Language of evaluation

 

1) (6 pts) The following points pertain to the book and lecture on evaluation. Which statements are true? MTF

(A)   Classifying data as irrelevant to a model means that the data could not possibly refute the model no matter how they turned out.

(B)   In considering guilt versus innocence of a suspect in a trial, the law specifies we adopt the view that the suspect be considered innocent until proven guilty. The model of innocence is an example of a null model.

(C)   In science and in many aspects of society, the scientific method dictates that we adopt strict thresholds for acceptance of a model. A model slightly below the acceptance threshold is considered just as refuted as one way below the threshold.

(D)   If data are consistent with a model, then they also support it.

(E)   A null model is part of every properly designed study. If the study does not have a null model, then it is not properly designed.

 

Correlations

2. (6 pts) Which of the following statements describe a (non-zero) correlation? Do not choose any option that describes a zero correlation or for which a correlation is undefined. If insufficient information is given to determine whether a correlation exists, treat it as if there is no correlation. MTF

(A) There are more parking spaces open at UT in the evening than in the morning.

(B) Teenagers are typically more trouble for parents than are younger children.

(C) Both sexes of lab mice have similar weights

(D) More people buy products endorsed by Michael Jordan than products endorsed by O.J. Simpson

(E) People living near a high voltage power line get leukemia more often than those living far from a high voltage power line

(F) In countries with a large percentage of children malnourished, measles kills about 20% of those infected. In countries with less than 1% of children malnourished, measles is fatal in less than 1% of infections.

(G) White Ford trucks sell better in summer than in winter

(H) Voter turnout is low in UT student elections and in Austin city government elections.

 

3. (4 pts) MTF ÒThird variableÓ models, which we introduced for correlations,

(A)   Explain why correlations are consistent with many causal models

(B)   Explain how X directly causes Y or explain how Y directly causes X

(C)   Explain how X and Y can be correlated when one does not cause the other.

(D)   Explain how X can cause Y (or Y cause X) through intermediate steps (e.g., X-> -> -> Y)

 

4. (4 points) Which of the following constitutes an example of inferring causation from correlation (i.e., in which a correlation leads someone to infer the causal basis of the correlation)? Base your answer only on the information provided. Do not circle answers that merely describe a correlation or which infer correlation from causation. MTF

(A)   The Centers for Disease Control has noted that the incidence of sexually transmitted diseases (STD) decreased after beer taxes were raised in the past. They have thus recommended that further increases in beer taxes is a way to reduce STD rates even further.

(B)   The greatest longevity is seen in adults with modest alcohol consumption (~1 drink/day). This pattern is present even when smoking tobacco is controlled.

(C)   Coca cola ads portray happy people drinking Coca cola. Viewers buy Coca cola to capture that happy feeling.

(D)   In the 1970s it was noted that prison populations contained an unusually high incidence of men with two Y chromosomes (they also had an X), in contrast to their low incidence in the non-prison population. It was suggested at that time that YY was a criminal chromosome type and that all men should be screened for YY to identify them before they committed crimes.

5. (5 pts) Mark all models(s) that are consistent with the information in the following graph. (That is, mark an answer if it cannot be ruled out using the information in the following graph.) Assume you have no data other than what is presented in this graph. MTF


 


(A) Increasing fluoride level causes a decline in tooth decay.

(B) Fluoride level is negatively correlated with tooth decay.

(C) Fluoride level is positively correlated with tooth decay.

(D) Fluoride level is correlated with tooth decay.

(E) Fluoride does not cause changes in tooth decay.

(F) Fluoride causes changes in tooth decay.

(G) After receiving mid-term results, students with poor scores increase their alcohol intake. (this option is not a mistake)

 

 

 

6. (5 pts) Consider the correlation that:

forests in areas with high rainfall have higher levels of insect pests than forests with low rainfall.

Which of the following causal models of this correlation use a Òthird variableÓ to explain the correlation? Do not mark an option if the correlation goes in the wrong direction (e.g., that regions of high rainfall have lower levels of insect pests). MTF

Causal model

Cause invoke a third variable?

rainfall promotes insect growth and reproduction, so that more pests are born when rainfall is higher

Fill in (A) if a 3rd variable is invoked

higher rainfall occurs in regions that also have warmer temperatures, and it is the warmer temperatures that lead to greater pest abundance

Fill in (B) if a 3rd variable is invoked

lower rainfall occurs in regions that also have more parasites of insects, and these parasites reduce levels of pests

Fill in (C) if a 3rd variable is invoked

higher rainfall promotes greater tree growth which in turn breeds larger populations of tree pests.

Fill in (D) if a 3rd variable is invoked

 

7) (4pts) Consider a positive correlation between variable Y and cancer rate. If X (not Y) is the true cause of cancer rate, which cancer rates are expected in cells 1 & 2 of the following table? Assume that no other variables besides X and Y are important. (one answer only)

 

 

Y:

 

 

present

absent

X:

present

low cancer

(1)

absent

(2)

high cancer

 

A) 1 is high, 2 is high

C) 1 is low, 2 is high

B) 1 is high, 2 is low

D) 1 is low, 2 is low

 

8. (4 pts;) New diagnostic methods have allowed us to detect cancers at earlier ages than in the past, and the average age of a person first diagnosed with cancer has decreased (they are now younger at first diagnosis). In addition, people have been living longer from the times of their diagnosis. That is, if A is the average age of a person when cancer is diagnosed and Y is the number of years a person lives beyond diagnosis, there is a (negative) correlation between A and Y (Y increases as A decreases). Note that the average age of death of a diagnosed person is A+Y. Which of the following models are consistent with this correlation? That is, which models are not rejected by this correlation? MTF

(A)   There has been no change in the age of death for people diagnosed with cancer.

(B)   There has been no change in the number of years a person lives beyond their diagnosis of cancer.

(C)   Late diagnosis causes a person to live to a later age than does early diagnosis.

(D)   Early diagnosis causes a person to live to a later age than does late diagnosis.

 

 

Controls

9. (6 pts) Each of rows (A)-(G) describe different treatments that could be applied to humans in testing factors that affect influenza infection rates. The treatments differ in which factors are present (indicated by Ò+Ó) or absent (-). Factor 1 is vitamins in the diet; factor 2 is vaccination; factor 3 is wearing gloves throughout the day; factor 4 is use of public transportation; factor 5 is use of public telephones.

Which two treatments would you want to compare to determine if factor 5 is correlated with differences in influenza infection rates when all other factors are controlled? In evaluating possible answers, pick any comparison that controls for all unwanted factors, and assume that these treatments differ only in the ways stated. Mark exactly two options, or option I if none apply. Each row (each option) describes a different set of conditions, so to know which factors would be applied in a treatment, you look across the row. If multiple combinations satisfy the problem, any correct combination will be accepted. (Two answers or None; options have changed).

 

factor

 

Option

 

1

2

3

4

5

(A)

+

-

-

+

+

(B)

-

+

+

-

-

(C)

-

-

+

-

+

(D)

-

-

-

+

-

(E)

+

+

-

-

-

(F)

+

-

+

-

-

(G)

+

+

-

-

+

(H)

+

-

+

-

-

(I)

No combination satisfies the request

 

10. (5 pts) An epidemiologist does a survey to evaluate whether a personÕs tendency to acquire a criminal habit later in life can be predicted from characteristics of their environment early in life. Only correlational data are gathered (no experiment). For such a study, which of the following options are true? MTF

A)    Characteristics that are the same for all individuals in the study (e.g., country of residence, if the study was confined to one country) would not be controlled for because they are not variables in the study.

B)    Variables that were not recorded about each individual in the study could not knowingly be controlled for.

C)    Variables that were recorded about each individual in the study could be analyzed so that they were controlled for.

D)    Suppose the goal was to assess whether the childhood neighborhood was a factor in development of a criminal habit. If four different neighborhoods were chosen for the study, then the random choice of which individuals within each of those neighborhoods to include in the study would ensure that all variables besides neighborhood would be controlled (on average).

 

 

Experiments

11. (4pts) The test of FC shown in the video was an experiment because (one only)

A) it included controls

C) it manipulated the normal FC environment

E) all of the above (A-D)

B) it was blind

D) it included several levels of replication

 

 

12. (5 pts) An advertiser does a study to compare the effect that the endorsing personality has on sales of a product. The goal is to compare endorsement of a product by a popular person has versus endorsement by someone who is unknown. The ad is video-taped using a famous person, then taped again by someone who is not famous. The same script is used for both people, and the same setting (backdrop, lighting, and camera angle) of the ad is used for both tapings. One ad is shown in Austin, the other in San Antonio, both on 30 January, 2006, and sales of the product are compared before and after the ad is released to determine which type of personality works best (famous vs. unknown). What factors are explicitly controlled for (matched) in the design of this study? Do not infer more than is given. MTF

(A) Gender of the personality

(D) Script

(B) Age of the personality

(E) City of response

(C) Setting in which the ad is taped

(F) timing of ad showing

 

13. (3 pts) In conducting an experiment that assigns subjects to either a control group or a treatment group, which feature of ideal data is most central to ensuring that this assignment controls for as many unwanted (or hidden) factors as possible? One answer.

(A) explicit protocol

(D) randomization

(B) replication

(E) blind

(C) standards

(F) manipulation

 

14 (6 pts). Which of the following studies describe experiments, regardless of whether the experiment was designed well or poorly and regardless of ethics. In each problem, the goal is given. The question is whether the option describes an experiment with respect to the goal. MTF

(A)   To test whether smoking causes lung cancer, you interview people about their smoking habits. You then identify two groups of individuals to observe for a period of two years: smokers and non-smokers. Yet during your observation period, some of the smokers have quit because they do not like the habit. At the end, your study has three groups: smokers who did not quit, smokers who quit, and those who never smoked. You then look for an association between lung cancer and level of smoking.

(B)   You change the spark plugs in your car to see if the engine runs more smoothly.

(C)   You quit smoking to see if food tastes better.

(D)   You quit smoking for health concerns and notice that food begins to taste better.

(E)   A study is done to determine whether alcohol increases the incidence of birth defects. The study design involves comparing babies of women who voluntarily drank during pregnancy with those who did not.

(F)   A researcher merely records the diets of a large number of pregnant women. He finds that women who took twice the recommended daily dose of vitamin A have a 1 in 57 chance of a child with birth defects.

15. (6pts) Prisoners of Silence video. It will be shown later in the semester that Douglas Bicklen of Syracuse University claimed that the experimental tests of FC were not valid because they used an intimidating test environment and asked for tasks that autistic children were not good at (e.g., word-finding). Which of the following statements about the experimental design are valid in refuting Bicklen's specific objection? MTF

(A)   Explicit protocol: specifying a protocol in advance of the test (especially if the child was allowed to practice) mostly eliminates any concern about an intimidating test environment.

(B)   Replication: by performing the test with multiple autistic children and multiple facilitators, it can be concluded that the testing environment was not intimidating.

(C)   Controls: the correct response was typed when child and facilitator were shown the same picture, demonstrating that the FC setting was operating as normal; this control result overcomes the objection Bicklen raised.

(D)  Blind: the fact that the experiments were done blindly meant that the child was unaware of the test and so could not be intimidated.

 

 

16. (6 pts) On-line survey and/or video of horoscope experiment. Which of the following options apply to the on-line survey given to the class and/or the videoÕs horoscope experiment, as shown or discussed in lecture? MTF

A)    Randomization. In light of the goal, the study would have been improved if the distribution of who got which survey had been randomized, both in our class and in the video.

B)    Blind was essential to the goal of the studies. If everyone had known the true purpose of the survey (and had known the design), it is likely that our class and the one shown would not have rated the personality description so highly.

C)    Replication. Similar outcomes have been obtained in previous years of Bio301D. Furthermore, a similar outcome of positive responses was shown in the video.

D)    Controls. It is ambiguous whether controls were present in the video.

 

 

 

17. (9 pts). Which of the following options were true of the Òchemicals are evilÓ lecture? MTF

A)      When tested using the standard models of (rodent) carcinogenicity, organic foods prove significantly safer than non-organic foods

B)      The FDA approval process for herbal remedies is approximately 1/10 as expensive as for drugs.

C)      If not sprayed with chemicals, most food plants are free of compounds that test positive as carcinogens in rodent models.

D)      Somewhat more than half of natural compounds tested in the rodent model are classified as carcinogens.

E)      Somewhat more than half of synthetic compounds tested in the rodent model are classified as carcinogens.

F)       Coffee contains many compounds that test positive in rodent carcinogenicity tests

G)      A slight increase in cancer rates has been observed in heavy coffee drinkers.

H)      The high temperatures in cooking can turn some safe compounds into harmful ones.

I)        The FDA drug approval process approved a drug that, upon marketing, was found to kill some people because of drug interactions.

 

 

 

18. (9 pts). Which of the following options were true of the Òinfectious diseaseÓ lecture? MTF

A)    Since the 1800s (or since 1900), most of the decline in infectious diseases can be attributed to medical interventions.

B)    A graph of the average expenditure (per person) of health care versus the healthy lifespan of individuals showed that the U.S. expenditures were approximately twice that of most other countries shown, but that the U.S. healthy lifespan was less than that of most other countries.

C)    Life expectancies in the U.S. increased throughout the 1900s. From the year 1900 into the 1990s, the most dramatic improvement in lifespan in the U.S. was from reduced infant mortality.

D)    In 1900, the 2 leading causes of death in the U.S. were infectious diseases, and several other infectious diseases were in the top 10.

E)    Toward the end of the 20th century (1980, 1998), neither of the top two leading causes of death in the U.S. were infectious diseases.

F)     By various estimates, hospital-acquired infections kill enough people to be one of the top 10 leading causes of death in the U.S.

G)    The recent trend in the U.S. (in the late 1900s) suggests a decline in hospital-acquired infections.

H)    Although antibiotic-resistant bacteria continue to be a problem in the U.S., there has been a recent surge in new drugs (to new targets in the bacteria) that suggests we will overcome drug resistance.

I) The 1918 flu was different from other flu in that it disproportionately killed young adults (e.g., ages 20-30).

 

19. (5pts) Key code, name, and box number. Fill in (A B) in scantron field 19 to indicate your key for this version of the exam.

Be sure your name and box number are correctly bubbled in on the scantron.

Your name is required on this exam form and the scantron form to receive credit for this test.