Homework 5
1. Textbook Questions
16.
This experimental study is not double blinded. Therefore bias can occur on both the side of
the researcher and the participant. The participant may feel more confident in the new suit
or feel pressured to perform better because the test is for marketing purposes. There is
also no control group, hence the experiment does not account for people simply swimming
faster the second time because they are more warmed up.
20.
a) This was an observational study.
b) Wine may not have any health benefits as proven by this study because the study merely
demonstrates a correlation between drinking wine, better health, and higher
socioeconomic status, but cannot demonstrate a causal relationship between drinking wine
and better health. Further, it is more likely that a higher socioeconomic status would
improve one’s health as one would have fewer stressors and more resources for medical
treatment.
c) These studies cannot prove causation because there is no way to account for lurking
variables, and there is no real control group.
28.
a) This is a multifactor experimental study.
b) The factors are gender and receiving a coupon or not.
c) The response variable is amount of expenditure.
d) no, because randomization, which is crucial to proper design, has not been mentioned
There is also no replication, so doing multiple weeks would improve it.
Blocking would also be helpful because it would randomize the treatment within the study.
30.
a) Subjects were randomly assigned to create a random distribution of lurking
variables. There is a good chance that people who would choose to eat a specific diet
are different in some way from people who would not.
b) The meals were prepared by dietitians to make sure they actually satisfied the
requirements of the diet and standardize meals
c) They needed the control group to compare the effect on blood pressure of the DASH
diet to a placebo diet. If the DASH diet group’s blood pressure was lower at the end
of the experiment than the beginning, it might not prove that the diet was effective
because the placebo effect and other factors could play a role in lowering their blood
pressure.
d) I would need to know the average reduction in blood pressure among the group fed
the control diet. I would also want to know more about the distribution of reduction
values. Mean is sensitive to outliers, so I would want to know the spread and the
IQR.
34.
a) The rebate works in gamers
b) There is an interaction effect because the slopes of the two lines are not the same.
c) Better targeted at gamers.
40.
2. Ethics In Action Page 300
Ethical Dilemma: Beth wants her training videos to improve the knowledge of all managers,
especially those in the Southeast where there are more problems. However, the managers
who received the training in the South East did worse than the managers in the same
region who received it.
The undesirable consequences of reporting that the training was unsuccessful would be
ignoring the West regional group, where the managers who received the training actually
did better than those in the same region. It may be that the managers in the South East
received incorrect prior training, thus the new training video confused them further.
Whatever the reason for the discrepancy, we do not want to discount the success the
training had on the managers in the West.
An ethical solution would be to report the differences in performance on the “test” by
region and work to develop more effective training for the managers in the South East.
Some strategies to gather data about their previous training could be an observational
study where they both self evaluate and are asked to share about their previous training.
Beth could also replicate the study to verify that the difference in performance is a true
pattern.
3. Okcupid
a. D
b. Probability male out of all pet owners
6552/14292
c. Probability cats only out of all females
3412/7740
d. D
6552/7018
4. Cell Phones and americans
a. A home or a cell phone
i. Home + cellphone – (home and cell phone)
There is a 95.41% probability that a randomly selected American has either a home or cell
phone
ii.
b. Neither a home nor a cell phone
i. 1- home or cell phone
There is a 4.59% probability that a randomly selected American has neither a home phone
nor a cell phone.
c. A cell phone but no home phone
i. Cell phone – (home and cell phone)
.83-.58= .25 = 25%
ii. Or is it cell phone * (no home phone)
.83* (1-.73) = .2241 = 22.41%
