College Statistics Quiz: Observational Studies Vs Experiments
20 questions · exam conditions
0:00
Observational Studies Vs ExperimentsQuestion 1 of 20

Researchers want to investigate the long-term effects of regular, vigorous exercise on cognitive function in old age. Which of the following is the most compelling reason for using a prospective observational study design rather than an experiment?

It is impractical and likely unethical to randomly assign individuals to a 'no exercise' group for several decades.
Observational studies are less expensive and can accommodate a much larger sample size than experiments.
It is difficult to accurately measure the response variable, cognitive function, in an experimental setting.
A retrospective study would be more effective for establishing a causal relationship between exercise and cognition.
← Back to quizzes

College Statistics Quiz

College Statistics Quiz: Observational Studies Vs Experiments

Practice Observational Studies Vs Experiments in College Statistics with focused quiz questions that help you check what you know, review explanations, and build confidence with test-style prompts.

What this quiz covers

This quiz focuses on Observational Studies Vs Experiments, giving you a quick way to practice the rules, question types, and explanations that matter most for College Statistics.

How to use this quiz

Try each quiz question before looking at the correct answer. Use the explanations to review missed ideas, then come back to similar questions until the pattern feels familiar.

All questions

Question 1

Researchers want to investigate the long-term effects of regular, vigorous exercise on cognitive function in old age. Which of the following is the most compelling reason for using a prospective observational study design rather than an experiment?

  1. It is impractical and likely unethical to randomly assign individuals to a 'no exercise' group for several decades. (correct answer)
  2. Observational studies are less expensive and can accommodate a much larger sample size than experiments.
  3. It is difficult to accurately measure the response variable, cognitive function, in an experimental setting.
  4. A retrospective study would be more effective for establishing a causal relationship between exercise and cognition.
Explanation: The correct answer is A. While other factors might be considered, the primary reason an experiment is not feasible is the ethical and practical challenge of enforcing a randomly assigned treatment (like vigorous exercise or no exercise) over a very long period, such as decades. Forcing a control group to be sedentary would be unethical, and ensuring compliance in the treatment group would be nearly impossible. (B) is a practical consideration but not the most important reason. Even with unlimited funds, the design described in (A) would be a major barrier. (C) is incorrect because measuring cognitive function is a challenge for both observational studies and experiments; it's not a reason to prefer one design over the other. (D) is incorrect because a retrospective study would be even weaker for establishing causation than a prospective study, as it would likely suffer from recall bias and poorer quality data on past exercise habits.

Question 2

A researcher recruited 100 volunteers with moderate anxiety from a university campus. She randomly assigned 50 volunteers to a new mindfulness meditation program and 50 to a waitlist control group. At the end of eight weeks, the meditation group showed a statistically significant reduction in anxiety scores compared to the control group. What is the most appropriate scope of inference from this study?

  1. The results can be generalized to the entire population of adults with anxiety, and the study provides evidence that the meditation program causes a reduction in anxiety.
  2. The results can be generalized to the population of students at that university, but it only shows an association, not a causal link, between meditation and reduced anxiety.
  3. The study provides evidence that the meditation program causes a reduction in anxiety for individuals like the volunteers, but the results may not generalize to all adults with anxiety. (correct answer)
  4. Neither generalization to a larger population nor a causal conclusion is warranted because the participants were volunteers.
Explanation: The correct answer is C. The study used random assignment, which allows for causal inference (the program caused the reduction in anxiety). However, the participants were volunteers from a single university campus, not a random sample of the entire population of adults with anxiety. Therefore, the results are most reliably applicable to individuals similar to those in the study, and generalization to a broader population is not warranted. (A) is incorrect because while a causal conclusion is appropriate, generalization to all adults with anxiety is not, due to the non-random sample. (B) is incorrect because it mistakenly claims a causal link cannot be established. Random assignment is the key to establishing causation. (D) is incorrect because causal conclusions are warranted due to random assignment. The use of volunteers limits generalizability but does not invalidate the causal link for the group studied.

Question 3

A teacher wants to determine if a new educational app improves student performance. In her class of 30 students, she tells them about the app and makes it available for them to use. At the end of the unit, she compares the test scores of the 10 students who used the app most frequently with the 10 students who used it least. Why is the teacher's conclusion that the app improved scores for the high-use group likely to be flawed?

  1. The study lacks a proper control group since all students were in the same class and had access to the app.
  2. The sample size of 30 students is not large enough to establish a statistically significant finding.
  3. This is an observational study because students were not randomly assigned to use or not use the app. (correct answer)
  4. The teacher's grading of the test could have been biased if she knew which students used the app.
Explanation: The correct answer is C. The fundamental flaw is that this is an observational study, not an experiment. Students self-selected into the 'high-use' and 'low-use' groups. It is highly plausible that the students who chose to use the app more were already more motivated, organized, or had higher prior ability. These confounding factors, rather than the app itself, could be the real reason for their higher test scores. (A) is related but less precise. Comparison groups were formed, but the method of forming them (self-selection) is the critical flaw. (B) addresses statistical power, not the underlying logical flaw in the study's design. A flawed design cannot be fixed by a larger sample size. (D) identifies a potential source of bias, but the confounding from self-selection is a more fundamental problem with the study's design for drawing a causal conclusion.

Question 4

A nutritionist compares the cardiovascular health of a group of people who have chosen to follow a ketogenic diet with a group who have chosen a standard diet. To minimize confounding, she carefully matches each individual in the keto group with an individual in the standard diet group based on age, sex, exercise level, and smoking status. Even with this matching, why can this study NOT establish that the ketogenic diet causes any observed differences in health outcomes?

  1. The sample was not randomly selected from the entire population, so the results cannot be generalized.
  2. There may be unmeasured systematic differences between people who voluntarily choose different diets. (correct answer)
  3. The study lacks a true control group because the standard diet group is not receiving a placebo.
  4. Matching variables after the fact is not a valid statistical technique for reducing bias in a study.
Explanation: The correct answer is B. Despite matching on several key variables, this is still an observational study because participants chose their own diets. There could be other systematic, unmeasured differences (confounding variables) between the two groups. For example, people who choose a ketogenic diet might also be more likely to engage in other health-conscious behaviors (e.g., supplement use, stress management) that are not accounted for by the matching variables. This prevents a firm conclusion about causation. (A) is incorrect because the issue of generalizability (due to lack of random sampling) is separate from the issue of establishing causation (which requires random assignment). (C) is incorrect because the standard diet group serves as a valid comparison group; the inability to use a placebo does not negate this, but the lack of random assignment is the critical flaw for causal inference. (D) is incorrect because matching is a valid and commonly used technique to reduce bias in observational studies, although it cannot eliminate it entirely.

Question 5

A researcher obtains a list of all 4,000 patients at a medical clinic who have Type 2 diabetes. She randomly selects 80 patients from this list to participate in a study. She then randomly assigns 40 of them to a new dietary program and 40 to continue with the standard recommendations. Based on this design, which of the following is the most appropriate scope of inference if the new program shows a significant benefit?

  1. The results can be generalized to all people with Type 2 diabetes, and a causal conclusion can be drawn.
  2. A causal conclusion can be drawn, but the results cannot be generalized beyond the 80 participants.
  3. The results can be generalized to all patients with Type 2 diabetes at that clinic, and a causal conclusion can be drawn. (correct answer)
  4. The results can be generalized to all patients with Type 2 diabetes at that clinic, but a causal conclusion cannot be drawn.
Explanation: The correct answer is C. This study design has two key features. First, the use of random assignment of patients to the dietary programs allows for a causal conclusion (i.e., that the new program caused the observed benefit). Second, the use of random sampling from the population of all 4,000 diabetic patients at the clinic allows the results to be generalized to that specific population. (A) is incorrect because the random sample was only from one clinic, so generalizing to all people with Type 2 diabetes is not warranted. (B) is incorrect because the random sampling allows for generalization to the clinic's population. (D) is incorrect because the random assignment allows for a causal conclusion.

Question 6

A horticulturist wants to test a new fertilizer's effect on the yield of tomato plants. She has 20 plants on the sunny east side of a greenhouse and 20 plants on the shady west side. She suspects that the amount of sunlight will affect yield and wants to account for this. Which of the following is the most appropriate experimental design?

  1. A completely randomized design: combine all 40 plants and randomly assign 20 to receive the fertilizer.
  2. An observational study: apply the fertilizer to all east-side plants and compare them to the unfertilized west-side plants.
  3. A randomized block design: within the east-side plants, randomly assign 10 to the fertilizer, and separately, do the same for the west-side plants. (correct answer)
  4. A matched-pairs design: create 20 pairs, with each pair consisting of one east-side and one west-side plant.
Explanation: The correct answer is C. This describes a randomized block design. The horticulturist has identified a potential confounding variable (sunlight). By creating 'blocks' (east side and west side) and then randomizing the treatment within each block, she can isolate the effect of the fertilizer from the effect of the sunlight. This design ensures that both the treatment and control groups have equal representation from the sunny and shady locations. (A) is not ideal because a completely randomized design might, by chance, assign more sunny plants to one group, confounding the results. (B) is a poor design because it confounds the effect of the fertilizer with the effect of sunlight, making it impossible to separate them. (D) is incorrect because matched pairs should consist of two experimental units that are as similar as possible. Pairing a sunny and a shady plant violates this principle.

Question 7

To evaluate a new online workplace wellness program, a company offers it on a voluntary basis to all employees. At the end of the year, analysts compare the average number of sick days for employees who participated in the program with those who did not. Why is this study not considered a true experiment?

  1. The company did not use a placebo for the non-participating group, making a valid comparison impossible.
  2. The employees were not randomly assigned to participate or not participate; they chose for themselves. (correct answer)
  3. The response variable, number of sick days, is not a sufficiently reliable measure of employee wellness.
  4. The study did not use random sampling to select the employees who were offered the program.
Explanation: The correct answer is B. The defining characteristic of an experiment is the random assignment of subjects to treatment groups. In this study, employees self-selected into the 'treatment' group (program participation) or the 'control' group (non-participation). This introduces potential confounding; for example, employees who are already more health-conscious might be more likely to both participate in the program and take fewer sick days, regardless of the program's effectiveness. (A) is incorrect because while placebos are used in some experiments, their absence is not what defines this as non-experimental. The core issue is the lack of random assignment. (C) is incorrect because the choice of response variable is a matter of measurement validity, not the fundamental design type of the study. (D) is incorrect because random sampling relates to generalizability to a larger population, not whether the study is an experiment or observational.

Question 8

A city government observed that neighborhoods with a higher number of public parks per capita also had significantly lower rates of reported crime. They concluded that building more parks would be an effective crime-reduction strategy. Which of the following represents the most significant flaw in this conclusion?

  1. The study was observational, and a confounding variable, such as average neighborhood income, could be associated with both more parks and lower crime. (correct answer)
  2. The study was an experiment that lacked a proper control group, such as randomly selected neighborhoods where no new parks were built.
  3. The sample size, consisting of neighborhoods within a single city, is too small to draw any conclusions about the relationship between parks and crime.
  4. The data on crime rates were based on reported crimes, which may not accurately reflect the true crime level in a neighborhood.
Explanation: The correct answer is A. This was an observational study because the city did not assign parks to neighborhoods; it only observed existing conditions. The primary issue with drawing a causal conclusion from an observational study is the potential for confounding variables. In this case, wealthier neighborhoods might have more resources for both public parks and crime prevention, or other factors related to wealth could lead to lower crime. Therefore, the observed association might be due to income, not the parks themselves. (B) is incorrect because the study was observational, not an experiment. No treatment was imposed. (C) is incorrect because while generalizability might be limited, the most significant flaw is the logical leap to causation, which is an issue regardless of sample size. (D) is incorrect because while data accuracy is a potential issue in any study, the fundamental flaw in the conclusion about causation is the study design itself (confounding), not the measurement of the response variable.

Question 9

A school district wants to test the effectiveness of a new math curriculum against the old curriculum. They have 10 elementary schools. They randomly select 5 schools to implement the new curriculum, and the other 5 continue with the old one. At the end of the year, they compare the average standardized test scores for all students in the two groups of schools. Which statement best describes this study?

  1. A well-designed experiment where the experimental units are the individual students.
  2. A prospective observational study because the students themselves were not individually assigned a curriculum.
  3. An experiment where the experimental units are the schools. (correct answer)
  4. A matched-pairs experiment, because each new-curriculum school is implicitly paired with an old-curriculum school.
Explanation: The correct answer is C. This is an experiment because a treatment (the new curriculum) was randomly assigned. However, the random assignment was done at the school level, not the student level. Therefore, the independent units that were randomly assigned are the schools, making them the experimental units. This is often called a cluster randomized trial. (A) is incorrect because the students were not the units of randomization. (B) is incorrect because the random assignment of the curriculum makes it an experiment, not an observational study. (D) is incorrect because the design is a completely randomized design (at the school level), not a matched-pairs design. A matched-pairs design would involve pairing similar schools first and then randomly assigning the curriculum within each pair.

Question 10

Ecologists studying the effect of a reintroduced wolf population on an elk herd measure elk population density in a national park for five years before and for five years after the wolves were introduced. They find that elk density significantly decreased after the reintroduction. What is the most significant limitation in concluding that the wolves caused the elk population's decline?

  1. The study is observational, and other factors, such as a change in climate or a disease outbreak, could have coincided with the wolf reintroduction. (correct answer)
  2. The sample size of one park is too small, so the results cannot be generalized to other ecosystems where wolves might be introduced.
  3. The ecologists were not blinded to the presence of the wolves, which could have introduced bias into their population density measurements.
  4. It is an experiment, but it lacks a proper control group, such as a similar park that did not have wolves reintroduced.
Explanation: The correct answer is A. This is a pre-post observational study (sometimes called a natural experiment). Since the researchers did not randomly assign the wolves to the park, they cannot rule out other factors that may have changed over the 10-year period. A harsh winter, a new plant disease affecting elk forage, or a disease among the elk are all examples of confounding variables that could have caused the decline. (B) discusses generalizability, which is a valid concern, but it is not the primary limitation for drawing a causal conclusion within this specific park. (C) mentions a potential source of measurement bias, but the larger, more fundamental threat to the causal conclusion is confounding. (D) is close, but classifying it as an experiment is less accurate than calling it observational. The core idea that a control group is missing is correct, but option A better articulates the consequence of that missing control group, which is the inability to rule out confounding variables.

Question 11

A pharmaceutical company tests a new drug for lowering cholesterol. They recruit 200 volunteers with high cholesterol from a national database of patients. They randomly assign 100 to receive the new drug and 100 to receive a placebo. The study is double-blind. The drug is found to be significantly more effective than the placebo. Which is the most appropriate conclusion?

  1. The study shows an association, but not a causal link, between the new drug and lower cholesterol.
  2. There is evidence that the new drug causes a reduction in cholesterol for patients similar to those in the study. (correct answer)
  3. The new drug causes a reduction in cholesterol for anyone with high cholesterol.
  4. The study design is flawed because the volunteers were not a random sample of the entire country's population.
Explanation: The correct answer is B. Because the study used random assignment to treatment groups, researchers can make a causal conclusion. Because the study was double-blind and used a placebo, the effect is likely due to the drug itself. However, the volunteers, while from a national database, are still a sample and may not perfectly represent all people with high cholesterol. Therefore, the conclusion should be tempered, applying to patients similar to the volunteers. (A) is incorrect because random assignment allows for causal inference. (C) is an overstatement. The conclusion is generalized too broadly to anyone with high cholesterol. (D) is incorrect because while the use of volunteers may limit generalizability, it does not 'flaw' the study's ability to draw a causal conclusion for the population represented by the sample.

Question 12

A researcher analyzes a large database of medical records and finds a statistically significant association between patients who were prescribed a specific antidepressant and a higher incidence of a particular cardiovascular disease. The researcher's conclusion that the antidepressant is a risk factor for the disease is likely invalid because the study is a(n)   and fails to account for  .

  1. experiment; the placebo effect
  2. survey; non-response bias
  3. prospective observational study; participant attrition
  4. retrospective observational study; confounding variables (correct answer)
Explanation: The correct answer is D. The study design is a retrospective observational study because it uses existing records to look back at past exposures (antidepressant prescription) and outcomes (cardiovascular disease). The main weakness of this design is the potential for confounding variables. For example, the underlying condition of severe depression (which prompted the prescription) might itself be linked to lifestyle factors (poor diet, lack of exercise) that increase the risk of cardiovascular disease. The drug might not be the cause at all. (A) is incorrect; this is not an experiment as no treatment was assigned by the researcher. (B) is incorrect; while it uses a database, its primary structure is not that of a survey, and non-response bias is not the central issue. (C) is incorrect; the study is retrospective (looking at past records), not prospective (following subjects into the future).

Question 13

A team of researchers conducts a survey of a random sample of 2,000 adults in a country. They find that individuals who report reading for pleasure for at least 30 minutes daily have, on average, higher incomes than those who do not. Which of the following is the most accurate statement about the conclusions that can be drawn?

  1. The results can be generalized to adults in the country, and there is evidence that reading for pleasure causes an increase in income.
  2. The results cannot be generalized, but there is evidence that reading for pleasure causes an increase in income for those surveyed.
  3. The results can be generalized to adults in the country, but the study only shows an association, not a causal link, between reading and income. (correct answer)
  4. No valid conclusions can be drawn because the study is observational and relies on self-reported data which is often inaccurate.
Explanation: The correct answer is C. The study used random sampling from the country's adult population, so the findings can be generalized to that population. However, the study is observational (a survey) because no treatment was assigned. Therefore, it can only establish an association. It is impossible to rule out confounding variables (e.g., higher education levels could lead to both more reading and higher income). (A) incorrectly claims causation. (B) incorrectly claims causation and also incorrectly states that the results cannot be generalized. (D) is too strong. While self-reported data has limitations, valid conclusions about association can still be drawn from well-conducted observational studies.

Question 14

In the design of a scientific study, what is the primary purpose of randomly assigning subjects to treatment groups?

  1. To ensure that the subjects in the study are representative of a larger population of interest.
  2. To reduce the impact of the placebo effect on the measurement of the response variable.
  3. To allow the use of probability and inferential statistics to analyze the study's results.
  4. To create treatment groups that are, on average, similar in all respects before the treatment is applied. (correct answer)
Explanation: The correct answer is D. The primary goal of random assignment is to balance the influence of all other variables—both those we can measure (like age, gender) and those we cannot (like genetic predisposition, motivation)—across the treatment groups. By making the groups as similar as possible at the outset, researchers can be more confident that any differences observed after the treatment are due to the treatment itself and not to pre-existing differences. (A) describes the purpose of random sampling, not random assignment. (B) describes the purpose of using a placebo and blinding, not random assignment. (C) is incorrect. While random assignment is a key assumption for many statistical tests used to analyze experiments, the fundamental purpose of the technique in the study design is to control for confounding, not merely to enable a calculation.

Question 15

An observational study found that individuals who regularly take vitamin C supplements have a lower incidence of the common cold. However, it was also noted that people who take supplements are often more health-conscious in general, engaging in more exercise and eating healthier diets. In this context, the general 'health-consciousness' of the individuals is best described as what type of variable?

  1. A response variable
  2. A confounding variable (correct answer)
  3. An explanatory variable
  4. A placebo variable
Explanation: The correct answer is B. A confounding variable is a variable that is associated with both the explanatory variable (vitamin C use) and the response variable (incidence of colds) and can create a spurious association. Here, 'health-consciousness' is likely associated with taking vitamins and also with behaviors that independently reduce the incidence of colds, thus confounding the relationship between vitamins and colds. (A) is incorrect; the response variable is the outcome being measured, which is the incidence of the common cold. (C) is incorrect; the explanatory variable is the factor being studied as a potential cause, which is vitamin C supplement use. (D) is incorrect; 'placebo' refers to an inert treatment in an experiment, not a type of variable in an observational study.

Question 16

Researchers wish to determine if a new online learning platform is more effective than traditional classroom instruction. They randomly select 100 students from a large university who need to take a specific course. They then allow these 100 students to choose whether to enroll in the online version or the traditional version of the course. Which statement correctly identifies the primary flaw in this study's design?

  1. The lack of random assignment to instructional method makes it an observational study, preventing a valid causal conclusion. (correct answer)
  2. The use of random sampling is irrelevant when the sample size is only 100 students.
  3. The study is a valid experiment, but the lack of a placebo group for the online platform is a significant weakness.
  4. The study is flawed because it only includes students from one university, limiting generalizability.
Explanation: The correct answer is A. Even though the students were randomly selected for inclusion in the study, the critical step of randomly assigning them to the treatments (online vs. traditional) was not performed. Instead, students chose their own group. This self-selection makes the study observational. Any difference in outcomes could be due to pre-existing differences in the students who choose each format (e.g., more self-disciplined students might choose the online option), thus confounding the results and preventing a causal conclusion. (B) is incorrect; random sampling is always relevant for generalizability, regardless of sample size. (C) is incorrect; the study is not a valid experiment due to the lack of random assignment. (D) identifies a limitation on generalizability, but it is not the primary flaw in the study's internal design for testing effectiveness.

Question 17

A researcher recruited 100 volunteers with moderate anxiety from a university campus. She randomly assigned 50 volunteers to a new mindfulness meditation program and 50 to a waitlist control group. At the end of eight weeks, the meditation group showed a statistically significant reduction in anxiety scores compared to the control group. What is the most appropriate scope of inference from this study?

  1. The results can be generalized to the entire population of adults with anxiety, and the study provides evidence that the meditation program causes a reduction in anxiety.
  2. The results can be generalized to the population of students at that university, but it only shows an association, not a causal link, between meditation and reduced anxiety.
  3. The study provides evidence that the meditation program causes a reduction in anxiety for individuals like the volunteers, but the results may not generalize to all adults with anxiety. (correct answer)
  4. Neither generalization to a larger population nor a causal conclusion is warranted because the participants were volunteers.
Explanation: The correct answer is C. The study used random assignment, which allows for causal inference (the program caused the reduction in anxiety). However, the participants were volunteers from a single university campus, not a random sample of the entire population of adults with anxiety. Therefore, the results are most reliably applicable to individuals similar to those in the study, and generalization to a broader population is not warranted. (A) is incorrect because while a causal conclusion is appropriate, generalization to all adults with anxiety is not, due to the non-random sample. (B) is incorrect because it mistakenly claims a causal link cannot be established. Random assignment is the key to establishing causation. (D) is incorrect because causal conclusions are warranted due to random assignment. The use of volunteers limits generalizability but does not invalidate the causal link for the group studied.

Question 18

Ecologists studying the effect of a reintroduced wolf population on an elk herd measure elk population density in a national park for five years before and for five years after the wolves were introduced. They find that elk density significantly decreased after the reintroduction. What is the most significant limitation in concluding that the wolves caused the elk population's decline?

  1. The study is observational, and other factors, such as a change in climate or a disease outbreak, could have coincided with the wolf reintroduction. (correct answer)
  2. The sample size of one park is too small, so the results cannot be generalized to other ecosystems where wolves might be introduced.
  3. The ecologists were not blinded to the presence of the wolves, which could have introduced bias into their population density measurements.
  4. It is an experiment, but it lacks a proper control group, such as a similar park that did not have wolves reintroduced.
Explanation: The correct answer is A. This is a pre-post observational study (sometimes called a natural experiment). Since the researchers did not randomly assign the wolves to the park, they cannot rule out other factors that may have changed over the 10-year period. A harsh winter, a new plant disease affecting elk forage, or a disease among the elk are all examples of confounding variables that could have caused the decline. (B) discusses generalizability, which is a valid concern, but it is not the primary limitation for drawing a causal conclusion within this specific park. (C) mentions a potential source of measurement bias, but the larger, more fundamental threat to the causal conclusion is confounding. (D) is close, but classifying it as an experiment is less accurate than calling it observational. The core idea that a control group is missing is correct, but option A better articulates the consequence of that missing control group, which is the inability to rule out confounding variables.

Question 19

A researcher analyzes a large database of medical records and finds a statistically significant association between patients who were prescribed a specific antidepressant and a higher incidence of a particular cardiovascular disease. The researcher's conclusion that the antidepressant is a risk factor for the disease is likely invalid because the study is a(n)   and fails to account for  .

  1. experiment; the placebo effect
  2. survey; non-response bias
  3. prospective observational study; participant attrition
  4. retrospective observational study; confounding variables (correct answer)
Explanation: The correct answer is D. The study design is a retrospective observational study because it uses existing records to look back at past exposures (antidepressant prescription) and outcomes (cardiovascular disease). The main weakness of this design is the potential for confounding variables. For example, the underlying condition of severe depression (which prompted the prescription) might itself be linked to lifestyle factors (poor diet, lack of exercise) that increase the risk of cardiovascular disease. The drug might not be the cause at all. (A) is incorrect; this is not an experiment as no treatment was assigned by the researcher. (B) is incorrect; while it uses a database, its primary structure is not that of a survey, and non-response bias is not the central issue. (C) is incorrect; the study is retrospective (looking at past records), not prospective (following subjects into the future).

Question 20

An athletic trainer wishes to test whether a new type of sports drink improves endurance. He recruits 40 student-athletes and randomly assigns 20 to receive the new drink and 20 to receive a placebo drink that is identical in taste and appearance. He then has them run on a treadmill until exhaustion and records the time. The trainer knows which drink each athlete received. Which of the following is the most significant potential source of bias remaining in this design?

  1. It is an observational study, so it cannot be used to determine causation.
  2. The sample size of 40 athletes is too small to detect a real effect.
  3. The study is single-blind, so the trainer's expectations could influence his interaction with the athletes. (correct answer)
  4. The lack of random sampling of athletes limits the generalizability of the findings.
Explanation: The correct answer is C. The study is an experiment with a placebo control, but it is only single-blind because the researcher (the trainer) is aware of who is in which group. This knowledge could unconsciously affect his behavior—for example, he might offer more verbal encouragement to the athletes he knows received the real sports drink. This is a form of experimenter bias. A double-blind design, where neither the participants nor the person administering the treatment knows the group assignments, would eliminate this. (A) is incorrect; it is an experiment due to random assignment. (B) is an issue of statistical power, not bias. A small sample might fail to find an effect, but it doesn't systematically skew the results in one direction. (D) is a limitation on the scope of inference (generalizability), not a source of bias within the experiment itself.