CH7

Types of Experimental Designs

1. COMPLETELY RANDOMIZED Experimental Designs 2. BLOCK Experimental Designs 3. MATCHED PAIRS Experimental Designs

Check Your Skills 7.23: An experimental design randomly assigns volunteer patients suffering from chronic pain due to arthritis to take either a new pain medication, aspirin (the traditional nonsteroidal anti-inflammatory treatment), or a placebo - An institutional review board discusses whether the experiment should include a placebo This is... a. An ethical issue b. An issue of lack of replication c. An issue of lack of realism

A

"Matched Pairs"

A term that refers to actual pairs of SIMILAR BUT DIFFERENT INDIVIDUALS

Statistical Significance

An OBSERVED EFFECT SO LARGE that IT WOULD RARELY OCCUR BY CHANCE - Use the Laws of Probability to determine whether a treatment's effect is larger than we would expect to see if only chance assignment to groups were operating

Clinical Trial

An experiment using HUMAN SUBJECTS for HEALTH-RELATED RESEARCH PURPOSES

Treatment

Any specific EXPERIMENTAL CONDITION APPLIED TO the SUBJECTS - If an Experiment has MORE THAN ONE FACTOR, this is a COMBINATION OF SPECIFIC VALYES OF EACH FACTOR

Q's to Ask Ourselves re: Experimental Ethics:

Are all experiments acceptable in the name of science, knowledge, or the public's greater good? - Both medical ethics and international human rights standards say that "the interests of the subject must always prevail over the interests of science and society" - The quoted words are from the 1964 Helsinki Declaration of the World Medical Association, the most respected international standard - The most outrageous examples of unethical experiments are those that ignore the interests of the subjects - Chief among them is the Tuskegee syphilis study Is it morally right to include a placebo in a study design when safe and efficient treatments are already available? - Because "the interests of the subject must always prevail," medical treatments can be tested in clinical trials only when there is reason to hope that they will help the patients who are subjects in the trials - Future benefits aren't enough to justify experiments with human subjects - So why is it ethical to give a control group of patients a placebo? • Well, we know that placebos often work • Once more, placebos have no harmful side effects other than withholding the benefits of a proven treatment • If we knew which treatment was better, we would give it to everyone • When we don't know, it is ethical to try both and compare them Should we stop an experiment if we find that one of the treatments shows early signs of adverse effects? And what if one of the treatments shows early signs of clear superiority? - Again, "the interests of the subject must always prevail" - Disregard of information gained in the course of experimentation or follow-up studies has been the basis of many private and class-action lawsuits against pharmaceutical companies and is fueling the current debate over the prescription of Vioxx, for instance Is it ever acceptable to leave subjects in the dark about the purpose and expected results of an experiment? - This question goes hand in hand with the first question on what experiments are morally acceptable - Subjects should be able to judge for themselves the acceptability of their participation - That's why all federally funded human experiments are now required to obtain the subject's informed consent in writing as well as the formal approval of an ethics committee - But can subjects truly grasp the purpose and consequences of an experiment? • And is complete informed consent achievable without compromising the experiment? • Some experiments, especially in psychology, must disguise their true purpose to obtain genuine responses from the participants Should subjects be rewarded for participation? - This is a tricky question related to informed consent and the freedom to choose participation - Poor and uninsured individuals could easily become primary targets for the riskiest experiments - In fact, it was once common to test new vaccines on prison inmates who gave their consent in return for good-behavior credit - The law now forbids medical experiments in prisons - And, of course, we must also learn to think globally because new treatments can be tested anywhere, including in parts of the world where subjects' rights have not yet fully matured

Check Your Skills 7.14: What electrical changes occur in muscles as they get tired? - Student subjects hold their arms above their shoulders until they drop - Meanwhile, the electrical activity in their arm muscles is measured This is... a. An observational study b. An uncontrolled experiment c. A randomized comparative experiment

B

Check Your Skills 7.22: The National Institutes of Health should be reluctant to use the results of the experiment described in Exercise 7.15 for its guideline on diets for Americans with high blood pressure because... a. The study used a matched pairs design instead of a completely randomized design b. Results from men with high blood pressure may not generalize to the population of all Americans with high blood pressure c. The study did not use a placebo and was not double-blind

B

Blocking is an Important Additional Principle of Statistical Design of Experiments

Blocking allowed the experimenters to include in their analysis an additional variable (smoke exposure in Ex 7.6 and field location in Ex 7.7) that was not an imposed treatment in the experiment - Blocks allow us to draw separate conclusions about each block and to assess whether the response to the treatments (such as the two asthma care options in Example 7.6) may depend on the block variable A wise experimenter will form blocks based on the most important unavoidable sources of variability among the subjects or experimental units Randomization will then average out the effects of the remaining variation and allow an unbiased comparison of the treatments

A Strong OBSERVED Association DOES NOT Imply Causation...

But a STATISTICALLY SIGNIFICANT Association in data from a well-designed experiment DOES imply Causation

Check Your Skills 7.21: To decide which call the mother goats in Exercise 7.20 should hear first, it is best to... a. Always start with the same call (for example, the different kid) for consistency b. Base the decision on how attached each mother goat was to her kid c. Use chance (for example, a coin toss)

C

Ex 7.6: An estimated 9.5% of children in the United States suffer from asthma - Children with persistent asthma should take daily preventive anti-inflammatory medication - Unfortunately, this recommendation is not always followed A study compared the effect of the usual at-home asthma care and a school-based asthma care program (in which school nurses administer treatment) on the asthma symptoms of elementary school children

Environmental tobacco smoke can trigger and worsen asthma symptoms, so the experiment used home exposure to tobacco smoke as a blocking variable Two separate randomizations were done: - One assigned the children exposed to smoke to the two treatments - The other assigned the children not exposed to smoke to the two treatments Figure 7.4 outlines the design of this experiment - Assignment to Block is NOT Random Subjects (530 children with asthma) → Exposed to smoke (285) → Random Assignment → Group 1 or 2 → Group 1 (140) → School-based care → Group 2 (145) → Usual care → Compare symptoms → Not exposed to smoke (245) → Random Assignment → Group 1 or 2 → Group 1 (125) → School-based care → Group 2 (120) → Usual care → Compare symptoms Note that no randomization was involved in making up the blocks - Instead, the groups of subjects differed in some way (home smoke exposure, in this case) that existed before the experiment began

Exercise 7.51: In the 1960s, children entering the Willowbrook State School, an institution for the developmentally disabled, were deliberately infected with hepatitis - The researchers argued that almost all children in the institution quickly became infected anyway - The studies showed for the first time that two strains of hepatitis existed - This finding contributed to the development of effective vaccines Despite these valuable results, the Willowbrook studies are now considered an example of unethical research - Explain why, according to current ethical standards, useful results are not enough to allow a study

Ex: informed consent is lacking

Placebo

In an Experiment, a CONTROL TREATMENT that is FAKE, but otherwise indistinguishable from the treatment in the experimental group - Ex: taking a SUGAR PILL

Experiments have Clear Advantages over Observational Studies

In an Experiment, we can study the effects of the specific treatments in which we are interested - By randomly assigning subjects to treatments, we can avoid confounding Consider the experiment described in Example 7.1: - If the researchers simply compared the medical records of infants treated with immediate probing with those of infants monitored with deferred treatment, it is possible that one group might also be found to have more severe symptoms or to have better access to health care and insurance Moreover, we can control the environment of the subjects to hold constant factors that are not directly of interest to us - In Example 7.2, for instance, the chrysanthemums were grown in greenhouses where factors such as soil, temperature, and humidity were controlled Experiment make it so we can study the combined effects of several factors simultaneously - The interaction of several factors can produce effects that could not be predicted from looking at the effect of each factor alone - Perhaps interrupting the night period inhibits flowering in chrysanthemums, but only if the light used during the interruption contains some red light Experiments are the preferred method for examining the effect of one variable on another - By imposing the specific treatment of interest and controlling other influences, we can pin down cause and effect

Lack of Realism

In which the SUBJECTS, TREATMENTS, or SETTING of an experiment MAY NOT REALISTICALLY DUPLICATE the CONDITIONS WE really WANT TO STUDY - This can limit our ability to apply the conclusions of an experiment to the settings of greatest interest - Ex: Carcinogenicity is often tested in extreme situations unlikely to reflect realistic human use

Experimental Units

Individuals on whom an Experiment is performed - Can be animals, bacteria, chemical compounds, elements, objects, etc. - Refers to NON-human Subjects

Ex 7.12: We know that mice and rats can differentiate human experimenters by smell - So, when it comes to study results, does it matter who the experimenter is? Researchers decided to investigate this question with a randomized experiment, using rodent facial grimacing indicative of pain as an observable behavior

Male and female adult mice that had never before participated in a lab experiment were injected with an irritating product in the ankle - Each mouse was then placed next to an experimenter sitting quietly in an empty room, and the mouse's facial expressions were recorded - Each mouse was tested only once, with a randomly assigned experimenter The data show that facial grimacing was significantly diminished when the experimenter was male - The same was true when the mice were left in the room with only a T-shirt previously worn by either a male or a female experimenter - When the study was replicated with lab rats, the findings were similar The researchers concluded that the sex of an experimenter can affect the behavioral response of rodents in a laboratory setting

COMPLEX Experimental Designs

On the 1 hand, the causal effect of one variable on another might be better isolated from influences of other factors - On the other hand, the data collection, analysis, and interpretation are likely to become much more complex

Subjects

Refers to the Individuals/Units studied in an Experiment when they are HUMANS - Experimental Units that are human beings

Simple Experimental Design

Subjects → Treatment → Measure response In the controlled environment of the laboratory, these may work well - However, field experiments and experiments with human subjects are exposed to more variable conditions and deal with more variable individuals - This type of design often yields worthless results because of confounding with lurking variables

Factors

The EXPLANATORY VARIABLES in an Experiment

Exercise 7.27: Does regular exercise reduce the risk of a heart attack? - Here are two ways to study this question - Explain clearly why the second design will produce more trustworthy data 1. A researcher finds 2000 men older than age 40 who exercise regularly and have not had heart attacks - She matches each with a similar man who does not exercise regularly, and she follows both groups for five years 2. Another researcher finds 4000 men older than age 40 who have not had heart attacks and are willing to participate in a study - She assigns 2000 of the men to a regular program of supervised exercise - The other 2000 continue their usual habits - The researcher follows both groups for five years

The first study is an observational, case-control study The second is an experiment (the exercise plan is assigned randomly)

Ex 7.8: Essential tremor is a neurological movement disorder characterized by involuntary rhythmic movement that typically interferes with the full use of the arms and hands - The most common movement disorder, it differs from Parkinson's disease in that it is typically action related - Bioengineers designed a noninvasive handheld device using active cancellation of tremor (ACT) technology to stabilize tremor-induced motion

To test the efficacy of the ACT device, the researchers recruited a sample of 15 subjects suffering from essential tremor - The subjects performed a feeding task (bringing a full spoon to their mouth) with the device attached to their dominant hand - The device was turned both on and off in random order during the task, without letting the subjects know when the switch occurred - Tremor amplitude was measured (in centimeters) with a precision accelerometer under both conditions This study uses a MATCHED PAIRS Design in which pairs of measurements are obtained for each subject - The order of the on and off conditions was randomized in case tremor amplitude changed during the course of the task - Not informing the subjects when the device was switched on or off meant that they could not anticipate the switch and change how they performed the task In a completely randomized design, the 15 subjects would have been assigned at random to two groups (ACT on and ACT off), so that only one tremor amplitude measurement would have been obtained for each subject - Some patients, however, have more severe conditions characterized by greater tremor amplitudes - The completely randomized design relies on chance to distribute the more severe cases roughly evenly between the two groups The matched pairs design, in contrast, compares each subject's tremor amplitude with the device on and with the device off - This makes it easier to see the effect of the ACT device on tremor amplitude despite differences in tremor severity - Figure 7.6 illustrates how data organization differs when data are collected using a completely randomized design (left) versus a matched pairs design (right)

REPEATED MEASURES Design

When the SAME INDIVIDUALS are used for BOTH TREATMENTS in an Experimental Design - This term also applies to experiments in which EACH SUBJECT is EXPOSED TO MORE THAN TWO TREATMENTS

Exercise 7.45: Does the herb Saint-John's-wort relieve major depression? A study examined this issue. Here are some excerpts from the published report: - Design: Randomized, double-blind, placebo-controlled clinical trial - Participants...were randomly assigned to receive either Saint-John's-wort extract (n = 98) or placebo (n = 102) - The primary outcome measure was the rate of change in the Hamilton Rating Scale for Depression over the treatment period a. Use the information provided to draw a diagram outlining the design of this study b. The report concluded that Saint-John's-wort is no more effective than a placebo at relieving symptoms of depression - Based on the information quoted, would you be inclined to trust this conclusion? - Explain your answer

b. Yes; the results appear trustworthy - The study was a comparative experiment with a placebo control group, subjects were randomly assigned to treatments, double-blinding was used to prevent bias, and the variable was clearly defined

Block

In a Block Design, a GROUP OF INDIVIDUALS that are KNOWN BEFORE THE EXPERIMENT TO BE SIMILAR IN SOME WAY that is EXPECTED TO AFFECT the RESPONSE TO the TREATMENTS - Individuals sharing the same characteristic are pooled, forming this A preexisting characteristic is used to form the basis of these - It represents a feature of the individuals rather than an imposed condition/treatment They are another form of CONTROL - They control for the effects of some outside variables by bringing those variables into the experiment to form themselves

Exercise 7.43: If the random assignment of patients to treatments in Exercise 7.42 did a good job of eliminating bias, possible lurking variables such as smoking history, asthma, and hay fever should be similar in all 4 groups - After recording and comparing many such variables, the investigators said that "all showed no significant difference between groups" Explain to someone who knows no statistics what "no significant difference" means - Does it mean that the presence of all these variables was exactly the same in all four treatment groups?

"No significant difference" does not mean the groups are identical but rather that observed differences are no bigger than we might expect from true random allocation alone

Principles of an Experimental Design

1. CONTROL the effects of Lurking Variables on the response - Most simply by comparing two or more treatments 2. RANDOMIZE - Use impersonal chance to assign subjects to treatments 3. Use ENOUGH SUBJECTS in each group to reduce chance variation in the results

Check Your Skills 7.17: In the experiment described in Exercise 7.15, the 240 subjects are labeled 001 to 240 - Software assigns an SRS of 60 subjects to Diet 1, then an SRS of 60 of the remaining 180 subjects to Diet 2, then an SRS of 60 of the remaining 120 subjects to Diet 3 - The 60 subjects who are left get Diet 4 This is a... a. Completely randomized design b. Block design, with four blocks c. Matched pairs design

A

Check Your Skills 7.18: A study of the relationship between dengue infection and miscarriage enrolled a group of women who had just experienced a miscarriage - The investigators matched each woman who had suffered a miscarriage to three women with viable pregnancies at a similar gestational stage, then asked them if they had had a recent dengue infection This is... a. An observational study with a case-control design b. An experiment with a block design c. An experiment with a matched pairs design

A

Ex 7.7: The soil type and fertility of farmland differ by location - To account for these differences, experiments on plant growth often divide the available planting area into smaller fields, which make up the Blocks

A test of the effect of tillage type (two types) and pesticide application (three application schedules) on soybean yields used small fields as blocks - Each block was further divided into six plots, and the six treatments were randomly assigned to plots separately within each block, as illustrated in Figure 7.5 - Field location, used to create the blocks, was a preexisting feature of the experimental units (the plots)

RANDOMIZED COMPARATIVE Experiment

An experiment that uses both the COMPARISON OF TWO/MORE TREATMENTS & the CHANCE ASSIGNMENT of subjects to treatments They are designed to provide good evidence that differences in the treatments actually cause the differences we see in the response - If we observe statistically significant differences among the groups, we have good evidence that the treatments actually caused these differences - They can produce data that GIVE GOOD EVIDENCE FOR a CAUSE-&-EFFECT RELATIONSHIP between the explanatory and response variables

MATCHED PAIRS Design

An experimental design that COMPARES EXACTLY TWO TREATMENTS, either by: - Using a SERIES OF INDIVIDUALS that are CLOSELY MATCHED TWO BY TWO - Using EACH INDIVIDUAL TWICE They require that the ASSIGNMENT OF the TWO TREATMENTS WITHIN EACH "PAIR" be RANDOMIZED to avoid a systematic bias They are especially useful when substantial differences between individuals exist/are suspected - This experimental design uses a form of blocking to compare just two treatments With this type of design, the experimenter chooses pairs of subjects that are as closely matched as possible: - These can be individuals of the same sex, age, weight, and stature, for example - Or they can be genetically related individuals such as twins or animals born in the same litter The experimenter then uses chance to decide which subject in a pair gets one specific treatment - The other subject in that pair gets the other treatment - That is, the random assignment of subjects to treatments is done within each matched pair, not for all subjects at once - Alternatively, sometimes a Crossover Design is used - Because the order of the treatments can influence the subject's response, we randomize the order for each subject

Apply Your Knowledge 7.13: We know that emotional states can be transferred to others via emotional contagion - But does emotional contagion occur outside of in-person interactions? A research article describes the findings of an experiment conducted by Facebook for internal purposes - The experiment manipulated the amount of emotional content in the posts from friends shown in the news feeds of 689,003 Facebook users - Based on random assignment, these Facebook users received news feeds that were either modified to have reduced positive expression, modified to have reduced negative expressions, or left intact - When positive expressions were reduced, Facebook users posted fewer positive words & more negative words; when negative expressions were reduced, the opposite pattern occurred a. How were participants selected for this experiment? - Here is what the article says: "People who viewed Facebook in English were qualified for selection into the experiment" - No notification was provided for those 689,003 qualified participants, and they were not given any opportunity to opt out of the study - Describe some ethical concerns about such an approach to human experimentation The article states that participation in the experiment "was consistent with Facebook's Data Use Policy, to which all users agree prior to creating an account on Facebook, constituting informed consent for this research" - Explain how agreeing to Facebook's terms and conditions at the time of account creation differs from the ethics requirements for human research referred to as the "Common Rule" c. The effect resulting from the emotional manipulation was very small, on the order of 1 fewer emotional word per 1000 words posted the following week - Does this fact affect the validity of the ethical concerns over this experiment? d. Many scientists voiced their concern in the media over the ethicality of this experiment, and the editor-in-chief of the scientific journal in which this article was published demanded that an editorial expression of concern be added to the article - How is this involvement from the scientific community different from the scientific community's reaction to publications about the Tuskegee syphilis study decades earlier (described in the Discussion on page 185)

Answers will vary a. One difference is the inability to refuse participation b. Not a truly informed consent c No - The experimenters did not know when they started the experiment that the effects would be small d. In this case, the scientific community voiced its concern clearly & immediately

Ex 7.9: Chocolate milk is gaining in popularity among athletes as a sports drink because it contains proteins in addition to the carbohydrates (sugar) and minerals found in commercial sports drinks such as Gatorade Researchers wanted to see if drinking chocolate milk would result in higher endurance among athletes than drinking Gatorade or mineral-enriched water

Athletes vary substantially in their physical abilities - They vary so much that individual variations across athletes might be more substantial than the effect of one drink over another Recognizing this fact, the researchers asked trained male cyclists to cycle to exhaustion in 3 different weeks: 1. Once drinking chocolate milk 2. Once drinking Gatorade 3. Once drinking mineral-enriched water - In a randomized order That is, each athlete was subjected to all 3 treatments, in a REPEATED-MEASURES design

Check Your Skills 7.15: Can changing a person's diet reduce high blood pressure? - Vegetarian diets and low-salt diets are both promising alternatives Men with high blood pressure are assigned at random to four diets: 1) normal diet with unrestricted salt 2) vegetarian diet with unrestricted salt 3) normal diet with restricted salt 4) vegetarian diet with restricted salt This experiment has... a. One factor—the choice of diet b. Two factors—normal/vegetarian diet and unrestricted/restricted salt c. Four factors—the four diets being compared

B

Check Your Skills 7.20: Can goats remember and identify the call of their kid (baby goat)? Researchers selected 9 mother goats who had weaned their kids nearly a year earlier - They measured the intensity of the mother's response when presented with the recorded call of their own kid and also with the recorded call of a different kid - They found that the mothers' responses were significantly more intense for their own kid This is... a. An observational study b. A matched pairs experiment c. A completely randomized experiment

B

Check Your Skills 7.16: The response variable in the experiment described in Exercise 7.15 is... a. The amount of salt in the subject's diet b. Which of the four diets a subject is assigned to c. Change in blood pressure after 8 weeks on the assigned diet

C

Check Your Skills 7.19: A medical experiment compares an antidepressant with a placebo for relief of chronic headaches - A total of 38 patients with chronic headaches are available to serve as subjects: 28 women and 10 men - The investigators randomly select 16 female patients to receive the antidepressant and they also randomly select 5 men to receive the antidepressant - The remaining patients will receive the placebo This experiment has... a. A completely randomized design b. A matched pairs design c. A block design

C

COMPARATIVE Experiment

Experimental design in which DATA FROM VARIOUS UNMANIPULATED SAMPLES/POPULATIONS are COMPARED, but in which variables are not controlled or even necessarily identified In Ex 7.3: - Some patients undergo the gastric-freezing treatment & other, similar patients undergo a sham procedure or no procedure at all - Comparing patients undergoing gastric freezing with untreated patients allows us to control for the effect of natural improvement - However, that alone would not tell us whether a lower pain rate in the Experimental Group (gastric freezing) is due to the benefits of the medical procedure itself or to the Placebo Effect - Only by comparing the gastric-freezing procedure with a sham procedure would we be able to conclude that gastric freezing itself is responsible for any observed improvement in patient condition - This remedies the confounding

DOUBLE-DUMMY Experiment

Experimental plan in which each participant uses the ACTIVE VERSION OF ONE TREATMENT along with a PLACEBO VERSION OF THE OTHER The use of a MATCHING PLACEBO to ensure blinding, when all patients receive TWO FORMULATIONS (drug & placebo) - Active tablet + placebo liquid compared to placebo tablet + active liquid

Experimental Ethics

For an experiment to be ethical (for humans) it needs consent, no harm, anonymity, & a debriefing - For animals it can't be prolonged pain Specifies what kinds of experiments should or should not be done Ethical issues are particularly important in biological experimentation because investigators deal with live forms - Whether the experimental units are humans, animals, or ecosystems, experimentation needs to be thought out not only to be statistically powerful and elegant but also to be responsible It is about all the questions we need to ask before, during, and after experimentation - But, unlike design, ethics doesn't always have clear-cut answers - Therefore, our objective here won't be to give you all the right answers but rather to help you ask all the right questions Human experimentation undoubtedly lends itself to particular ethical scrutiny - All experiments involving humans, even the least intrusive ones, raise ethical issues - Many experiments involving babies and young children have, for instance, helped us understand the nature and importance of attachment between a parent and a child - Yet we might legitimately ask whether the subjects themselves benefited from these experiments and whether they suffered any negative impacts - These same questions become paramount in the area of medical experimentation, where benefits and risks are clear and potentially great

Laws of Probability

Give a MATHEMATICAL DESCRIPTION OF CHANCE BEHAVIOR - The basis for hypothesis testing and confidence interval estimation

Ex 7.13: The carcinogenicity of a product is its propensity to induce cancer when used, ingested, or simply present in the environment - Saccharin has been used as an artificial sweetener for almost a century Because of experimental evidence of carcinogenicity in rats, in 1981 saccharin was listed in the U.S. Report on Carcinogens as reasonably anticipated to be a human carcinogen

However, closer examination of the observed urinary bladder cancers in rats indicate that the biological mechanism of this cancer is specific to the rat urinary system and related to saccharin consumption at concentrations unlikely to represent realistic human consumption Saccharin was officially delisted from the Report on Carcinogens in 2005, with the conclusion that "the factors thought to contribute to tumor induction by sodium saccharin in rats would not be expected to occur in humans"

Comparison Alone Isn't Enough to Produce Strong Conclusions

If the treatments are given to groups that differ markedly when the experiment begins, bias will result In Ex 7.4: - If we select all the patients most likely to improve to make up the gastric-freezing treatment group, we will see better results in the treatment group even if the procedure itself makes no contribution to the improvement Any personal choice would bias our results in the same way that volunteers bias the results of opt-in polls - The solution to the problem of bias is the same for experiments and for survey samples: RANDOMIZATION

CROSSOVER Design

In a Matched Pairs Design, sometimes EACH "PAIR" CONSISTS OF just ONE SUBJECT, who GETS BOTH TREATMENTS ONE AFTER THE OTHER - EACH SUBJECT makes up ONE PAIR of EXPERIMENTAL RESULTS Subjects are exposed to 2+ conditions in random order - Subjects "serve as their own control"

EXPERIMENTAL Group

In an Experiment, a GROUP of individuals RECEIVING a TREATMENT WHOSE EFFECT WE SEEK TO UNDERSTAND - The group exposed to the treatment In Ex 7.3: - The group that underwent gastric freezing

CONTROL Group

In an Experiment, it serves as a BASELINE WITH WHICH the EXPERIMENTAL GROUP is COMPARED - The group that goes UNTREATED In Ex 7.3: - It would be the group of individuals that undergo no treatment or a sham treatment

Randomization

In an Experiment, using IMPERSONAL CHANCE TO SELECT the GROUPS - Solves the problem of Bias

SINGLE-BLIND Experiment

In many medical studies, only the STATISTICIAN WHO DOES the RANDOMIZATION KNOWS WHICH TREATMENT EACH PATIENT IS RECEIVING In Ex 7.11 - The medical practitioner performing the real and sham acupuncture had to know which treatment each subject received

Ex 7.16: Children can't give fully informed consent, so the usual procedure is to ask their parents A study of new ways to teach reading is about to start at a local elementary school, so the study team sends consent forms home to parents - Many parents don't return the forms Can their children take part in the study because the parents did not say "no," or should we obtain data from only those children whose parents returned the form and said "yes"? - What about research into new medical treatments for people with mental disorders? - What about studies of new ways to help emergency room patients who may be unconscious?

In most of the latter cases, there isn't time to get the consent of the family - Does the principle of informed consent bar realistic trials of new treatments for unconscious patients? These are questions without clear answers - Reasonable people differ strongly on their opinions on all of them - Not everything about informed consent is simple

COMPLETELY RANDOMIZED Experimental Designs

In which ALL the INDIVIDUALS are ALLOCATED AT RANDOM AMONG ALL the TREATMENTS - The simplest experimental design assigns the individuals (subjects or experimental units) to the treatments completely at random Ex 7.5, about acupuncture for migraines, describes an experiment with this design, & Figure 7.3 illustrates the randomization process - Note that it is not necessary for this design to assign the same number of individuals to each treatment - It might be advantageous or more practical to assign a larger number of individuals to one particular treatment - This is the case for the acupuncture group in Ex 7.5 They can compare any number of treatments - They can also have more than one factor The chrysanthemum-flowering experiment of Ex 7.2 has two factors: photoperiod and light wavelength - Their combinations form the 12 treatments outlined in Figure 7.1 - This design assigns individual chrysanthemum plants at random to these 12 treatments - Once the layout of treatments is set, the randomization needed for this design is tedious but straightforward, as illustrated in Ex 7.5 They are the simplest statistical designs for experiments - They illustrate clearly the principles of comparison, randomization, and adequate number of subjects However, experiments sometimes need additional sophistication - Ex: to deal with substantial individual-to-individual variability or potentially confounding variables

DOUBLE-BLIND Experiment

In which NEITHER the SUBJECTS NOR the PEOPLE WHO ITNERACT WITH THEM KNOW WHICH TREATMENT EACH SUBJECT IS RECEIVING - This is obviously necessary whenever the investigator evaluates the experimental outcome In clinical trials, how a doctor interacts with a patient can also influence the patient's perception of his or her prognosis, thereby creating or reinforcing a placebo effect - This design type CONTROLS the INFLUENCE OF these HUMAN INTERACTIONS BY PREVENTING POTENTIAL BIAS

BLOCK Design

In which the RANDOM ASSIGNMENT OF INDIVIDUALS TO TREATMENTS is CARRIED OUT SEPARATELY WITHIN EACH BLOCK - Individuals sharing the same characteristic are pooled, forming a Block They can have BLOCKS OF ANY SIZE

Ex 7.15: A study wanted to examine the impact of the shape of the glass (straight or curved) on how fast adult social drinkers finish a beer or a soft drink

Informing participants of this objective would likely change their natural drinking behavior - So the researchers obtained special approval from their university's Faculty of Science Research Ethics Committee to disguise the purpose of their study The researchers told participants that the study examined "the effects of alcohol consumption on wordsearch performance," and they implemented the deception by having the participants actually complete a wordsearch after they had finished their drinks

Matched Pairs & Repeated-Measures designs use the principles of comparison of treatments and randomization - However, the randomization is not complete; we do not randomly assign all the subjects at once to the treatments

Instead, we randomize only within each matched pair - Or, when using the same individuals for different treatments, we randomize the order of the treatments for each individual This differs from a completely randomized design, which produces groups that are independent: - That is, the individuals in one group are completely unrelated to the individuals in the other group(s)

"Common Rule"

Principles that, by law, ALL STUDIES FUNDED BY the FEDERAL GOVERNMENT MUST OBEY 1. The organization that carries out the study must have an INSTITUTIONAL REVIEW BOARD that reviews all planned studies in advance to protect the subjects from possible harm 2. All individuals who are subjects in a study must give their INFORMED CONSENT in writing before data are collected - Subjects must be informed about the nature of a study and any risk of harm it may bring 3. All individual data must be kept CONFIDENTIAL - Only statistical summaries for groups of subjects may be made public - Any breach of confidentiality is a serious violation of data ethics - The best practice is to separate the identity of the subjects from the rest of the data at once

Ex 7.5: Acupuncture is an ancient traditional Chinese medicine that uses very thin needles inserted at specific body points to relieve pain and other conditions Researchers investigated whether acupuncture can help prevent migraine attacks - Patients suffering from migraines were randomly assigned to either acupuncture, sham acupuncture (needles inserted at nontraditional anatomic points), or a waiting list for acupuncture - The sham acupuncture acted as a placebo - The waiting-list group acted as a control (no medical treatment) The patients were asked to record headache days in diaries for 12 weeks - Figure 7.3 outlines the design of this study in graphical form

Random Assignment → Group 1 (145 subjects) → Treatment 1 (acupuncture) → Group 2 (81 subjects) → Treatment 2 (sham acupuncture) → Group 3 (76 subjects) → Treatment 3 (waiting list) → Record Headache Days The selection procedure is exactly the same as it is for sampling: label and select 1. Label the 302 patients as 001 to 302 2. Select - Use software, such as the Simple Random Sample applet, or use a table of random digits, such as Table A at the back of the book (reading successive three-digit groups) - The first 145 labels encountered select the acupuncture group - Ignore repeated labels and groups of digits not used as labels - After the first group is selected, the next 81 labels encountered select the sham acupuncture group - The remaining 76 labels make up the waiting-list group This design is: - Comparative because it compares 3 treatments (acupuncture, sham acupuncture, and waiting list) - Randomized because the subjects are assigned to the treatments by chance The "flowchart" outline presents all the essentials: randomization, the sizes of the groups and which treatment they receive, and the response variable - In this study, the researchers chose to place about half the patients in the acupuncture group and to split the other half between the two control treatments (sham and waiting list)

Logic Behind a Randomized Comparative Experimental Design

Random assignment of subjects forms groups that should be similar in all respects before the treatments are applied - Exercise 7.53 uses the Simple Random Sample applet to demonstrate this process - The random assignment also allows us to use probability laws to conduct statistical inference on the data collected Comparative design ensures that influences other than the experimental treatments operate equally on all groups Therefore, differences in average response must be due either to the treatments or to the play of chance in the random assignment of subjects to the treatments

Ex 7.11: A study compared the effect of acupuncture and the established drug flunarizine in preventing migraines among adult subjects suffering from recurring migraines - Patients can obviously tell the difference between needles and pills, but they cannot tell the difference between real and sham acupuncture or between a flunarizine pill and a placebo pill

So the researchers gave all subjects both types of treatment (needles and pills), in a DOUBLE-DUMMY design The subjects were randomly assigned to 2 groups: 1. One group received real acupuncture and placebo pills for four weeks 2. The other group received sham acupuncture and actual flunarizine pills for four weeks - All subjects kept diaries to record migraine occurrences and intensity

Apply Your Skills 7.1: The post-lunch dip is the drop in mental alertness after a midday meal - Does an extract of the leaves of the ginkgo tree reduce the post-lunch dip? A study assigns healthy people age 18 to 40 to take either ginkgo extract or a placebo pill - After lunch, they are asked to read seven pages of random letters and place an X over every e - We count the number of misses What are the individuals, the treatments, and the response variable in this experiment?

Subjects: the healthy people aged 18 to 40 Factor: the pill given to the subject Treatments: ginkgo or placebo Response variable: the number of e's missed by each subject

Experiments have Some Weaknesses in Detail

The ENVIRONMENT of an experiment can influence the outcomes in unexpected ways Although experiments are the gold standard for obtaining evidence of cause and effect, really convincing evidence usually requires that the study be REPLICATED successfully by its investigators as well as by independent investigators in different locations - All researchers are required to keep detailed records of their experiments and must provide these records if other teams have been unsuccessful at replicating the published results Always be very cautious in your interpretation of an experiment if the likely SOURCES OF BIAS have NOT ALL been carefully CONTROLLED FOR The most serious potential weakness of experiments is LACK OF REALISM - Lack of realism can limit our ability to apply the conclusions of an experiment to the settings of greatest interest - Most experimenters want to generalize their conclusions to some setting wider than that of the actual experiment - Ex: experiments in the health sciences aim to produce results that can be generalized to the entire population or to all individuals in the entire population diagnosed with a particular condition • This is very difficult to do when women and minorities are underrepresented or when potential differences between sexes or ethnic groups are not investigated Statistical analysis of an experiment cannot tell us how far the results will generalize - The ability to extend conclusions beyond the experimental setting depends on a thorough discussion of the biological merits of the experiment and the representativeness of the subjects used

Exercise 7.49: Researchers from Yale, working with medical teams in Tanzania, wanted to know how common infection with the human immunodeficiency virus (HIV), which causes acquired immunodeficiency syndrome (AIDS), is among pregnant women in that African country - To do this, they planned to test blood samples drawn from pregnant women Yale's institutional review board insisted that the researchers get the informed consent of each woman and tell her the results of the test - This is the usual procedure in developed nations - The Tanzanian government did not want to tell the women why blood was drawn or tell them the test results - The government feared panic if many people turned out to have an incurable disease for which the country's medical system could not provide care - The study was canceled Do you think that Yale was right to apply its usual standards for protecting subjects?

The Tanzanian government and the researchers both presented serious arguments - However, the study's purpose was entirely academic - Without patients' informed consent or any benefit to them, the study was not ethically justified

Tuskegee Syphilis Study

The Washington Evening Star published these lines in a front-page article in the summer of 1972: - "For 40 years, the U.S. Public Health Service has conducted a study in which human guinea pigs, not given proper treatment, have died of syphilis & its side effects" - "The study was conducted to determine from autopsies what the disease does to the human body" The article would end what was perhaps the most controversial human experiment ever in the United States & result in drastic changes to federal regulations related to scientific and medical experimentation requiring human subjects. The Public Health Service & the Tuskegee Institute began a study of syphilis in 1932 - Nearly 400 poor black men with syphilis from Macon County, Alabama, were enrolled in the study - They were never told they had syphilis & were never treated for it; even after penicillin became known as a highly effective cure by 1943 & widely available by 1947 - Also, they were never told that they were part of a medical experiment Syphilis is a sexually transmitted disease caused by the spirochete bacterium Treponema pallidum - Its pathology progresses in stages, starting with the development of hard ulcers primarily around the genitalia, which is followed by a transient, widespread rash - If the person does not receive treatment, the third stage of syphilis may develop, damaging the nervous system, eyes, heart, blood vessels, liver, bones, & joints, causing blindness, dementia, & paralysis - Untreated men can transmit the disease to their female partners, who can then transmit it to their future offspring • Approximately 40% of pregnancies where the mother has untreated syphilis result in stillbirth (baby dead at birth) • If the baby survives birth, it then has a 40% to 70% chance of being infected with syphilis • If left untreated, infected newborns may later become developmentally disabled, have seizures, or die because of the disease How could a study designed to deny treatment to uninformed subjects have occurred? - The historical context can help us understand the logical journey - At the end of the 19th century, philanthropists dreamed of black economic development, & the Tuskegee Institute was founded to help develop local schools, businesses, & agriculture - Its mission was later expanded to fight the devastating impact of a syphilis pandemic that affected nearly one-third of the African American population in the South - At that time, diagnosis & the currently known treatment were offered to participants - When the stock market crashed in 1929, funds for the endeavor were cut - In 1932, the Tuskegee philanthropic effort was recycled into a study of the impact of untreated syphilis - In exchange for participating in the study, the men were given free medical exams, free meals, & free burial insurance - The study was originally intended to run for 6 months, but it lasted 40 years until a whistleblower turned to the press The revelation of the Tuskegee syphilis study was the spark that initiated profound & long-lasting reforms of the ethical code of conduct for research involving human subjects in the United States - In 1974, the National Research Act was signed into law, creating the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research - Today, all federally supported studies using human subjects must be reviewed by institutional review boards, which decide whether study protocols meet ethical standards, & researchers must get voluntary informed consent from all study participants - Complementing this approach, training for researchers now emphasizes being aware of the ethical issues related to human experimentation - Ex: the National Institutes of Health now requires that all pre-baccalaureate, predoctoral, & postdoctoral trainees supported by institutional training grants receive instruction in the responsible conduct of research

Case-Control vs. Matched Pairs

The case-control design applies to observational studies (no treatment is imposed) - The matched pairs design applies to experiments Also, the selection of control individuals in case-control studies does not necessarily follow a strict matching process, and investigators may select more than one control for each case

Ex 7.14: Researchers and drug manufacturers have historically avoided testing new drugs on women - Yet these same drugs have been routinely prescribed regardless of gender - Of course, an untested drug could potentially be harmful to the fetus if a woman were to become pregnant In drug studies, this concern can easily be alleviated by enrolling only women willing to use contraceptives over the timeline of the clinical trial

The major reasons for omitting women from clinical research are variations in women's menstrual cycle and hormonal changes over time with menopause or oral contraceptive use - Response to treatment might be influenced by these hormonal fluctuations, confounding the results - Or the treatment itself might interfere with concurrent use of oral contraceptives Hormonal fluctuations during the menstrual cycle and the use of an oral contraceptive have indeed been shown to affect how drugs are metabolized - Ex: This is true of such common drugs as acetaminophen, aspirin, diazepam, and even caffeine - Some drugs, such as the antibiotic rifampin, also reduce the effectiveness of oral contraceptives The question is: Is this a good reason to exclude women from clinical trials? - Absolutely not! - We know that samples are representative of only the population from which they are taken - Clinical trials based entirely on men ensure drug effectiveness and safety for men alone - Not including women in clinical research does not protect them: It merely delays the assessment of efficacy and side effects until the drug is made available to the public Since the mid-1990s, the Food and Drug Administration (FDA) has required that clinical trials for drugs targeting both men and women include participants of both sexes, in proportions reflecting the actual proportions of men and women in the target population

Placebo Effect

The phenomenon when subjects genuinely felt better but did so because the care & attention they received made them feel better overall, not because of the treatment they were given - Experimental results caused by expectations alone - Any effect on behavior caused by the administration of an inert substance or condition, which the recipient assumes is an active agent - Psychological improvements due to participant expectation - The PSYCHOLOGICAL IMPACT OF UNDERGOING ANY KIND OF TREATMENT The effect is particularly strong when the response variable is pain - The mechanisms of this mind-body effect are unknown but are not without biological foundation - A neural response to this effect has been documented with neuroimaging techniques as early in the sensory process as the spinal cord - We also know that the nervous system interacts with the immune system and that depression, for instance, suppresses the immune response

Ex 7.2: Photoperiod, the relative lengths of light and dark periods in a 24-hour cycle, is a common environmental cue for flowering, allowing plants to bloom during an appropriate time of the year Chrysanthemums (Chrysanthemum morifolium) are known as short-day plants, blooming during times of longer nights Plant physiologists grew chrysanthemum plants in controlled greenhouses under different combinations of photoperiods (short day, long day, continuous light, and interrupted night) and light wavelengths (blue light, red light, and blue + red light) - The plants were kept in these conditions for 5 weeks and examined regularly to assess whether flowering had occurred during this time

This experiment has 2 factors (explanatory variables): 1. Photoperiod, with 4 values (short day, long day, continuous light, and interrupted night) 2. Light wavelength, with 3 values (blue, red, and blue + red) The 12 combinations of both factors form 12 treatments - Figure 7.1 shows the layout of the treatments The individuals (or experimental units) are chrysanthemum plants - The response variable is flowering, a categorical variable

Ex 7.3: Gastric freezing was introduced in the 1960s by a prominent surgeon as a way to relieve ulcer pain - Patients would swallow a deflated balloon that was then filled with a refrigerated liquid - The idea was that the refrigerant would cool the stomach lining, reduce acid production, and reduce pain - The procedures performed showed that gastric freezing did reduce ulcer pain, and the treatment was recommended based on this evidence

This experiment has a very simple design - A group of subjects (the patients) were exposed to a treatment (gastric freezing), and the outcome (pain reduction) was observed Here is the design: - Subjects → Gastric freezing → Pain reduction Some critical pieces of information are missing from this experiment - What would have happened if we had left the patients alone? - Would their pain have abated even without intervention? - What if patients were inclined to say they were feeling better (maybe because they expected to get better, or to please the experimenter)? - And what if patients genuinely felt better but did so because the care and attention they received made them feel better overall, not because of the cooling of their stomach lining? • This is a real possibility, and similar psychological improvements have been documented in many human experiments • We call it the placebo effect None of these questions can be addressed by the simple experiment - The effect of gastric freezing is confounded with the effect of 2 lurking variables: spontaneous, natural improvement and psychological improvement due to medical care - Figure 7.2 shows the confounding in picture form - Because of confounding, subject improvement cannot be unequivocally attributed to the effect of gastric freezing

Ex 7.10: In Ex 7.8 we saw how researchers used a matched pairs design to test the efficacy of the ACT device in patients with essential tremor - They concluded that the device significantly reduced spoon tremor, based on accelerometer data and on clinical evaluation by a neurologist - This means that the tremors were so much smaller with the ACT-on condition that chance alone would not likely produce such results

This finding is not quite enough, however, to convince us that the ACT device is effective - What if the subjects or the neurologist were somehow biased (consciously or not) by the knowledge of the device being activated? Such bias is unlikely in this case because the study was DOUBLE-BLIND: As the researchers reported,... - "Subjects and the evaluating neurologist were not told whether the device was on or off" - "The device prototype also does not emit a sound or cause a palpable sensation that can be detected by the user when turned on" Who turned the device on & off? - A person not otherwise involved with the experiment &, therefore, unlikely to influence the results

Ex 7.4: The simple experiment outlined in Example 7.3 was followed up many years later with a randomized, controlled experiment in which patients undergoing gastric freezing were compared with similar patients undergoing the same procedure except that the liquid pumped into the balloon was not refrigerated

This second group was a placebo group The comparative study found that the condition of 28 of the 82 patients who received the gastric-freezing treatment improved, while the condition of 30 of the 78 patients in the placebo group improved - This comparative experiment demonstrated that cooling the stomach lining was not more effective in reducing ulcer pain than a sham procedure - The treatment was then abandoned

Ex 7.1: Congenital nasolacrimal duct obstruction, or blocked tear duct, is a fairly common condition in infants, characterized by excessive tear production and mucus discharge in one or both eyes - The condition can be treated immediately in office by surgically probing the blocked tear duct after application of a topical anesthetic - An alternative approach is to defer treatment and continue observation until either the condition resolves itself or medical care becomes clearly necessary

To find out which approach is more effective, researchers assigned infants between six and nine months of age to 2 groups: 1. Immediate probing (82 infants) 2. Deferred treatment (81 infants) The infants were examined by a doctor up to 18 months later to determine if and when the condition had fully resolved - The assignment to one treatment group or the other was random, and the 2 groups did not differ substantially in demographics or initial severity of symptoms - To avoid possible confounding health factors, all infants enrolled in the study were otherwise healthy This experiment compares 2 treatments: 1. Immediate probing 2. Deferred treatment There is a single factor: "treatment type" The subjects are infants diagnosed with blocked tear duct The main response variable is the time of symptom resolution - (The researchers also recorded all related medical care and associated costs)

Even with an excellent experimental design, we can never say that any difference between two groups is entirely due to differences in the treatments applied

Variables, by definition, vary from individual to individual - We would not trust an experiment like that of Ex 7.5 if it had just 1 patient in each group because the results would depend too much on which group got lucky and received a patient with less frequent headaches, for example If we assign many subjects to each group, however, the effects of chance will average out, and there will be little difference in the average responses in the three groups unless the treatments themselves cause a difference - "USE ENOUGH SUBJECTS TO REDUCE CHANCE VARIATION" is the third big idea of statistical design of experiments

While it is Important to Question the Necessity & Ethicality of Human & Animal Experimentation...

We also must CONSIDER the IMPLICATIONS of NOT PERFORMING CAREFULLY DESIGNED EXPERIMENTS - There are many unfortunate examples of pharmaceutical or surgical methods that have been or still are widely used without going through this important preliminary stage In such a situation, the whole population becomes unsuspecting test subjects of an "experiment" that has: - No guidelines - No safeguards - No way to reach informative conclusions about the benefits & dangers of the drug or procedure Ex: - In the 1950s & 1960s, Lobotomy (the severance of the frontal lobes from the rest of the brain) became a popular method for dealing with psychotic/mentally ill individuals • After thousands of lobotomies were performed in the U.S. alone, it became clear that the procedure did provide improvement for some patients, but others showed no improvement at all, & the rest suffered irreversible, lifelong psychological damage - Gastric freezing for the treatment of stomach ulcers (Ex 7.3) & mammary artery ligation for the treatment of angina (Ex 15.12, page 390) are two medical procedures that were widely used • Until Randomized Comparative Experiments showed that their results were not any better than those obtained with placebos, leading to the procedures being abandoned - The use of routine Episiotomy to speed up labor & prevent natural tearing of the perineum during childbirth has been questioned by multiple studies, & its use has been declining rapidly - Fortunately, these drastic practice reversals have led to a greater awareness of the necessity for careful experimentation to legitimize the use of medical & surgical procedures in humans

Exercise 7.35: An ultramarathon, as you might guess, is a footrace longer than the 26.2 miles of a marathon - Runners commonly develop respiratory infections after an ultramarathon Will taking 600 milligrams of vitamin C daily reduce these infections? - Researchers randomly assigned ultramarathon runners to receive either vitamin C or a placebo - Separately, they also randomly assigned these treatments to a group of nonrunners the same age as the runners - All subjects were watched for 14 days after the big race to see if infections developed a. What is the name for this experimental design? b. Use a diagram to outline the design

a. A block design

Apply Your Knowledge 7.5: The changing climate will probably bring more rain to California, but we don't know whether the additional rain will come during the winter wet season or extend into the long dry season in spring and summer Kenwyn Suttle of the University of California at Berkeley and his coworkers carried out a randomized controlled experiment to study the effects of more rain in either season - They randomly assigned plots of open grassland to 3 treatments: added water equal to 20% of annual rainfall either during January to March (winter) or during April to June (spring), and no added water (control) - Thirty-six circular plots of area 70 square meters were available, of which 18 were used for this study - One response variable was total plant biomass, in grams per square meter, produced in a plot over a year a. Outline the design of this experiment, following the model of Figure 7.3 b. Number all 36 plots and choose 6 at random for each of the 3 treatments - Be sure to explain how you did the random selection

a. Assign 6 plots to each treatment b. If using Table A, label 01 to 36 and take two digits at a time

Exercise 7.33: A study examined the impact of cell phone use on brain metabolism in 47 healthy adults - Here is a quote from the published abstract: "Cell phones were placed on the left and right ears and positron emission tomography with (18F) fluorodeoxyglucose injection was used to measure brain glucose metabolism twice, once with the right cell phone activated (sound muted) for 50 minutes (ON condition) and once with both cell phones deactivated (OFF condition)" This experiment has a matched pairs design - How can you tell? b. What is the advantage of using a matched pairs design rather than a completely randomized design in this context? c. The abstract reports that "metabolism in the region closest to the antenna (orbitofrontal cortex and temporal pole) was significantly higher for ON than OFF conditions" - Explain in simple terms what "significantly higher" means in this context

a. Each subject was scanned twice (ON and OFF conditions) b. Different subjects might have different brain activities - The matched pairs design allows comparison of the ON and OFF conditions for each subject c. The difference could not easily be explained by chance variations alone

Exercise 7.25: The most common treatment for breast cancer discovered in its early stages was once removal of the breast - Today, the usual practice is to remove only the tumor and nearby lymph nodes, with the surgery then being followed by radiation To study whether these treatments differ in their effectiveness, a medical team examines the records of 25 large hospitals and compares the survival times after surgery of all women who have had either treatment a. What are the explanatory and response variables? b. Explain carefully why this study is not an experiment c. Explain why confounding will prevent this study from discovering which treatment is more effective - (The current treatment was recommended after several large randomized comparative experiments)

a. Explanatory: type of operation - Response: survival time b. Not an experiment: existing records are simply examined c. The type of operation is most likely decided based on the size of the tumor - Larger tumors would be much more dangerous to remove, regardless of type of operation

Exercise 7.41: Observational studies had initially suggested that vitamin E may reduce the risk of heart disease - Careful experiments, however, showed that vitamin E has no effect, at least for women According to a commentary in the Journal of the American Medical Association, vitamin E is yet another treatment that may have been "promising in epidemiologic and observational studies but failed to deliver in adequately powered randomized controlled trials" - The author attributes this reversal to a healthy-user bias, "the healthy lifestyle behaviors that characterize individuals who care enough about their health to take various supplements [and] are actually responsible for the better health" A friend who knows no statistics asks you to explain this commentary a. What is the difference between observational studies and experiments? b. What is a "randomized controlled trial"? (We'll discuss "adequately powered" in Chapter 15.) c. How does a healthy-user bias explain how people who take vitamin E supplements have better health in observational studies but not in controlled experiments?

a. In an observational study, we observe subjects who have chosen to take supplements and compare them with others who do not take supplements - In an experiment, we assign some subjects to take supplements and others to take a placebo b. Treatments are assigned at random, and a control group is used as a basis for comparison to observe the effects of the treatment c. Subjects who choose to take supplements are more likely to make healthy lifestyle choices - When random assignment is used, some of those subjects will take the supplement and some will take the placebo

Apply Your 7.3: In the United States, marketing "healthy" food and supplements does not require demonstrated health benefits, though any specific health claim must be backed by scientific evidence - False and unsubstantiated advertising claims are investigated by the Federal Trace Commission (FTC) The makers of PomWonderful juice and POMx supplements had made a number of advertising claims such as: - "SUPER HEALTH POWERS! 100% PURE POMEGRANATE JUICE" - "Backed by $25 million in medical research" - "Proven to fight for cardiovascular, prostate and erectile health" After conducting an investigation, the FTC ordered "POM's future disease treatment and prevention claims to be supported by at least one randomized, well-controlled human clinical trial" a. A clinical trial is an experiment using human subjects for health-related research purposes - Explain why the FTC insists that they should be "randomized" b. Explain the importance of a clinical trial being "well-controlled" in the context of POM products c. The FTC also prohibited POM makers from misrepresenting scientific evidence, stating, for instance, that "the erectile dysfunction claims are false and unsubstantiated because the study on which the company relied did not show that POM Juice was any more effective than a placebo" - Explain the reasoning behind this argument

a. Randomization prevents bias in the assignment to conditions b. Without a control group serving as comparison, whatever effect would be found with POM consumption could be completely unrelated to POM consumption c. If a positive outcome occurs whether men consumed POM or a placebo, then the outcome cannot be attributed to POM consumption

Apply Your Knowledge 7.9: A psychologist wants to know if the difficulty of a task influences our estimate of how long we spend working at it She designs two sets of mazes that subjects can work through on a computer - One set has easy mazes and the other has hard mazes - Subjects work until told to stop (after six minutes, but subjects do not know this). They are then asked to estimate how long they worked - The psychologist has 30 students available to serve as subjects a. Describe the design of a completely randomized experiment to learn the effect of task difficulty on estimated time spent on the task b. Describe the design of a matched pairs experiment using the same 30 subjects c. Describe the design of a block experiment considering men and women separately

a. Randomly assign 15 students to Group 1 (easy mazes) and the other 15 to Group 2 (hard mazes) - Compare each group's time estimates b. Each student does the activity twice, once with the easy mazes and once with the hard mazes, in a randomly determined order - Compare each student's easy and hard time estimates c. Randomly assign half the men and half the women to Group 1, and the remaining subjects to Group 2 - Compare each group's time estimates

Exercise 7.31: Doctors identify "chronic tension-type headaches" as headaches that occur almost daily for at least six months - Can antidepressant medications or stress management training reduce the number and severity of these headaches? - Are both together more effective than either alone? a. Create a diagram like Figure 7.1 to display the treatments in a design with two factors: medication (yes or no) and stress management (yes or no) - Then outline the design of a completely randomized experiment to compare these treatments b. The headache sufferers named here have agreed to participate in the study - Randomly assign the subjects to the treatments - If you use the Simple Random Sample applet or other software, assign all the subjects - If you use Table A, start at line 130 and assign subjects to only the first treatment group Abbott Decker Herrera Lucero Richter Abdalla Devlin Hersch Masters Riley Alawi Engel Hurwitz Morgan Samuels Broden Fuentes Irwin Nelson Smith Chai Garrett Jiang Nho Suarez Chuang Gill Kelley Ortiz Upasani Cordoba Glover Kim Ramdas Wilson Custer Hammond Landers Reed Xiang

a. Randomly assign the subjects to four groups: - Group 1, antidepressants and stress management - Group 2, stress management only - Group 3, antidepressants only - Group 4, control b. Using Table A; - Group 1 = Chai, Hammond, Herrera, Xiang, Irwin, Hurwitz, Reed, Broden, Lucero, Nho

Exercise 7.37: The law allows marketers of herbs and other natural substances to make health claims that are not supported by evidence - Brands of ginkgo extract claim to "improve memory and concentration" A randomized comparative experiment found no evidence for such effects - The subjects were 230 healthy people older than age 60 years - They were randomly assigned to receive ginkgo or a placebo pill - All the subjects took a battery of tests for learning and memory before treatment started and again after six weeks a. Following the model of Figure 7.3, outline the design of this experiment b. Use the Simple Random Sample applet, other software, or Table A to assign half the subjects to the ginkgo group - If you use software, report the first 20 members of the ginkgo group (in the applet's "Sample bin") and the first 20 members of the placebo group (those left in the "Population hopper") - If you use Table A, start at line 103 and choose only the first 5 members of the ginkgo group

a. Randomly sample half of the subjects and give them the treatment; everyone else gets the placebo pill - Then compare the results b. Using Table A: 170, 005, 227, 118, 007

Apply Your Knowledge 7.7: The timing of peanuts' first introduction in children's diet could affect allergic reactions, so physicians have debated whether to recommend early exposure or avoidance A study enrolled 640 infants with severe eczema and split them into two groups based on whether they also had a diagnosed peanut allergy - Within each group, the researchers randomly assigned the infants to either completely avoid peanuts or consume peanuts in small amounts regularly until they reach 60 months of age - At the end of the study, an allergy test was conducted to establish which children had a peanut allergy a. What are the individuals, the treatments, and the response variable in this experiment? b. What is the design of this experiment? - Outline this design as in Figure 7.4 c. Of the 640 infants in the study, only 98 had a diagnosed peanut allergy at the beginning of the study - Explain why you think the researchers decided to split the infants into two groups before randomly assigning them to the conditions

a. The individuals are infants with severe eczema - The treatments are peanut consumption and peanut avoidance - The response variable is peanut allergy at 60 months of age b. Block design c. A pre-existing peanut allergy could be influential, so to have these subgroups evenly represented in both treatments

Apply Your Knowledge 7.11: Dark chocolate can dilate arteries by down-regulating the oxidative stress A study investigated whether dark chocolate improves walking autonomy in patients with peripheral artery disease (PAD) - The study compared the maximum walking distance (in meters) for 20 PAD patients after they ate 40 g of dark chocolate and also after they ate 40 g of milk chocolate, on two different weeks - The published findings mention that "an individual not involved in the study randomly allocated the participants to a treatment sequence starting with dark chocolate or with milk chocolate, and kept this information in a sealed envelope. The authors and laboratory technicians were unaware of the treatment allocation." a. What are the individuals, the treatments, and the response variable in this experiment? b. What is the design of this experiment? - What type of blinding, if any, was used? - Explain your answer

a. The individuals are patients with PAD - The treatments are milk chocolate and dark chocolate - The response variable is maximum walking distance b. Matched pairs design (repeated measures) - Single blind (subjects can taste the difference between dark chocolate and milk chocolate)

Exercise 7.47: How do you deliver oxygen to cells in emergency situations when the lungs have failed? A team of researchers devised a new method that encapsulates oxygen within microparticles of fat so that oxygen can be injected directly into the bloodstream without risking an embolism - The method was tested experimentally in rabbits and shown to be successful - But how do you then test it in humans? a. One step would involve giving the oxygen injections to healthy volunteers to test its safety - What are some ethical challenges to using healthy volunteers to test an unknown medical procedure? b. Another step would require giving the oxygen injections during real-life emergencies involving oxygen deprivation - Obtaining informed consent from the patient is likely to be impossible, and there may not be enough time to obtain informed consent from family members - Would it be ethical to test the procedure anyway?

a. The procedure might be harmful in humans b. It might be ethical to test the procedure without consent if it was believed that the patient would not survive otherwise

Exercise 7.39: Here is an excerpt from a 2007 NBC news report: - Millions of women suffer from severe mood swings and hot flashes associated with menopause and other causes of estrogen depletion. . . . Lipov is treating menopausal women by using an injection which he believes can reset the brain's thermometer to reduce or eliminate hot flashes. . . . Lipov is the first and only doctor in the U.S. to target hot flashes - To make the procedure more safe, he first injects a dye into the targeted area - Complications, like seizures, are rare and so far Lipov's patients have had no problems - The hot flash relief lasts from weeks to months, Lipov said, and almost all of the 22 women he's used the technique on have gotten some benefit a. What is the design of this study? b. Can we trust the results of this study? - Explain why or why not

a. This is a nonrandomized, noncomparative study b. No, because subjects were not randomly selected (they were hand-picked by Lipov) and the lack of comparison means that results could be a placebo effect

Exercise 7.29: Bird songs have been hypothesized to be a secondary sexual character signaling an individual's health status Researchers designed an experiment in which they randomly assigned male collared flycatchers (Ficedula albicollis) to two groups - One group received an immune challenge in the form of an injection of sheep red blood cells, and the other group received a placebo injection - The researchers then recorded the changes in song rate (in strophes per minute) after the injection for the 15 males in the immune challenge group and the 12 males in the placebo group a. Outline the design of this experiment b. Label the birds from 1 through 27 - Use software or Table A at line 131 to carry out the randomization your design requires

a/ Completely randomized design: 12 birds randomly selected for the placebo group, with the remaining 15 making up the immune challenge group (or the other way around) b. Using Table A to select the 12 birds in the placebo group: 05, 19, 04, 25, 20, 16, 18, 07, 13, 02, 23, 27