1.Writing paper 1 instruction:
Do a full six-step evaluation of the causal study labeled as “writing 1 martial”, including a
diagram of the design of the study and a diagram of the data:
Part of the data (in italics and noted within the PDF of the media report of the study) had to
be added from unpublished information from the clinical trial itself because the otherwise
very detailed report by the journalist writing about it in the New York Times left it out. The
study is unpublished because the results were just announced on November 16th.
Publication is expected in 2020. So, this is real news!
After you write your six-step evaluation, add a 7th "step" to your report to offer a brief
reflection (no more than a paragraph) indicating what you could not have done in your
evaluation had the italicized information I added to the newspaper report not been included.
Grading Breakdown:
Your report will be graded out of 20 points: 2 points each for steps 1-6 of the six-step
evaluation. Your 7th "step" reflection is worth 2 points. And there are 2 points (1 each) for
grammar/composition and argument coherence. Two points for your diagram of the study
design. Two points for your diagram of the data. There is no need for a bibliography since we
know what study you are discussing.
Your report should be 3-5 pages long (double spaced), including diagrams. So, that means 2-3
pages for your six-step evaluation plus a page with the two diagrams. Some people write a bit
more per step than others, so that's why there is a range of lengths from 3-5 pages.
Upload your report as a single PDF including the diagrams as part of your report PDF to
Canvas.
2. Extra credit from writing paper:
write a six step evaluation of the heart attack data in the story used in Writing 1, using
the exact margin of error calculations (appendix, p. 185 ATTACHED).
Your evaluation should, like Writing 1, include a full six step evaluation and also two
diagrams: one for the study design and one for the data. You do not need to do the 7th step
reflection as asked for in the main Writing 1 assignment.
Include your evaluation and both diagrams in a single PDF
Surgery for Blocked Arteries Is Often Unwarranted, Researchers Find
By Gina Kolata Nov. 16, 2019, The New York Times
The findings of a large federal study on bypass surgeries and stents call into question the medical
care provided to tens of thousands of heart disease patients with blocked coronary arteries,
scientists reported at the annual meeting of the American Heart Association on Saturday.
The new study found that patients who received drug therapy alone did not experience more
heart attacks or die more often than those who also received bypass surgery or stents, tiny wire
cages used to open narrowed arteries.
That finding held true for patients with several severely blocked coronary arteries. Stenting and
bypass procedures, however, did help some patients with intractable chest pain, called angina.
“You would think that if you fix the blockage the patient will feel better or do better,” said Dr.
Alice Jacobs, director of Cath Lab and Interventional Cardiology at Boston University. The
study, she added, “certainly will challenge our clinical thinking.”
This is far from the first study to suggest that stents and bypass are overused. But previous
results have not deterred doctors, who have called earlier research on the subject inconclusive
and the design of the trials flawed.
Previous studies did not adequately control for risk factors, like LDL cholesterol, that might have
affected outcomes, said Dr. Elliott Antman, a senior physician at Brigham and Women’s Hospital
in Boston. Nor did those trials include today’s improved stents, which secrete drugs intended to
prevent opened arteries from closing again.
With its size and rigorous design, the new study, called Ischemia, was intended to settle questions
about the benefits of stents and bypass.
“This is an extraordinarily important trial,” said Dr. Glenn Levine, director of cardiac care at
Baylor College of Medicine in Houston.
The results will be incorporated into treatment guidelines, added Dr. Levine, who sits on the
guidelines committee of the American Heart Association.
The participants in Ischemia were not experiencing a heart attack, like Senator Bernie Sanders,
nor did they have blockages of the left main coronary artery, two situations in which opening
arteries with stents can be lifesaving. Instead, the patients had narrowed arteries that were
discovered with exercise stress tests.
With 5,179 participants followed for a median of three and a half years, Ischemia is the largest
trial to address the effect of opening blocked arteries in nonemergency situations and the first to
include today’s powerful drug regimens, which doctors refer to as medical therapy.
All the patients had moderate to severe blockages in coronary arteries. Most had some history of
chest pain, although one in three had no chest pain in the month before enrollment in the study.
One in five experienced chest pain at least once a week.
All participants were regularly counseled to adhere to medical therapy. Depending on the
patient’s condition, the therapy variously included high doses of statins and other cholesterollowering drugs, blood pressure medications, aspirin and, for those with heart damage, a drug to
slow the heart rate. Those who got stents also took powerful anti-clotting drugs for six months to
a year.
Patients were randomly assigned to have medical therapy alone or an intervention and medical
therapy. Of those in the intervention group, three-quarters received stents; the others received
bypass surgery.
The number of deaths among those who had the invasive procedure (stents or bypass), n=388,
was 145, compared to 144 among the patients who received medication alone (the conservative
procedure), n=389. The number of patients who had heart attacks was 276 in the stent and
bypass group, compared with 314 in the medication group, an insignificant difference. [Italicized
information added from the unpublished clinical trials report.]
An angiogram of the arteries of a
patient who was part of the trial, prior
to receiving a stent. NYU Langone
Health
The same patient, after receiving a
stent. NYU Langone Health
Dr. Judith Hochman, senior associate dean of clinical sciences at N.Y.U. Langone Health and
chair of the study, had expected that those with the most severe chest pain and blockages would
benefit from stents or bypass.
But “there was no suggestion that any subgroup benefited,” she said.
Ischemia’s results are consistent with current understanding of heart disease. Researchers have
learned that a patient with a narrowed artery may have plaques not just in a single blocked area,
but throughout the coronary arteries.
There is no way to predict which of those plaques will break open and cause a heart attack.
Stents and bypass treat only areas that are obviously narrowed, but medical therapy treats the
entire arterial system.
Yet when a cardiologist sees a blockage, the temptation for doctor and patient alike is to get rid
of it quickly, said Dr. David Maron, director of preventive cardiology at Stanford University, the
study’s other co-chair.
When an exercise stress test indicates a narrowing, most doctors send patients to a cardiac
catheterization lab to look for blockages, Dr. Maron said. If there is a blockage, the usual practice
is to open it with a stent.
If stenting is not feasible — because of the configuration of the patient’s arteries, for example —
bypass surgery is usually the next step.
Patients with abnormal stress tests should talk to their doctors about the options, Dr. Maron said.
If a patient has chest pain despite taking recommended medications, a stent or bypass might help
improve quality of life.
Still, he said, patients have time to make considered decisions.
“You don’t have to rush to the cath lab because, OMG, you will have a heart attack soon or drop
dead,” Dr. Maron said. “If you have had no angina in the last month, there is no benefit to an
invasive strategy.”
Stenting costs an average of $25,000 per patient; bypass surgery costs an average of $45,000 in
the United States. The nation could save more than $775 million a year by not giving stents to
the 31,000 patients who get the devices even though they have no chest pain, Dr. Hochman said.
But the conventional wisdom among cardiologists is that the sort of medical therapy that patients
got in Ischemia is just not feasible in the real world, said Dr. William E. Boden, scientific
director of the clinical trials network at VA Boston Healthcare System, who was a member of the
study’s leadership committee.
Doctors often say that making sure patients adhere to the therapy is “too demanding, and we
don’t have time for it,” he said.
But getting a stent does not obviate the need for medical therapy, Dr. Boden noted. Since patients
with stents need an additional anti-clotting drug, they actually wind up taking more medication
than patients who are treated with drugs alone.
About a third of stent patients develop chest pain again within 30 days to six months and end up
with receiving another stent, Dr. Boden added.
“We have to finally get past the whining about how hard optimal medical therapy is and begin in
earnest to educate our patients as to what works and is effective and what isn’t,” Dr. Boden said.
1.Writing paper 1 instruction:
Do a full six-step evaluation of the causal study labeled as “writing 1 martial”, including a
diagram of the design of the study and a diagram of the data:
Part of the data (in italics and noted within the PDF of the media report of the study) had to
be added from unpublished information from the clinical trial itself because the otherwise
very detailed report by the journalist writing about it in the New York Times left it out. The
study is unpublished because the results were just announced on November 16th.
Publication is expected in 2020. So, this is real news!
After you write your six-step evaluation, add a 7th "step" to your report to offer a brief
reflection (no more than a paragraph) indicating what you could not have done in your
evaluation had the italicized information I added to the newspaper report not been included.
Grading Breakdown:
Your report will be graded out of 20 points: 2 points each for steps 1-6 of the six-step
evaluation. Your 7th "step" reflection is worth 2 points. And there are 2 points (1 each) for
grammar/composition and argument coherence. Two points for your diagram of the study
design. Two points for your diagram of the data. There is no need for a bibliography since we
know what study you are discussing.
Your report should be 3-5 pages long (double spaced), including diagrams. So, that means 2-3
pages for your six-step evaluation plus a page with the two diagrams. Some people write a bit
more per step than others, so that's why there is a range of lengths from 3-5 pages.
Upload your report as a single PDF including the diagrams as part of your report PDF to
Canvas.
2. Extra credit from writing paper:
write a six step evaluation of the heart attack data in the story used in Writing 1, using
the exact margin of error calculations (appendix, p. 185 ATTACHED).
Your evaluation should, like Writing 1, include a full six step evaluation and also two
diagrams: one for the study design and one for the data. You do not need to do the 7th step
reflection as asked for in the main Writing 1 assignment.
Include your evaluation and both diagrams in a single PDF
was the fact that 61% of American youth (boys and girls, all age groups combined) reported
not bullying any other schoolmates over the previous school term, while 85% of Swedish
youth reported not bullying.
Step 1. The Real-World Population. The population sampled consisted of school-attending
children, both boys and girls, aged 11, 13, and 15 years, who were capable of completing the
entire HBSC questionnaire translated into their primary language and whose school classes
(or class equivalents) were chosen randomly to receive the HBSC questionnaire. There should
be little difference between the population actually sampled and the population of all school-
attending children (both sexes) at these ages in the participating countries.
Step 2. The Sample Data. Among the young people surveyed, 61% of American youth
reported not bullying any schoolmates during the previous school term, while 85% of
Swedish youth reported not bullying. A total of 5168 American young people completed the
survey; 3802 Swedish youths completed it.
Step 3. The Statistical Model. The data can be understood as treating two variables. The
first variable is nationality of the youths sampled with values American and Swedish. The sec-
ond variable is self-reported bullying of other schoolmates at any time during the previous
school term, with values None and one or more times. The data suggest a weak positive cor-
relation between being a Swedish youth and self-reporting no instances of bullying other
schoolmates during the previous school term.
Step 4. Random Sampling. The HBSC employed a cluster sampling procedure that sam-
pled from all students expected to be in the target age groups in school classes or class equiv-
alents. This method, though administratively practical for a study of this scope, is not as precise
as simple random sampling for equal sample sizes. However, the HBSC study designers used
elaborate statistical techniques to generate larger sample sizes in order to ensure 95% confi-
dence levels for each participating country or region. They also chose school classes or class
equivalents using simple random sampling. Schools of all types in each participating nation or
region were included in the population from which classes or class equivalents were drawn.
Steps were taken to ensure accurate and consistent translations of the questionnaire into each
home language. With these procedures, the sampling methods employed approach the preci-
sion of a simple random sample.
Step 5. Evaluating the Hypothesis. The rule-of-thumb margins of error for the sample
of 5168 American youths and 3802 Swedish youths are both about 2%. (This value includes
correcting for the less precise method of cluster sampling.) So the observed difference is sta-
tistically significant. The margins of error are both 0.02, so their sum is 0.04. The difference
in observed frequencies is.85 - .61 = 24. So the difference in sample frequencies is statisti-
cally significant and we have evidence of a correlation between being a Swedish school-
attending youth and not bullying, as compared to being a school-attending American youth.
Its minimal estimated strength is only 2, however [(-85 – 02) – (.61 +.02)—a relatively
weak correlation.
Step 6. Summary. Given the fit between the design of the HBSC and a random sampling
model (with the appropriate corrections in place for the cluster sampling technique), and
given the difference between the two observed sample frequencies, we can be confident that
the population sampled exhibits a weak but statistically significant positive correlation
between being a Swedish school-attending youth and not bullying schoolmates, as compared
to being a school-attending American youth. The same judgment applies to the population of
interest.
6.6 A PROGRAM FOR EVALUATING
STATISTICAL HYPOTHESES
We now set out our program for evaluating statistical hypotheses and apply it to some of
the other data from the HBSC.
The Program
Step 1. The Real-World Population. Identify the real-world population actually sampled
in carrying out the study. Note any important differences between the population sampled
and the population of interest.
Step 2. The Sample Data. Identify the real-world sample and the particular data from that
sample whose relevance for hypotheses about the population you wish to evaluate. Be sure to
include sample sizes when available.
Step 3. The Statistical Model. Identify the relevant variables and the values of these vari-
ables. Then identify the statistical model of the population that is appropriate for evaluation
in the light of the data already identified. If the data have the form of a correlation, give a
clear statement of the statistical hypothesis asserting the existence of the corresponding cor-
relation in the real-world population.
Step 4. Random Sampling. How well does a random sampling model represent the actu-
al process by which the sample was selected from the population? Possible answers: (a) very
well, (b) moderately well, (e) somewhat well, or (d) not very well. Explain the factors relevant
to your answer.
Step 5. Evaluating the Hypothesis. Assuming a random sampling model is applicable, what
can you reasonably conclude about the real-world population? If a correlation is possible, is there
good evidence for the hypothesis stated in Step 3? Estimate the strength of the correlation.
Step 6. Summary. In the light of your answers in all previous steps, give a summary state-
ment of how well the statistical data support your evaluation in Step 5. Possible answers: (a)
very well, (b) moderately well, (e) somewhat well, or (d) not very well. Note the major fac-
tors supporting your answer.
While you work through the program, it is useful simultaneously to diagram the
study following the examples of Figures 6.1, 6-2, 6.3, or 6.4. This might even be done
on a separate sheet of paper. During Step 1, draw a box representing the population sam-
pled. If relevant, surround this box with a larger box representing the population of inter-
est. Label your boxes. During Step 2, add a smaller box below the first one to represent
the sample. Label this box with relevant parts of the data such as sample sizes and observed
frequencies. For Step 3, add the details to your picture of the population necessary to rep-
resent the appropriate statistical hypothesis. At Step 5, draw in the appropriate intervals
representing estimates of population ratios. Where appropriate, indicate statistically sig-
nificant differences. At the end of your analysis, you should have a complete diagram of
the study as it applies to the specific data you are considering.
Nationality and Bullying
Now let us use this program to evaluate other statistical hypotheses suggested by the HBSC.
In particular, let us look at some data on nationality and reported bullying. Among that data
APPENDIX
FORMULA FOR MARGIN OF ERROR
For those who can appreciate the mathematics, there is a fairly simple formula for cal-
culating an approximate value for the margin of error, ME, as a function of both the sam-
ple size, n, and the observed sample frequency, f. The formula, which gives the value of
one standard deviation, is
1 SD = [(f) (1 - f)/n]"2
C2006 Coupe Al Rigle Reserved. Met die weer in parte deg, the profonde desde la
de les despeses destay affect the villain experience Cerage Learning reserves the right way me selfie seg
EXERCISES
185
To obtain our standard margin of error, which produces a 95% confidence interval,
simply multiply the result by 2. For a 99% confidence interval, multiply by 3. An hour
spent with a hand calculator will give you a good idea of how the margin of error varies
when f has values near 1 or zero. However, because this formula is too complex to com-
pute in our heads, we shall not use it in our evaluations.
FORMULA FOR STATISTICALLY
SIGNIFICANT DIFFERENCES
There is also a relatively simple formula for determining whether a difference in
observed sample frequencies is or is not statistically significant. Here we designate the two
observed sample frequencies as f, and f, respectively. The corresponding sample sizes we
designate as n, and n. The formula, which yields the value of one standard deviation in
the difference between the two sample frequencies, is
1 SD = [f, (1 - f)/n, + f (1 - fx/n,]'2
Differences in sample frequencies greater than twice this amount are statistically sig-
nificant at the .05 level. Differences in sample frequencies greater than three times this
amount are statistically significant at the .01 level.
FORMULA FOR CORRECTLY ADDING
MARGINS OF ERROR
There is also a simple formula for converting the sum of individual margins of error at
the 95% confidence level to the margin of error for the difference while keeping the
confidence level at .95. The correct 95% margin of error for the difference is the square
root of the sum of the squares of the two individual margins of error. That is,
ME (difference) = [ME, + ME, 1/2
In the examples of Figures 6.3 and 6.4, the 95% margin of error for the difference works
out to be .11, not .15.
EXERCISES
Analyze these reports following the six-point program for evaluating statistical hypotheses developed
in the text. Number and label your steps. Be as clear and concise as you can, keeping in mind that
you must say enough to demonstrate that you do know what you are talking about. A simple yes
or no is never a sufficient answer. Some of these reports are taken directly from recent magazine or
newspaper articles and are presented here with only minor editing.
EXERCISE 6.1

Purchase answer to see full
attachment