1 It is often unethical to experiment with human
health.
2 Measurements in ecological studies do not entail follow-up of
individuals over time. Therefore, ecological studies are cross-sectional.
1 person-time
2 This is an aggregate-level observation.
3 This means the effects of dietary fat could be mixed-up with the
effects of other differences tin the populations.
4 A synonym for aggregation bias is the ecological fallacy.
5 Aggregation bias occurs when an association seen in
aggregate data does not hold for data based on individual data.
6 This occurred because high literacy (northern) states had high
immigration rates. This is an example of the ecological fallacy.
7 Farr failed to account for the fact that people living a lower
elevations were more likely to drink contaminated water.
8 True
9 False. Aggregative properties are summaries of characteristics of
smaller units.
10
(1) contextual variable
(2) integral variable
(3) contagion variable
11 Larger hospitals may have more thorough or complete reporting..
1 False. "Longitudinalness" depends on whether you can place
events over time. (The temporal relation between occurrence and measurement is
referred to as proximity.)
2 Social explanations require classifications and classifications
require definitions of characteristics.
3 Notiones vulgares are crudely formed concepts about social and
natural phenomena without objective reflection and testing.
4 Alternative hypotheses concerning the apparent protective effects of
marriage on suicide: (a) the marriage environment is protective (b) people who
get married have characteristics that make them less likely to commit suicide.
It is possible both effects are operative.
5 The Marine Hospital Service
6 Many people believed pellagra was contagious.
7 See text pp. 200 - 202.
8
(1) prevalence-incidence bias
(2) detection bias
(3) reverse-causality bias (cart-before-the-horse bias)
1 A cohort is is a group of people having a common characteristic.
2 In cohort studies, the investigator selects people based
on exposure status. In non-randomized trails, the investigator assigns treatments
to people.
3 The calculation of mortality rates in open populations does not entail follow-up of individuals over
time.
4 TB's long latent period required methods that follow individuals over
extended periods of time. Such methods are needed for the study of chronic
diseases.
5 Framingham is a town in Massachusetts. It is the seat of the innovative,
long-term follow-up study that has yielded many important findings about heart
disease.
6 Bradford Hill and Richard Doll
7 Approx. 70 percent of light smokers survived to
age 70.
8 The proximity of prospective studies in concurrent. The proximity of
retrospective studies is historical (nonconcurrent).
9 Because the college was compiled from a roster the covered employment
history from the 1920s. (The study in question was completed in 1954.)
10 The investigator need not wait the many years needed for chronic
diseases to develop.
1 By studying the exposure experience of only a small fraction of
non-cases.
2 10 years
3 Levin did his study in a referral hospital. The frequency of smoking was high
in the control series perhaps because some were referred for smoking-related chronic
conditions.
4 T
5 doubles
6 True
7 Blank
8 Blank
9 c
10 True
11 Blank
12 Blank
13 Blank
14 k-1
15 Allows you to gather higher-quality information and more information on study subjects (p. 212, �5).
16 To allow for sensible selection of controls.
17 True.
18 Hospital catchment areas are open.
19 To allow for a small sample size adjustment. It is especially useful when there are zeros in table
cells.
20 Publicity could have produced a spurious association by stimulating
reporting of exposed cases while
unexposed cases went undiagnosed and unreported.
21 True.
22 1
23 True.