Biology 101G Ch.1 The Process of Science

You hear a news report about a new asthma treatment. What would you want to know before you asked your doctor if this treatment was right for you?

1. Was the drug tested in a randomized clinical trial? 2. How many participants were in the trial? 3. Was there a significant difference between the effect of the new drug and the treatment used in the control group? 4. Did any of the researchers have financial ties to the manufacturer of the new asthma drug?

Most statistical tests report a p value that determines whether or not the results are statistically significant (i.e., not produced by chance). Usually the cutoff for p values is 0.05: if the p value is less than 0.05 the results are considered to be statistically significant. The graph below shows data from a 2012 study published in the New England Journal of Medicine. The study examined the impact of the drug Tofacitinib on ulcerative colitis. From the data shown, what dose(s) of Tofacitinib is/are significantly better than the placebo in treating ulcerative colitis?

15 mg

Experiment

A carefully designed test, the results of which will either support or rule out a hypothesis.

Correlation

A consistent relationship between two variables.

Randomized Clinical Trial

A controlled medical experiment in which subjects are randomly chosen to receive either an experimental treatment or a standard treatment (or a placebo).

Placebo

A fake treatment given to control groups to mimic the experience of the experimental groups.

Testable

A hypothesis is testable if it can be supported or rejected by carefully designed experiments or observational studies.

Ststistical Significance

A measure of confidence that the results obtained are "real" and not due to chance.

Peer Review

A process in which independent scientific experts read scientific studies before they are published to ensure that the authors have appropriately designed and interpreted the study.

Hypothesis

A tentative explanation for a scientific observation or question.

Scientific Theory

An explanation of the natural world that is supported by a large body of evidence and has never been disproved.

Anecdotal Evidence

An informal observation that has not been systematically tested.

Faslifiable

Describes a hypothesis that can be ruled out by data that show that the hypothesis does not explain the observation.

ou carry out a clinical trial to test whether a new drug relieves the symptoms of arthritis better than a placebo. You have four groups of participants, all of whom have mildly painful arthritis (rated 7 on a scale of 1 to 10). Each group receives a daily pill as follows: control (group 1) - placebo; group 2- 15 mg; group 3- 25 mg; group 4- 50 mg. At the end of 2 weeks, participants in each group are asked to rate their pain on a scale of 1 to 10. The mean pain rating of the participants was 6.5 for the placebo, 6.0 for 15 mg of the drug, 4.5 for 25 mg of the drug, and 4.5 for 50 mg of the drug. What is your next step?

Run a statistical analysis to determine if the differences are significant.

Control Group

The group in an experiment that experiences no experimental intervention or manipulation.

Experimental Group

The group in an experiment that experiences the experimental intervention or manipulation.

Dependent Variable

The measured result of an experiment, analyzed in both the experimental and control groups.

Sample Size

The number of experimental subjects or the number of times an experiment is repeated. In human studies, sample size is the number of participants.

Science

The process of using observations and experiments to draw conclusions based on evidence.

Epidemiology

The study of patterns of disease in populations, including risk factors.

Looking at Infographic 1.4 (Sample Size Matters), you see that both graphs show a positive impact of caffeine on memory. However, the data in the graph on the right carry more weight. Why is that? If you read a study that reported only the data in the left graph, would you find the relationship to be compelling? Why or why not?

The study plotted in the graph on the left has far fewer participants. Because of the smaller sample size, the data are not as convincing as those plotted in the right panel, with a larger sample size. Larger sample sizes give more conclusive results.

Independent Vaiable

The variable, or factor, being deliberately changed in the experimental group.

You are working on an experiment to test the effect of a specific drug on reducing the risk of breast cancer in postmenopausal women. Describe your control and experimental groups with respect to age, gender, and breast cancer status.

There are many possibilities. Given the parameters, both the experimental and control groups would have to be postmenopausal females. The control group would not receive the drug; the experimental group would receive the drug (the drug is the independent variable). The main dependent variable is breast cancer.

Design a randomized clinical trial to test the effects of caffeinated coffee on brain activity. Design your study so that the results will apply to as many people in as many scenarios as possible.

There are many possibilities. The independent variable will be caffeinated coffee. The dependent variable can be any measure of brain activity (e.g., memory, alertness, problem solving). The basic design will involve two groups: caffeinated-coffee drinkers and non-caffeinated-coffee drinkers. To have the results be broadly applicable, the participants should include both sexes and a variety of ages, ethnicities, and racial groups.

Following the prompts below, design a clinical trial to test the impact of a particular intervention on a specific aspect of human health. You will need to use everything you have learned to do this. a. From scientific articles or press releases from health organizations you have read, or from your own experiences, what observation(s) can you start with? b. Do some investigation to generate a testable hypothesis. c. Design the trial. Consider sample size, whether or not you will use a placebo, what level of blinding you will use, and possible independent and dependent variables.

There are many possible options. The key points are to make observations, consult sources to develop a hypothesis and design a strong trial, including controls, to test the hypothesis.

What is the importance of statistical analyses?

They can be used to support or reject the hypothesis. They can be used to determine whether any observed differences between two groups are real or a result of chance.

How can two different studies investigating the same thing (e.g., the relationship, if any, between caffeinated coffee and memory) come to different conclusions?

They may have had different sample sizes. They may have used different types of participants (e.g., participants of different ages or professions). They may have used different amounts of caffeine. They may have evaluated memory differently (e.g., long-term vs. short-term memory).

The mother of a friend is a self-described "coffee addict." She recently received a diagnosis of Parkinson disease. Does her experience negate the results of the JAMA study described in this chapter? Why or why not?

This is a case of anecdotal evidence, presenting a single person's experience in the absence of a carefully designed experiment and a large sample size. As such, this person's experience does not negate the results of the study published in JAMA.

In which of the following would you have the most confidence? a. a randomized clinical trial with 15,000 subjects b. a randomized clinical trial with 5,000 subjects c. an epidemiological study with 15,000 subjects d. an endorsement of a product by a movie star

a. a randomized clinicla trail with 15,000 subjects

A scientist who reads an article in a scientific or medical journal can be confident that the report has been peer reviewed. a. What is a "peer-reviewed" report? Is an article in a daily newspaper a peer-reviewed article? b. What is the role of a peer reviewer of a scientific article? c. Why do scientists place so much value on the peer-review process?

a: A "peer-reviewed" scientific report is one that has been read and evaluated by other scientists. A typical newspaper has not been peer reviewed. b: A peer reviewer considers the experimental design, the data, and the analysis of the data. The reviewer's job is to ensure that the conclusions do not overreach the data and that the study is well designed and without bias. c: The peer-review process helps to ensure the integrity and quality of scientific publications, so that they can inform the scientific process moving forward.

There are many misconceptions about breast cancer and its causes. In the late 1990s, there were rumors that antiperspirants cause breast cancer. There are still catalogues that offer alternative underarm hygiene products that purport to reduce the risk of breast cancer. One viral e-mail claimed that by blocking perspiration, antiperspirants prevent the body from purging toxins, instead forcing the body to store the toxins in lymph nodes in the underarm area near breast tissue. The e-mail stated that men were less likely to develop breast cancer from antiperspirants because their underarm hair trapped most of the product away from direct contact with skin. And as men are less likely to shave their underarms, they are less likely to have shaving nicks through which antiperspirants can enter the body. a. Read the abstracts of the two articles for which links are provided below. Darbre, 2005 http://is.gd/pPLxwZ Harvey and Everett, 2004 http://is.gd/ycqDD8 From the abstracts, and from any other investigation you do, name the components of underarm deodorants and antiperspirants that have been identified as possible culprits in causing breast cancer. b. Briefly comment on the strengths and weaknesses of each study (consider sample size, control groups, and overall study design). c. From what you read in the abstracts and from other research you do (cite any additional reliable sources that you consulted), do you think that use of antiperspirants or deodorants or both is a consistent risk factor for breast cancer? Has your opinion about underarm hygiene changed? Explain how and why your opinion has either changed or remained consistent, referring to the abstracts that you have reviewed.

a: The Darbre abstract points to aluminum, while the Harvey and Everett abstract points to parabens. b: Darbre implicates aluminum on the basis of finding breast cancer in areas of the breast near where antiperspirants are applied. There is no indication of sample size, and as this is simply correlation, it does not suggest causation. The paraben study looks at levels of parabens in breast tumors from 20 patients. This is a small sample size, and there does not appear to be an investigation of paraben levels in normal (control) breast tissue. c: Answers will vary, but should be supported by evidence.

You may have seen advertisements on television that show beautiful people with clear skin who claim that a specific skin care product is "scientifically proven" to reduce acne. The product reportedly gave these people glowing, clear skin. a. Is their testimony alone strong enough evidence for you to act on? Why or why not? b. What kind of scientific evidence would persuade you to spend money on this product? Explain your answer.

a: The testimonials of these few people are not convincing. They represent a small sample size, and we don't know if their evidence is biased. There is no strong experimental trial. b: A large, randomized trial to compare outcomes in people who use the product and people who do not would provide more convincing reasons to use the product, particularly if the dependent variables specifically included number of pimples, degree of redness, and other objective measures of skin clarity.

Of the following, which is the earliest step in the scientific process? a. generating a hypothesis b. analyzing data c. conducting an experiment d. drawing a conclusion e. asking a question about an observation

asking a question about an observation

In the studies of coffee and memory discussed, the independent variable is _____________ and the dependent variable is _____________.

caffeine;memory

When scientists carry out an experiment, they are testing a

hypothesis

From what you have read in this chapter, would you say a 21-year-old Caucasian female can count on caffeinated coffee to reduce her risk of Parkinson disease?

no, because participants in that peer-reviewed study were Japanese-American males; it cannot be inferred that the same results would hold for Caucasian females

Can an epidemiologist who finds a correlation between the use of tanning beds and melanoma (an aggressive form of skin cancer) in college-age women conclude that tanning beds cause skin cancer?

no; correlation is not proof of causation

You are listening to a news report that claims a new study has found convincing evidence that a particular weight-loss product is much more effective than diet and exercise. What can you infer about "convincing" evidence in this case?

that statistical tests showed significantly more weight loss in the participants who used the weight-loss product than those who relied on diet and exercise.

In a controlled experiment, which group receives the placebo?

the control group

You carry out a clinical trial to test whether a new drug relieves the symptoms of arthritis better than a placebo. You have four groups of participants, all of whom have mildly painful arthritis (rated 7 on a scale of 1 to 10). Each group receives a daily pill as follows: control (group 1) - placebo; group 2- 15 mg; group 3- 25 mg; group 4- 50 mg. At the end of 2 weeks, participants in each group are asked to rate their pain on a scale of 1 to 10. What is the independent variable in this experiment?

the drug