Exam 2, Biology 301D, 17 October 2001 Printed
name ________________
(2pts) Social
security number and name. Bubble in the scantron with your name and your
social security number (your SSN goes(in the first 9 bubbles of the scantron ID
field). (For good measure, put your name on this hard copy too, in case your
key code does not carry over.)
(2pts) Key
Code: One of the exam questions will indicate a key code. Make sure you fill those in as indicated on
the question indicated.
When
finished, turn in both the Scantron
and hard copy. You may write on this hard copy, but your grade will be
determined by the Scantron form.
None, one, all, or
any combination of individual answers may apply to a question unless stated otherwise.
1-3. Which of the following
statements are currently accepted (from lecture and the book). (None, one or
many)
1. (4 points).
Which are true? (easy)
(A) High doses are used as models of low doses when
calculating the cancer risk from radiation exposure
(B) The four abstract models of the relationship
between cancer rate and radiation exposure (i.e., the graphical models: linear, threshold, accelerating,
decelerating) all exhibit the general property that the highest cancer
rates occur at the highest exposures.
(C) The use of a single measure of radiation
exposure for different types of radiation (rem or rad) is a false model, at
least because the different types of radiation have different biological
effects.
(D) The
physical models underlying our understanding of the cancer risk caused by
radiation include (i) different types of cancer, (ii) different types of
radiation, and (iii) different doses.
2. (4 points).
Which are true? (medium difficulty)
(A) The type of cancer most commonly studied in
conjunction with radiation has been solid-tumor cancers (e.g., lung, colon)
because of its relatively short time-lag between radiation exposure and
appearance of the cancer (5 years).
(B) Mice have been used as a model of humans to test
which of the abstract models (linear, threshold, ...) best fits the
relationship between cancer and radiation because this relationship appears to
be insensitive to differences between mice and humans.
(C) Living in Denver
increases your background level of radiation from at least two sources
(cosmic, rocks) compared to living in Austin.
(D) We do not try to
estimate elevated cancer rates from people given medical X-rays on a yearly
basis because the effects of this excess radiation are too small to measure.
3. (5 points).
Which are true? (hardest)
(A) Smoking increases radiation levels to the
lungs by 80X-100X.
(B) Recent work suggests
that low levels of radiation may actually reduce cancer rate (radiation
hormesis).
(C) A doubling of exposure
to radiation doubles cancer rate.
(D) The radioactive “mom”
advertisement warned about the dangers of exposure to high levels of radiation
in some waste sites.
(E) If a million people
were exposed to a single dose of radiation equivalent to 10 times the US
average background level, the yearly number of excess cancer deaths from that
exposure would be at least several hundred, starting a few years after the exposure.
In the following
questions, “fingerprints” refers to the prints made by your fingers, not to DNA
typing.
4. (4 pts) Which of the
following are true? (mostly facts)
(A) Since the death
penalty was reinstated in 1976, slightly more than 1% of people sentenced to
die have been released because they were later shown to be innocent.
(B) The use of fingerprint
evidence in US courts spanned nearly 90 years in the 1900s. Into the 1990s, the main professional
organization of fingerprint experts in the US resisted the use of
objective criteria for deciding whether two fingerprints matched.
(C) The use of hair matching (by methods that do not involve DNA) has
been bolstered recently by the creation of a database that encodes hair
characteristics from over 150,000 different hair samples
(D) The use of dog olfaction (sniffing) to identify items of clothing
belonging to specific individuals has no meaningful basis (is nonsense).
5. (5
pts) Which of the following are
true? (facts connected to course
themes)
(A) A proficiency test is
a test involving the use of standards.
It establishes a human and technical error rate.
(B) When tests of
fingerprint experts were finally conducted in the 1990s, false positive rates
by fingerprint experts were found to be 1% or less.
(C) The routine use of DNA tests before trial
in the last few years of criminal investigation has shown that in approximately
10% of cases, the prosecution’s “prime” suspect was the wrong person.
(D) Of the 4 types of
errors (RPA, sampling, human & technical, bias), the types most common in
leading to wrongful convictions are bias and human & technical.
6. (4 pts)
Which of the following are true?
(more facts connected to course themes)
(A) In 62 cases of
convictions that were later overturned by DNA evidence (you may accept that 62
is correct), the most common factor leading to wrongful conviction was mistaken
identity.
(B) Unintentional bias is
only a minor problem in witness recognition of a suspect, because even though
witnesses can make mistakes, those witnesses who identified the wrong person
have done so in their first attempts at identification and were not influenced
by subtle factors introduced after the investigation began.
(C) False confessions have
been a common factor in wrongful convictions (more than 5% of the time)
(D) The use of blind
samples in lab tests is warranted chiefly to avoid fraud and is not a contributor
to unintentional bias, because honest technicians are not influenced by knowing
sample identity.
Data: Error
(7-10). (3 pts each) For each of
the following descriptions, indicate the types of error present (the underlined phrase identifies
the error). Mark a type of error only if it is definitely present. Do not
assume any more than what is explicitly mentioned in the problem. One answer only for each question,
though an option may be used to answer more than one question.
|
|
(A)
Rounding, Precision and accuracy |
|
Types
of Error: |
(B)
Sampling |
|
|
(C)
Human and technical |
|
|
(D)
Bias |
|
|
(E)
No error is indicated |
7. (3 pts) Time magazine once
reported on an amusing issue with a toilet paper company. Company ads boasted of at least 500 sheets
per roll, but someone had determined that the rolls were a few sheets
short. This discovery led to meetings
of company officials, with lots of hand-wringing. Then someone else went and counted the number of sheets and found
that 500 was correct. The person who
had carried out the first “count” had measured the length of a few sheets to
determine the length per sheet. He then
measured the length of the entire roll and divided by his estimated length per
sheet to estimate the number of sheets in a roll. His number came up short because his length per sheet was not
exact. What kind of error is
represented by the fact that his measure of the length of a single sheet
was not good enough to calculate the exact number of sheets from the length of
the roll? (one only)
(A) (B) (C)
(D) (E)
8. (3 pts) John and Paul race on foot for 100 meters to
determine who is fastest. They hire an
atheletic service to photograph the race and time each of them. They run
together, side-by-side, and John wins the race, arriving at the finish line 2
meters ahead and 0.3 seconds ahead of Paul.
What kind of data error is indicated by John’s time being
shorter (John’s running faster) than Paul’s? (one answer only)
(A) (B) (C) (D) (E)
9. (3 pts)? I want to know what fraction of students
enrolled in class actually show up for lecture, so that the class enrollment
limit can be set as high as possible but without overfilling the classroom
during lecture. (We will assume that
exams are given in a different room). A
count of students in class one day reveals 103 attending. I contemplate using that number but then
deicide that the number may not be subject to error. So I count the number of students attending during the next two
lectures and observe 82 and 150. Although my counts on each day may differ
slightly from the actual numbers, these differences between days are real. What
kind of error is revealed by the variation between days in the number of
students in class as a model of how many students attend lecture? (one answer only)
(A) (B) (C)
(D) (E)
10.
(3 pts) The supermarket at which you shop sometimes rings up a different price
at the checkout stand than the price listed in the isle for that item. You decide to do a brief study of this
pattern. You detect 6 instances of
price discrepancies, and 5 of them involve a higher price at checkout. Feeling incensed about this, you do even
more research and find 15 more cases of price differences, and 13 of those are
higher at the checkout than in the isle.
What type of error is indicated by the fact that, when there is a
price discrepancy, the price at checkout is usually the higher one ?
(one answer only)
(A)
(B) (C) (D) (E)
(11-17).
For each of the following statements, mark the appropriate letters that describe
the data design features present. Base your answer only on the information
given. That is, mark a data feature only if it is explicitly present in the
problem description. One or many bubbles (unless indicated).
|
(A) explicit protocol |
(C) standards |
(E) blind |
|
(B) replicates |
(D) random |
(F) none |
11. (2 pts). Which features are important in reducing
bias? (A) (B) (C) (D) (E) (F)
12. (2 pts). Which features are important in
reducing/detecting human and technical error?
(A) (B) (C) (D) (E) (F)
13. (2 pts). Which features are important in reducing
sampling error? (A) (B) (C) (D) (E) (F)
14. (3 pts)
A pesticide is developed by a chemical company. To test its effectiveness at
killing crop pests, the company conducts a trial following the procedure of a
trial conducted in the previous year: the pesticide is applied to four plots,
half of which are heavily infested with insect pests. After 2 days, a farmer
evaluates only the plots that were heavily infested at the outset.
(A) (B) (C) (D) (E) (F)
15. (3 pts)
A farmer wants to know if using fertilizer will increase the amount of wheat
harvested. To test this hypothesis, he divides his field into 64 plots. Each of
these plots is a square 200m x 200m. To determine whether or not to use fertilizer
in a plot, a coin is flipped. At the end of summer, the yield of wheat is
measured in each plot. To determine that his machinery is harvesting the actual
amount of wheat per plot, he harvests half of one plot by hand and compares
that yield with the yield from the machine's harvest of the other half
(A) (B) (C) (D) (E) (F)
16. (3 pts)
Scientists often make a distinction between "objective" data which
typically requires little interpretation by the person gathering the data (such
as a student's score on a multiple choice exam), and "subjective"
data which may involve substantial interpretation (such as a student's score on
an essay exam). Which ideal data feature is often required in order for
subjective data to be reliable, but is often not necessary for objective data
to be reliable? Exactly one answer is correct.
(A) (B) (C) (D) (E) (F)
17. (3 pts)
A middle school student decides to determine if Fridays on the 13th day of the
month are unlucky. She solicits records
from emergency-room hospital admissions and police arrest reports for all four
Friday-the-13ths during a two year period.
To establish a base-line level of expected admissions and arrests, she
also obtains records from hospitals and police for the Fridays immediately
before and after each Friday-the-13th.
To guard against errors in her analysis of the data, she has the
analysis repeated by a second person, who also does not know which group of
admissions is for Friday-the-13th and which is not.
(A) (B) (C) (D) (E) (F)
18-20.
Do-it-yourself protocol. You are conducting an external review/test of a DNA lab. Your job is
to send two tubes to the lab, with labels. There are several options for the
content of and label on a tube. You must decide which contents to send and how
to label the tubes so that the features of ideal data requested in the question
are present from the lab's perspective. If a tube has a person's name on it,
the lab can assume that the tube contents belong to the name of the person on
the label. If a tube is labeled with a number, the contents are unknown to the
lab but known to you. Your options for tube contents and tube labels are:
|
option |
tube label |
Contents in the tube are
from |
Blood type
|
Gender |
|
(A)
|
Sam
Edwards |
Sam
Edwards |
AB |
Male |
|
(B)
|
Holly
Berry |
Holly
Berry |
O |
Female |
|
(C)
|
Marty
Ray |
Marty
Ray |
O |
Male |
|
(D)
|
#13 |
Wayne
Crill |
A |
Male |
|
(E)
|
#21 |
Wayne
Crill |
A |
Male |
|
(F)
|
#100 |
June
Scott |
B |
Female |
|
(G) |
#243 |
Linda
Davis |
O |
Female |
|
(H) |
#17 |
Joel
Sachs |
AB |
Male |
|
(I)
No combination of tubes can satisfy the protocol |
||||
In
the following questions, choose two letters among options (A)-(H) to describe
the two tubes that will be sent to the lab. If it is possible to satisfy the protocol, the question will require
exactly two letters and only
two letters -- one for each tube. Thus, the answer for a question might be
(A) & (B), or it might be (D) & (F). If more than one pair of options
are possible correct answers, fill in only one correct pair of options. Thus,
if (A) & (B) is one acceptable answer, and (C) & (D) is another
acceptable answer, fill in either (A)&(B) or (C)&(D), but not both.
If a factor (such as identity, blood type, gender) is not specified in
the protocol, then that factor will be ignored in grading the answer. Alternatively,
if a protocol cannot be satisfied with options for tubes (A)-(H), fill in (I).
18. (2 pts) Choose two tubes to
guarantee replication of blood type but not for individual.
two tubes: (A) (B) (C) (D) (E) (F) (G) (H) (I)
19. (2 pts) The lab has a way
to identify the gender (sex) of a sample. Make the protocol fully blind to the
lab and replicated for gender but not replicated for an individual or blood
type.
two tubes: (A) (B) (C) (D) (E) (F) (G) (H) (I)
20. (2 pts) Make the protocol
fully blind to the lab and not replicated in any way (neither gender, blood
type, nor individual).
two
tubes: (A) (B) (C) (D) (E) (F) (G) (H) (I)
(21-23). Each of items (A) through
(J) describes a feature of Department of Transportation drug tests. The details of the drug testing rules are
presented in the Federal Register 49 CFR PART 40 and described in a 100+ page
pamphlet "Guidelines for Implementing the FHWA anti-drug program."
(A) Each manager whose employees are tested is
required to include 3 known blank samples (lacking drugs) for every 100 unknowns;
managers with more than 2000 drivers must also include some samples known to
contain drugs.
(B) The laboratory initially tests all samples with
a rapid screen; positive samples are retested more comprehensively.
(C) When a sample is sent to the laboratory for
testing, it is labeled with a code, rather than the name of the person being
tested.
(D) Split sample collection is required -- the
partitioning of the original sample into two vials -- so that the second sample
can be retained for retesting.
(E) The rules specify that any retesting (in the
event of a positive sample) is to be done by a different laboratory than did
the original test.
(F) The regulations require that any driver involved
in an accident be tested for drugs and alcohol.
(G) Lab results are sent to the medical review
officer before being returned to the office originating the test.
21. (3 pts) Which feature(s)
represent standards in some part of the drug-testing protocol? (none, one, or
many)
(A) (B) (C) (D) (E) (F) (G)
22. (3 pts) Which feature(s),
if followed, ensure that DOT drug tests of an individual's sample are done
blind? (none, one, or many)
(A) (B) (C) (D) (E) (F) (G)
23. (3 pts) Which feature(s)
provide replication to guard against human and technical error? (none, one, or
many)
(A) (B) (C) (D) (E) (F) (G)
DNA Typing
(24-26) Which of the following violations of ideal data were discussed in class
or the book? An option is correct only
if the example was discussed AND it is a valid example of the type of violation
listed. For example, an option that
listed the failure to replicate a test as a violation of randomization would
not be considered correct. In most
cases, you will need to rely on information from the book and lecture to
provide details needed to answer the question.
24. (5 pts). Which are true? (general questions about DNA typing)
(A) When done properly,
DNA results have a low probability of producing a chance (random) match between
the forensic sample and a person not associated with the crime. This probability is often less than
1/million.
(B) DNA can be extracted
from most tissues and from only trace amounts.
For example, enough DNA can be recovered from the back of a postage
stamp to determine whether a particular suspect licked the stamp.
(C) Lack of
standards: most current DNA labs refuse
to be subjected to external blind proficiency tests.
(D) Human and technical
error: the best guess of current errors
due to such problems as sample mixup is a few percent (e.g., 2%).
25. (5 pts). Which are
true? (questions about individual cases in DNA typing)
(A) Violation of explicit
protocol: the lab in the Castro case
violated its own measurement criteria for declaring a match
(B) Lack of blind: the Castro case revealed one or more
instances in which knowledge of the sample identity likely led to technicians
ignoring inconsistencies in the data.
(C) Lack of blind: letters from the Chicago Police Dept.
requesting DNA testing from the FBI lab provided suggestive information about
the suspect’s guilt.
(D) Lack of
randomization: In the Castro case, the
failure to randomize the order in which standards and Castro’s sample were
scored caused the bar code to be misread in one test.
26. (8 pts) The following points
pertain to the introductory lectures on evaluation. Which statements are true?
(none, one, or many).
(A) Data are considered inconsistent with models because all models are false.
(B) If data are consistent with a model, it means
that they support it.
(C) If data support a model,
they are also consistent with it.
(D) Classifying data as irrelevant
to a model means that the data could not possibly refute it no matter how they
turned out.
(E) Classifying data as irrelevant to a model means that the data were not gathered
according to the Ideal Data template
(F) Data gathered on the incidence of business
failures in Austin would likely be irrelevant to models about the origins of
the AIDS virus.
(G) In considering guilt versus innocence of a
suspect in a crime, the law specifies that we adopt the view that the suspect
is considered innocent until proven guilty (consider this statement true). The
model of innocence is an example of a null
model because it is the model we accept until evidence forces us to reject it.
(H) In science and in many aspects of society, our
willingness to reject a null model increases as evidence accumulates against
that model. For example, in class, our acceptance of the safety of a vaccine
(against the null model that the vaccine was not safe) increased as the sample
of successful trials increased from 10 individuals to millions of individuals.
(I) In a Pepsi versus Coca-Cola taste
test, one would use either null model that Pepsi tastes best or that Coca-Cola
tastes best.
27.(5
pts) Which of the following describe non-zero correlations?
(A) on a per-car basis, red cars are involved in
more accidents than cars of other colors
(B) people who drink approximately 1 oz of alcohol
per day have higher survival rates than non-drinkers
(C) towns that have the most churches sell the least
liquor
(D) mice carry
diseases that can infect humans
(E) The people you encounter on Mondays tend to be
the same as those you encounter on Wednesdays and Fridays, but different from
those you encounter on Tuesdays and Thursdays
(F) The average score on exam 2 is the same as the
average on exam 1
28.
(5 pts) Which points about correlations are true?
(A) Correlations are consistent with many abstract
models
(B) a zero correlation means that there is only one
variable
(C) correlations are either positive, negative, or
zero
(D) advertisements commonly use correlations in
attempting to get people to buy a product
(E) if we know there is a correlation between
variables X and Y (height and weight), it means that we can predict the average
value of Y once we know the value of X
(F) if we know there is a correlation between height
and weight of people, it means that we know a specific person’s weight, we also
know that person’s height.
(G) It is not possible to determine whether a correlation
exists when one of the variables is “country of residence,” because a value
cannot be assigned to this kind of variable.
29. (2 pts). Key code. Fill in (A) and (B) on the scantron field for question 29 to indicate your key code.