Chapter 5 STAT 1312

Do multivitamin supplements improve health? To answer this question, researchers recruited 2000 adults. All were provided with supplies of capsules and were asked to take one capsule per day. One thousand of the adults received capsules that were multivitamin supplements and one thousand received capsules that were a placebo. The researchers used an extensive questionnaire to assess the health of all participants at the start of the study and after two years into the study. Outline the design of this study using a diagram.

(Left to right) (left) Random Assignment (top) Group 1 (bottom) Group 2 (next top) Treatment 1 (next bottom) Treatment 2 (right) Compare health through questionnaire Extended Answer: The diagram should represent the essential information about the design: random assignment to groups, one group for each treatment, the number of subjects in each group, what treatment each group gets, and the response variable that the researchers compare. At the first step, 2000 adults were randomly divided into two equal groups. The first group of 1000 adults took multivitamin supplements, and the second group took placebos. As a result of the experiment, information about the health of each person was obtained through a questionnaire. After that, the data for each group were compared. The diagram correctly displaying all stages of the experiment is shown. A plan has stages of the experiment as nodes. There are two branches from the root, Random assignment. The first node is labeled Group one. 1000 adults. The second node is labeled Group two. 1000 adults. Group one node is linked with node labeled Treatment one. Multivitamin supplements. Group two node is linked with node labeled Treatment two. Placebo. Both Treatment one and Treatment two nodes are linked with the end node, Compare health through questionnaire.

Can aspirin help prevent heart attacks? The Physicians' Health Study, a large medical experiment involving 22,000 male physicians, attempted to answer this question. One group of about 11,000 physicians took an aspirin every second day, while the rest took a placebo. After several years, the study found that subjects in the aspirin group had significantly fewer heart attacks than subjects in the placebo group. Use a diagram to outline the design of the Physicians' Health Study. (When you outline the design of an experiment, be sure to indicate the size of the treatment groups and the response variable. The diagrams in Figures 5.2 and 5.3 are models.) Please place group one above group two in your experiment design.

(Left to right) Random assignment --->11000 physicians(group 1) & 11000 physicians(group 2) Group 1 ---> Aspirin Group 2 ---> Placebo Aspirin & Placebo ----> Observe heart attacks Extended Answer: The goal of the Physicians' Health Study was to determine if aspirin lowered the chances of heart attacks. The researchers began the study by randomly assigning two equal groups of 11000 male physicians. The researchers gave the first group aspirin and gave the second group a placebo. Researchers recorded the number of heart attacks in each group. The study brought the resulting number of heart attacks together for comparison after several years..

For which of the following studies would it be possible to conduct a randomized comparative experiment?

A study to determine if taking Tylenol dulls your emotions. Extended Answer: Randomized comparative experiments are used to investigate a cause-and-effect relationship. Researchers can use randomized comparative experiments to determine if Tylenol dulls a person's emotions. The explanatory variable would be Tylenol and the response variable would be a person's records of their emotions. The treatment groups could be different doses of Tylenol. Researchers could create a control group by giving one group of subjects a placebo. Subjects in the treatment groups and in the control group could score their emotions using a scale from 1 to 5. The researchers could compare the emotional scores of the treatment groups to the emotional scores of the control group to see if there is a difference in reported emotions. It would not be possible to conduct an experiment to investigate the possible connection between a person's birth month and the length of a person's life. Researchers conduct experiments by assigning a treatment to determine a potential cause-and-effect relationship. Researchers cannot assign a birth month to a subject. This is an example of an observational study. A person's sex may impact a person's salary but researchers cannot investigate this claim using an experiment. Researchers cannot assign a sex to a person. A person's sex can only be observed. There may also be a link between the wealth of parents and the wealth of their children, but wealth cannot be assigned in an experiment. Researchers would not be able to randomly assign subjects to a treatment.

An article in Newsweek reported that, to investigate how an unhappy marriage can affect an individual's health, scientists recruited 43 healthy couples between 24 and 61 years old who had been married for at least three years. The researchers asked couples to discuss touchy topics likely to spark disagreement, such as money or in‑laws, and taped the conversations. They used this footage to analyze verbal and non-verbal modes of conflict, including eye rolling. The team also took blood samples from the couples before and after arguing, and found those who were most hostile toward their spouses had higher levels of LPS-binding protein, a biomarker for a leaky gut. Scientists found the highest levels of LPS‑binding protein in participants who had the nastiest fights and a history of mood disorders such as depression. The biomarker was also linked to inflammation in the body. Couples choose to argue and engage in hostile behavior when discussing touchy subjects. And anger and unhappiness that can lead to fighting may be symptoms of a physiological or mental health problem. Label the diagram to illustrate your explanation.

Causes?: increased levels of hostility, physiological or mental health problems ----> higher levels of LPS- binding protein Extended Answer: The goal of the study was to investigate marital fighting and overall health. The researchers analyzed the blood of married couples and found that those who were fighting had evidence of a leaky gut. However, physiological or mental health problems may be a symptom of fighting. Therefore, physiological or mental health problems may be causing increased levels of hostility and higher levels of LPS‑binding protein. Lurking variables are variables that the researchers did not include in the original study but influence the other variables of the study. Physiological or mental health problems may be causing increased fighting or higher levels of LPS‑binding protein. Researchers did not include physiological and mental health problems in the study. The physiological and mental health problems of the subjects were a potential lurking variable. A study has confounding variables if two or more variables have indistinguishable influences on a response variable. People who have physiological and mental health problems may have hostile marital fights because of these underlying issues. The physiological and mental heatlh problems of an individual might cause them to have higher amounts of LPS‑binding protein. Therefore, physiological or mental health problems are a confounding variable. The researchers conducted the study to determine whether increased levels of hostility led to higher levels of LPS‑binding protein. You can illustrate the relationship between hostility and the protein by pointing an arrow from hostility towards higher levels of the protein. Physiological or mental health problems may be causing increased levels of hostility and higher levels of LPS‑binding protein. You can demonstrate this relationship by pointing arrows from these health problems to the variables of increased hostility and higher protein.

A study closely followed a large social network of 12067 people for 32 years, from 1971 until 2003. The researchers found that when a person gains weight, close friends tend to gain weight, too. The researchers reported that obesity can spread from person to person, much like a virus. Explain why the fact that when a person gains weight, close friends also tend to gain weight does not necessarily mean that weight gains in a person cause weight gains in close friends. In particular, identify some lurking variables whose effect on weight gain may be confounded with the effect of weight gains in close friends. Place the text in the appropriate bins to illustrate your explanation.

Explanatory Variable: Weight Gain Response Variable: Weight gain in friends Lurking Variable: Shared Activities Extended Answer: The researchers found that when a person gains weight, close friends tend to gain weight, too. So, in their opinion, a person's weight gain causes weight gain in friends. Therefore, weight gain is considered to be an explanatory variable and weight gain in friends is considered to be a response variable. From the previous step, the lurking variable is shared activities (drinking, going out to eat, lack of exercise). This factor affects any person's weight gain in a group of friends. Therefore, this variable has an important effect on the relationship between one's weight gain and the weight gain in friends. The following figure illustrates these connections. A diagram shows the relationships among three variables: weight gain (explanatory variable), weight gain in friends (response variable), and shared activities (lurking variable). There is an arrow pointing from the weight gain to weight gain in friends with the word Causes? above it. There are also two arrows pointing away from shared activities; one points to weight gain and the other points to weight gain in friends.

People who eat lots of fruits and vegetables have lower rates of colon cancer than those who eat little of these foods. Fruits and vegetables are rich in "antioxidants" such as vitamins A, C, and E. Will taking antioxidants help prevent colon cancer? A clinical trial studied this question with 864 people who were at risk for colon cancer. The subjects were divided into four groups: daily beta-carotene, daily vitamins C and E, all three vitamins every day, and daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. What are the explanatory and response variables in this experiment?

Explanatory variable: Vitamin Regimen Response variable: Development of Colon Cancer Extended Answer: Explanatory variables are variables that cause the change in the study. Response variables are variables that measure the change caused by the explanatory variables. The researchers gave the subjects different antioxidants. Researchers gave one group of subjects a treatment of daily beta-carotene, another group daily vitamins C and E, a third group all three vitamins, and a fourth group a placebo. Because the researchers expect the treatments to cause a change in the subject's chances of developing colon cancer, the vitamin regimen is the explanatory variable. The researchers followed each subject over four years. Researchers recorded the number of subjects that developed colon cancer in each treatment group to see if the treatments caused a different number of subjects to develop colon cancer. As the researchers were using the subject's colon cancer to measure the effects of the treatments, the development of colon cancer was the response variable.

People who eat lots of fruits and vegetables have lower rates of colon cancer than those who eat little of these foods. Fruits and vegetables are rich in "antioxidants" such as vitamins A, C, and E. Will taking antioxidants help prevent colon cancer? A clinical trial studied this question with 864 people who were at risk for colon cancer. The subjects were divided into four groups: daily beta-carotene, daily vitamins C and E, all three vitamins every day, and daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. Suggest some lurking variables that could explain why people who eat lots of fruits and vegetables have lower rates of colon cancer. The results of the experiment suggest that these variables, rather than the antioxidants, may be responsible for the observed benefits of fruits and vegetables. Select the variables that represent potential lurking variables.

Fruits and vegetables contain fiber; this could account for the benefits of those foods. Persons who eat lots of fruits and vegetables may have healthier diets overall. Extended Answer: Previous studies have demonstrated a connection between fiber and lower rates of colon cancer. Fruits and vegetables are high and fiber. It may be the fiber in fruits and vegetables that reduces the rate of colon cancer instead of the antioxidants. Researchers have associated healthy lifestyles with lower rates of cancer. People who have healthy diets tend to eat more fruits and vegetables. It may be the healthier lifestyles, not the antioxidants, that reduces the risk of colon cancer. The goal of the clinical trial was to determine whether the antioxidants would reduce the rates of colon cancer. Lurking variables are variables that are not a part of the original study. Because researchers included antioxidants in the original study, antioxidants are not a lurking variable. Fruits and vegetables may reduce the chance of heart disease. However, the research is investigating the relationship between fruits, vegetables, and rates of colon cancer, not the relationship between fruits, vegetables, and heart disease. It has been speculated that pesticides lead to more cases of colon cancer. If it is true that pesticides lead to more cases of colon cancer, and if pesticides are on fruits and vegetables, pesticides would explain higher rates of colon cancer, not lower rates of colon cancer.

A large study used records from Canada's national health care system to compare the effectiveness of two ways to treat prostate disease. The two treatments are traditional surgery and a new method that does not require surgery. The records described many patients whose doctors had chosen one or the other method. The study found that patients treated by the new method were significantly more likely to die within eight years. (a) Further study of the data showed that this conclusion was wrong. The extra deaths among patients treated with the new method could be explained by lurking variables. What lurking variables might be confounded with a doctor's choice of surgical or nonsurgical treatment? For example, why might a doctor avoid assigning a patient to surgery? Choose the best explanation of the described fact.

If a patient has a little chance to survive during an operation, a doctor might choose not to recommend surgery. Extended Answer: (a) The experiment does not exclude the influence of external factors. Each patient is not randomly assigned to the first or second treatment since a doctor assigns a patient to surgery or alternative methods. A doctor takes into account many factors to assign a patient to treatment. These factors could be age, common health status of a patient or another. So, there are several lurking variables. In a serious case, when the patient has little chance of surviving, a doctor might choose not to recommend surgery. The surgery method might be seen as an unnecessary measure, bringing expense and a hospital stay with little benefit to the patient. Patients may have chosen the non‑surgical treatment out of fear of surgery, but it does not explain the extra deaths among patients treated with the new method. The non‑surgical treatment can be in effect a placebo, but the question statement says that further study of the data showed that the conclusion about the effectiveness of the new method was wrong. Therefore, after excluding lurking variables, patients treated by the new method were not significantly more likely to die within eight years. The new method is not a placebo. It is possible that only wealthy patients can afford the surgery if state‑paid surgeries are not performed. Socioeconomic status can be a lurking variable, but Canada has universal health care and this factor is unlikely to be significant. (b) The diagram should represent the essential information about the design: random assignment to groups, one group for each treatment, the number of subjects in each group, what treatment each group gets, and the response variable that the researchers compare. In this experiment, 300 prostate patients were randomly divided into two equal groups. The first group of 150 patients were assigned to surgery and the second group were assigned to the alternative method. As a result of the experiment, observe the recovery of patients in each group. The diagram correctly displaying the design of the experiment is shown. A plan has stages of the experiment as nodes. There are two branches from the root, Random assignment. The first node is labeled Group one. 150 patients. The second node is labeled Group two. 150 patients. Group one node is linked with node labeled Treatment one. Surgery. Group two node is linked with node labeled Treatment two. Alternative. Both Treatment one and Treatment two nodes are linked with the end node, Observe recovery.

Does church attendance lengthen people's lives? One study of the effect of attendance at religious services gathered data from 2001 obituaries. The researchers measured whether the obituaries mentioned religious activities and length of life. Choose the correct statement about the study.

In this study, mention of religious activities is the explanatory variable and length of life is the response variable. Extended Answer: To determine the type of the variable, use the definitions of the explanatory and response variables. An explanatory variable is a variable that might explain or causes changes in the response variable. A response variable is a variable that measures an outcome or result of a study. In this study, the researchers were interested in how religious activities can change length of life. Therefore, the explanatory variable is the mention of religious activities, not length of life nor the 2001 obituaries. The researchers examined the changes in life expectancy based on church attendance. This means that length of life is the outcome or result of interest. Therefore, length of life is the response variable.

An article in Newsweek reported that, to investigate how an unhappy marriage can affect an individual's health, scientists recruited 43 healthy couples between 24 and 61 years old who had been married for at least three years. The researchers asked couples to discuss touchy topics likely to spark disagreement, such as money or in‑laws, and taped the conversations. They used this footage to analyze verbal and non-verbal modes of conflict, including eye rolling. The team also took blood samples from the couples before and after arguing, and found those who were most hostile toward their spouses had higher levels of LPS-binding protein, a biomarker for a leaky gut. Scientists found the highest levels of LPS‑binding protein in participants who had the nastiest fights and a history of mood disorders such as depression. The biomarker was also linked to inflammation in the body. Couples choose to argue and engage in hostile behavior when discussing touchy subjects. And anger and unhappiness that can lead to fighting may be symptoms of a physiological or mental health problem. Explain why these facts make any conclusion about cause and effect untrustworthy.

Physiological or mental health problems may be causing increased levels of hostility and higher levels of LPS-binding protein. Thus, the association between increased levels of hostility and higher levels of LPS-binding protein is confounded by the lurking variable of physiological or mental health problems.

Can aspirin help prevent heart attacks? The Physicians' Health Study, a large medical experiment involving 22,000 male physicians, attempted to answer this question. One group of about 11,000 physicians took an aspirin every second day, while the rest took a placebo. After several years, the study found that subjects in the aspirin group had significantly fewer heart attacks than subjects in the placebo group. What do you think the term "significantly" means in "significantly fewer heart attacks?"

Statistically significant means " unlikely to have occured by chance if there was no difference between the aspirin and placebo groups." Extended Answer: An event is statistically significant if its effect is so large that it would rarely occur by chance. The conclusion of the study is that the physicians who took aspirin had significantly fewer heart attacks than the physicians who did not take aspirin. The effect of aspirin was so large that the researchers concluded that it was unlikely that the difference occurred by chance. Statistically significant results are not results that are likely to occur by chance. Statistically significant results are effects that are larger than what researchers could expect due to random chance.

Researchers identified 2587 10th grade students in the Los Angeles area who did not have significant symptoms of ADHD (Attention-Deficit/Hyperactivity Disorder). These students were followed for approximately two years. The frequency with which each student used digital media over the two‑year period was recorded. At the end of the two‑year period, students were again tested for symptoms of ADHD. Researchers found that the higher the frequency of digital media use, the more likely the student was to have developed symptoms of ADHD at the end of the study. Identify potential confounding variables and explain why confounding prevents us from concluding that the more one uses digital media, the more likely one is to develop ADHD.

Potential confounding variables: sleep disorders, cognitive disability, parental involvement Confounding makes it difficult to determine whether digital media use or the confounding variables cause an increase in ADHD symptoms. Extended Answer: Confounding variables are lurking or explanatory variables whose influence on the response (dependent) variable cannot be distinguished from the explanatory (independent) variable being studied. Parental involvement, the presence of a cognitive disability, and sleep disorders are all examples of potential confounding variables. For instance, if a sleep disorder is associated with both a higher level of digitial media usage and ADHD, the effect of the sleep disorder on ADHD cannot be distinguished from the effect of digital media usage alone. Similarly, parental involvement, or lack of it, and cognitive disabilities, could be associated with higher levels of digital media usage and symptoms of ADHD. Race and the age of any siblings, are unlikely to confound the association between digital media use and the development of symptoms of ADHD. Confounding variables make it difficult or impossible to determine what, if anything, caused the increase in ADHD symptoms, because the researchers cannot distinguish between the effect of digital media usage and the confounding variables on ADHD symptoms.

People who eat lots of fruits and vegetables have lower rates of colon cancer than those who eat little of these foods. Fruits and vegetables are rich in "antioxidants" such as vitamins A, C, and E. Will taking antioxidants help prevent colon cancer? A clinical trial studied this question with 864 people who were at risk for colon cancer. The subjects were divided into four groups: daily beta-carotene, daily vitamins C and E, all three vitamins every day, and daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. Outline the design of the experiment. (The diagrams in Figures 5.2 and 5.3 are models.) Please place the groups and their treatments in ascending order.

Random Assignment ---> groups 1-4 of subjects ----> groups of treatment -----> Observe colon cancer Extended Answer: The goal of the clinical trial was to determine if antioxidants would help prevent colon cancer. The researchers divided 864 subjects into four groups. The first group received daily beta-carotene. The second group received vitamins C and E daily. The third group received beta-carotene, vitamin C and vitamin E every day. The fourth group received a daily placebo. After four years, the researchers counted and compared the number of colon cancer cases in each group.

Does a lower pitch of a voice in an ad lead consumers to envision a bigger product? To test this, researchers had students listen to a radio advertisement for the new Southwest Turkey Club Sandwich at a ficititious sandwich chain, Cosmo. Half of the students were randomly assigned to hear the ad spoken at a high pitch and the other half at a low pitch. In all other respects, the ads were identical and no clues were given as to the size of the sandwich. After hearing the ad, students were asked to rate the perceived size of the sandwich on a 7 point scale, randing from −3 (much smaller than average) to +3 (much larger than average). Sketch a randomized comparative experimental design for this experiment. The design should show where randomization occurs, the groups, the treatment each group recieves, and how the outcomes are analyzed. Not all descriptions will be used.

Random assignment ---> groups of students ---> treatments ---> compare ratings of perceived size of the sandwich Extended Answer: One treatment is hearing the ad spoken in a low‑pitched voice and the other treatment is hearing the ad spoken in a high‑pitched voice. It does not matter which is called treatment 1 and which is called treatment 2. We will arbitrarily refer to the high‑pitched ad as treatment 1 and the low‑pitched ad as treatment 2. The first step is to randomly assign half of the subjects (the students) to one group. The remaining students are thereby randomly assigned to the other group. Group 1 receives treatment 1 and group 2 receives treatment 2. The final step is to compare the values of the response variable and see if the ratings of the perceived size of the sandwich depends on which group the subject was assigned to.

Researchers identified 2587 10th grade students in the Los Angeles area who did not have significant symptoms of ADHD (Attention-Deficit/Hyperactivity Disorder). These students were followed for approximately two years. The frequency with which each student used digital media over the two‑year period was recorded. At the end of the two‑year period, students were again tested for symptoms of ADHD. Researchers found that the higher the frequency of digital media use, the more likely the student was to have developed symptoms of ADHD at the end of the study. What are the explanatory and response variables?

Response: ADHD Explanatory: digital media usage Extended Answer: Researchers observed that a higher frequency of digital media usage over a two year period was related to increased symptoms of ADHD in a population that initially had no significant symptoms of ADHD. Since the digital media usage is being viewed as a possible cause of the ADHD symptoms, you should consider this to be the explanatory variable, and because the interest is in how much of a change there is in the ADHD symptoms with a higher exposure of digital media, you should consider ADHD to be the response variable. The 2587 10th grade students are the sample. The two‑year period was the duration of the observational study and the lack of ADHD symptoms at the beginning of the study was the initial condition of the subjects.

Can aspirin help prevent heart attacks? The Physicians' Health Study, a large medical experiment involving 22,000 male physicians, attempted to answer this question. One group of about 11,000 physicians took an aspirin every second day, while the rest took a placebo. After several years, the study found that subjects in the aspirin group had significantly fewer heart attacks than subjects in the placebo group. Identify the experimental subjects, the explanatory variable and the values it can take, and the response variable.

Subject: 22000 Male Physicians Explanatory Variable: Medication Response Variable: Health Values of Explanatory Variable: Aspirin, Placebo Values of Response Variable: Heart attack, no heart attack Extended Answer: The goal of the study was to determine if aspirin lowers the chance of a heart attack. The researchers gave 22,000 males a medication. Those males were the subjects. The researchers expected the medication to cause an effect in the subjects. The medication was therefore the explanatory variable. The researchers gave subjects either aspirin or a placebo. Since the explanatory variable is medication, aspirin and placebo are values of the explanatory variable. The researchers observed the subjects for several years after treatment. The researchers tracked the health of the subjects. Since the researchers wanted to know if the health of the subjects changed in response to the treatment, health is the response variable of the study. Researchers measured health by recording information about a subject's heart attacks. The subject either had a heart attack or did not have a heart attack. These two options represent the values of the response variable.

People who eat lots of fruits and vegetables have lower rates of colon cancer than those who eat little of these foods. Fruits and vegetables are rich in "antioxidants" such as vitamins A, C, and E. Will taking antioxidants help prevent colon cancer? A clinical trial studied this question with 864 people who were at risk for colon cancer. The subjects were divided into four groups: daily beta-carotene, daily vitamins C and E, all three vitamins every day, and daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. Assign labels to the 864 subjects. Use Table A, starting at line 127, to choose the first five subjects for the beta‑carotene group. Enter the labels for the first five subjects in order. Use a single space between the labels. If you wish, you may also use commas between the labels.

Subjects: 439 099 725 330 643 Extended Answer: To use a table of random digits, first label the subjects. There are 864 subjects in the experiment, so three digits are required per subject. Assign 001 to the first one, 002 to the second, and continue until all 864 subjects have a number. The digits in line 127 are 43909 99477 25330 64359 40085 16925 85117 36071 The spaces do not mean anything. They are just there to make the table easier to read. Separating the digits into groups of three gives 439 099 947 725 330 643 594 008 516 925 851 173 607 1 These are the labels from which you will chose the sample. Skip groups of three numbers that are greater than 864, skip 000, and if you were to encounter a group of three digits that you already encountered, skip them as well. Unless the population size is less than 10, you will need to use leading zeros. Otherwise you won't have a simple random sample.

A study, mandated by Congress when it passed No Child Left Behind in 2002, evaluated 15 reading and math software products used by 9424 students in 132 schools across the country during the 2004‑2005 school year. It is the largest study that has compared students who received the technology with those who did not, as measured by their scores on standardized tests. There were no statistically significant differences between students who used software and those who did not. Choose the meaning of "no statistically significant differences" in plain language.

The difference in scores for students who used software and those who did not is so small that it is likely to have occurred due to chance alone. Extended Answer: An observed effect of a size that would rarely occur by chance is called statistically significant. According to the definition, the fact that there were no statistically significant differences between students who used software and those who did not means that the observed difference in standardized test scores between the two groups is small and could plausibly be explained by chance alone. If the difference between two groups of students was large, it would be unlikely due to chance and there would be a statistically significant difference in scores between students who used software and those who did not. There is no any information about the mean scores for both groups. Thus, the fact that there were no statistically significant differences between two groups of students does not mean that the students who used software and those who did not have the same mean scores on the standardized tests. Flaws in an experimental design may indeed make it impossible to generalize the results to a larger population. However, this is not the meaning of no statistically significant difference.

Does regular exercise reduce bone loss in postmenopausal women? Here are two ways to study this question. Which design will produce more trustworthy data? A researcher finds 1000 postmenopausal women who exercise regularly. She matches each with a similar postmenopausal woman who does not exercise regularly, and she follows both groups for five years. Another researcher finds 2000 postmenopausal women who are willing to participate in a study. She assigns 1000 of the women to a regular program of supervised exercise. The other 1000 continue their usual habits. The researcher follows both groups for five years. Select all statements that are true.

The effect of a lurking variable, such as a condition that might affect someone's desire to exercise and their bone loss, is likely to be larger in 1 and less in 2 due to randomization. 1 is an observational study and 2 is an experiment. The effect of a lurking variable, such as a condition that might affect someone's desire to exercise and their bone loss, is likely to be larger in 2 and less in 1 due to randomization. Extended Answer: According to the definition, an observational study observes individuals and measures variables of interest but does not intervene in order to influence the responses. An experiment deliberately imposes some treatment on individuals in order to observe their responses. In the first case, the researcher observes both groups of postmenopausal women. The researcher does not intervene in the process, both samples can be created early. It is an observational study. The researcher in the second study finds a big group of postmenopausal women and randomly assigns each individual to the treatment. It is an experiment because the researcher imposes some treatment on individuals and intervenes the process. So, 1 is an observational study and 2 is an experiment. In the first study, the sample is self‑selected and so, the results of this study can be bias. There are lurking variables which can influence the study. They can be health status, age, social status, marital status, etc. Both samples in the second study are randomly selected. If there are external factors, the randomization of the samples helps to prevent their influence on the results. Therefore, the effect of a lurking variable, such as a condition that might affect someone's desire to exercise and their bone loss, is likely to be larger in 1 and less in 2 due to randomization. If the women who do regular exercise reduce a bone loss, it is not possible to say that it is caused by regular exercise. It might be just the placebo effect. Many patients respond favorably to any treatment, even a placebo. Thus, the researcher could divide the women between two groups to reduce the placebo effect. The first group will do regular exercise. The women in the second group will continue their usual habits receiving a placebo. Thus, a placebo could be practical in 2.

People who eat lots of fruits and vegetables have lower rates of colon cancer than those who eat little of these foods. Fruits and vegetables are rich in "antioxidants" such as vitamins A, C, and E. Will taking antioxidants help prevent colon cancer? A clinical trial studied this question with 864 people who were at risk for colon cancer. The subjects were divided into four groups: daily beta-carotene, daily vitamins C and E, all three vitamins every day, and daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. What does "no significant difference" mean in describing the outcome of the study?

The observed differences were no more than what might reasonably occur by chance. Extended Answer: The result of a study is statistically significant if the studied effect is so large that it would rarely occur by chance. The researchers found no significant difference between the amounts of colon cancer in the treatment groups. There were almost certainly different numbers of subjects that developed cancer in each group, but these observed differences were smaller or equal to what you could expect due to random chance.

Does a lower pitch of a voice in an ad lead consumers to envision a bigger product? To test this, researchers had students listen to a radio advertisement for the new Southwest Turkey Club Sandwich at a ficititious sandwich chain, Cosmo. Half of the students were randomly assigned to hear the ad spoken at a high pitch and the other half at a low pitch. In all other respects, the ads were identical and no clues were given as to the size of the sandwich. After hearing the ad, students were asked to rate the perceived size of the sandwich on a 7 point scale, randing from −3 (much smaller than average) to +3 (much larger than average). Could the researchers have used a placebo in this experiment? Explain.

The researchers could have assigned students to watch an ad with no audio, thus controlling for pitch. Extended Answer: A placebo treatment is a dummy treatment that is not designed to cause a specific effect in the subjects. Researchers use a placebo group to compare to treatment groups. Because the researchers were testing to see if the pitch of a voice had an effect on the perceived size of the sandwich, it is conceivable that the researchers could have assigned students to watch an advertisement with no audio. Removing the audio removes the potential effect of the pitch of the voice of the narrator. If researchers use a numbering system to prevent themselves from knowing the assignment of the treatments, the researchers are blinding themselves from knowing the treatment. Blinding may reduce bias, but blinding is not a treatment and is therefore not a placebo. The subjects could listen to both advertisements. However, the researchers assigned the advertisements to cause an effect in the subjects. Having the subjects listen to both advertisements is therefore not a placebo. The researchers could redesign the experiment so that each subject is listening with headphones. However, the assignment of headphones is a part of the design of the experiment and is not a form of a treatment. Giving headphones to the subjects might help to eliminate potential biases but the headphones do not represent a placebo treatment.

Researchers identified 2587 10th grade students in the Los Angeles area who did not have significant symptoms of ADHD (Attention-Deficit/Hyperactivity Disorder). These students were followed for approximately two years. The frequency with which each student used digital media over the two‑year period was recorded. At the end of the two‑year period, students were again tested for symptoms of ADHD. Researchers found that the higher the frequency of digital media use, the more likely the student was to have developed symptoms of ADHD at the end of the study. Explain carefully why this study is not an experiment.

The study is not an experiment because the researchers did not assign digital media usage as a treatment to measure a difference of effects. Extended Answer: The study is not an experiment because the researchers did not intervene by imposing a treatment (digital media usage) in order to see what would happen. Additionally, the study observed 2587 10th grade students in the Los Angeles area who did not have significant symptoms of ADHD over a two year period, making it an observational study of symptom free students, not an experiment. Students were neither chosen at random for the sample nor randomly assigned to treatment groups for comparison. Randomization produces groups of subjects that should be similar, on average, in all respects before treatments are applied. The treatment groups also need to be compared to some kind of control group, whether it is a placebo or other type of common exposure such as the current standard treatment. The ADHD symptoms and digital media usage are confounded because additional lurking variables, such as poor health or parental divorce, were not controlled. Comparative design exposes all groups to similar conditions, other than the treatments they receive. This ensures that any additional lurking variables operate equally on all groups and, on average, groups differ only in the treatments they receive. Therefore, differences in the response variable must be due to the effects of the treatments indicating a cause and effect relationship.

Does church attendance lengthen people's lives? One study of the effect of attendance at religious services gathered data from 2001 obituaries. The researchers measured whether the obituaries mentioned religious activities and length of life. The study is

an observational study. Extended Answer: To determine the type of the described study, use the following definitions. An observational study observes individuals and measures variables of interest but does not intervene in order to influence the responses. The purpose of an observational study is to describe some group or situation. A sample survey is a kind of observational study which chooses a group of individuals from a specific population and uses the group to get information about the entire population. An experiment deliberately imposes some treatment on individuals in order to observe their responses. The purpose of an experiment is to study whether the treatment causes a change in the response. A randomized comparative experiment compares two or more treatments, uses chance to decide which subjects get each treatment, and uses enough subjects so that the effects of chance are small. In this study, the effect of religious activities was measured, but no treatment was actively imposed in order to observe the response. Thus, the study cannot be any kind of experiment. A sample survey is a kind of observational study. Thus, the option "neither an experiment nor an observational study but, instead, a sample survey" does not make sense. Note that there is too little information about the choice of obituaries to understand whether this study can be a sample survey or not. The effect that was caused by church attendance on longevity was investigated without any interference of the researchers. Therefore, the given study is an observational study.

A study closely followed a large social network of 12067 people for 32 years, from 1971 until 2003. The researchers found that when a person gains weight, close friends tend to gain weight, too. The researchers reported that obesity can spread from person to person, much like a virus. Explain why the fact that when a person gains weight, close friends also tend to gain weight does not necessarily mean that weight gains in a person cause weight gains in close friends. In particular, identify some lurking variables whose effect on weight gain may be confounded with the effect of weight gains in close friends. A likely lurking variable is

shared activities (drinking, going out to eat, lack of exercise). Extended Answer: Because close friends have many shared activities, similar lifestyle can have a similar effect on every person in the group. Poor habits such as drinking, eating out or lack of exercise can cause weight gain on any given person and on close friends who share these activities with the person. Thus, shared activities can be lurking variables in the described research. Location does not have a direct influence on a group of friends' weight gain so it cannot be a lurking variable in the given situation. Since there is no empirical evidence suggesting that one's gender is a strong factor in weight gain, it cannot be a lurking variable. The placebo effect is a phenomenon in which some people experience a benefit after the administration of a treatment with no known positive effects. This concept is not applicable to the described research at all.

Does church attendance lengthen people's lives? One study of the effect of attendance at religious services gathered data from 20012001 obituaries. The researchers measured whether the obituaries mentioned religious activities and length of life. People who are active in religious activities are less likely to smoke or drink excessively than people who are not active in religious activities. In the described study

smoking and excessive drinking can be both lurking variables. Extended Answer: To determine the lurking variables, use the following definition. A lurking variable is a variable that has an important effect on the relationship among the variables in a study but is not one of the explanatory variables studied. Both smoking and excessive drinking, which were not mentioned in the study, can be involved in cause‑and‑effect relationships between the explanatory variable and the response variable. Because people who are engaged in religious activities are less likely to smoke or drink excessively, and smoking or drinking excessively can negatively affect one's length of life, they can be lurking variables of the study.

Does a lower pitch of a voice in an ad lead consumers to envision a bigger product? To test this, researchers had students listen to a radio advertisement for the new Southwest Turkey Club Sandwich at a ficititious sandwich chain, Cosmo. Half of the students were randomly assigned to hear the ad spoken at a high pitch and the other half at a low pitch. In all other respects, the ads were identical and no clues were given as to the size of the sandwich. After hearing the ad, students were asked to rate the perceived size of the sandwich on a 7 point scale, randing from −3 (much smaller than average) to +3 (much larger than average). What is the explanatory variable?

the pitch of a voice in an ad (high versus low) Extended Answer: Researchers were testing to see if the lower pitch of a voice in an ad led to consumers envisioning a bigger product. The researchers expected the pitch of the voice to cause a change in the perceived size of the product. Because the researchers expected the pitch to cause a change in the outcome variable, the pitch of a voice in the ad is the explanatory variable. Because the Southwest Turkey Club Sandwich itself was not changing and was therefore not causing a change in the study, the sandwich was not a variable. The rating of the size of the sandwich was the effect that researchers suspect might be due to the pitch of the voice reading the ad. The effect is a study is not the explanatory variable. The students in the sample are the subjects of the experiment. The explanatory variable was measured in the subjects, so the subjects cannot be the explantory variable.

Does a lower pitch of a voice in an ad lead consumers to envision a bigger product? To test this, researchers had students listen to a radio advertisement for the new Southwest Turkey Club Sandwich at a ficititious sandwich chain, Cosmo. Half of the students were randomly assigned to hear the ad spoken at a high pitch and the other half at a low pitch. In all other respects, the ads were identical and no clues were given as to the size of the sandwich. After hearing the ad, students were asked to rate the perceived size of the sandwich on a 7 point scale, randing from −3 (much smaller than average) to +3 (much larger than average). What is the response variable, and what values does it take?

the rating of the perceived size of the sandwich; point values for the size could range from - 3 to + 3 Extended Answer: Researchers tested the hypothesis that a lower‑pitch voice in an ad led consumers to envision a bigger product. The researchers used the rating of the perceived size of the sandwich to measure the outcome of the study. Therefore, the rating is the response variable. The values it takes are the point values for the size, which range from −3 to 3. The design of the experiment did not measure the effect of the change in pitch. The researchers assigned the change in the pitch. The design of an experiment may effect a response variable but the design does not measure the outcome variable. The pitch, low or high, was a treatment that the researchers assigned to the subjects. Therefore, the pitch in the ad was the explanatory variable. The students in the sample may be assigned to the high pitch or the low pitch group. However, the assignment of groups does not measure the perceived size of the sandwich.

Does church attendance lengthen people's lives? One study of the effect of attendance at religious services gathered data from 2001 obituaries. The researchers measured whether the obituaries mentioned religious activities and length of life. In the described study, researchers found that there was a statistically significant difference in longevity between those whose obituary mentioned religious activities and those whose obituary did not. Those whose obituary mentioned religious activities lived more than 5 years longer. Statistical significance here means

the size of the observed difference in longevity is not likely to be due to chance. Extended Answer: Differences among the effects of the treatments so large that they would rarely occur by chance is called statistically significant. According to the definition of the statistically significant result, the observed difference in longevity is most likely to happen due to the effect of the church attendance. Thus, the term does not mean that the obtained difference is likely to be due to chance. The study demonstrated that the observed difference in the length of life between people whose obituary mentioned religious activities and people whose obituary did not mention religious activities is substantial, but this information has no connection with probability of occurrence of the observed difference. The only meaning of the term "statistically significant" that matches the definition is that the size of the observed difference in length of life is very unlikely to be due to chance.