
95% CI for y = e(0.3916, 0.6106)
= (1.48, 1.84). Interpretation should address the parameter.
(C) Test of
H0: y = 1.
Show all work.
Yates' continuity corrected chi-square test
c2Yates'
= (|932 - 767|-0.5)2 / 767 +
(602 - 767|-0.5)2 / 767 +
(|4994
- 5159|-0.5)2 / 5159
+ (|5324 - 5159|-0.5)2 / 5159
=
35.281
+ 35.281
5.245
+ 5.245
= 81.052
df = 1; c = sqrt(81.092) = 9.00; p = off chart (less than .00001)
SPSS output was distributed via WebCT.
(17.8) BD1
y^1 = (75)(386) / (280)(29) = 3.57
y^2 = (51)(386) / (87)(29) = 7.80
y^3 = (45)(386) / (22)(29) = 27.23
Yes, there is evidence of a dose-response relation between alcohol consumption and esophageal cancer risk.
(17.14) y^ = 6.1431;
ln(y^) = 1.815;
se lny^ = 0.3794;
95% CI for lny = 1.815 � (1.96)(0.1794) =
1.815 � 0.7436 = (1.071, 2.559)
95% CI for OR = (2.92, 12.92); Chi-sqMcN, corrected = 24.50, df =
1; Chi = 4.95; p < .0000001
Interpretation: Data show a significant odds ratio of 6.14 (conjugated estrogen
use associated with 6-fold risk of cervical cancer). The confidence interval
places the odds ratio parameter between 2.9 and 12.9.