AP Statistics : Data Collection
Study concepts, example questions & explanations for AP Statistics
All AP Statistics Resources
Example Questions
Example Question #1 : How To Identify Sources Of Bias In An Experiment
A major chocolate company wants to test the effects of adding more sugar to their standard chocolate bar to see if customers enjoy it more.
They select 10 subjects to randomly participate in a taste test. They bring in samples of their original product, which is sold in tiny squares that says the company's name on them, and samples of the increased-sugar versions, which are plain chocolate squares of the same size. The company asked participants to taste both chocolates and rank how much they like them on a scale of 1 to 10.
Which of the following represents a possible source of bias in the study?
The original recipe has the company's name on it, but the new sample does not.
There are not many taste-testers who are qualified to evalutate food.
Some people may wish they were eating a different dessert.
There is no bias.
People may not all like chocolate to the same degree.
The original recipe has the company's name on it, but the new sample does not.
The presence of the company's name on the original sample may be a soure of bias. If people already have a preexisting opinion about the brand, they may rate that chocolate as better or worse based on those opinions rather than flavors.
Example Question #92 : Data
Let us suppose a company wants to evaluate whether a new medical device works better than current devices. It conducts a small experiment to assess the effectiveness of the new device. To conduct the experiment, the company randomly assigns one group to the new medical device, which requires users to stay well hydrated, and the other group to the old device.
How should we control for confounding variables?
The group receiving the new device should simply receive the device without being asked to stay hydrated.
Participants should be able to choose which device is right for them.
The group receiving the old device should also be required to stay hydrated.
The group receiving the old device should also be required to stay hydrated.
When comparing the effectiveness of a treatment, one should try to ensure that only the treatment varies across groups. In this case, the new device is compared to an old device. However, the new device also requires that users stay well hydrated. If we observe any positive effects from the new device, we won't know whether the new device is effective, or if merely staying well hydrated is actually what is effective. To rule out this confounding variable, we should also ask the group using the old machine condition to stay hydrated as well.
Example Question #1 : How To Identify Confounding Factors In An Experiment
An experiment was done by medical researchers to determine the association between drinking caffeine and severity of lung cancer. Results showed that there was a high association between the two variables. Which of the following could be a potential confounding variable in the experiment?
Lung Cancer
None of these are potential confounding variables
Caffeine Consumption
Medical Researchers
Smoking
Smoking
A confounding variable is one that could potentially have an effect on both the independent and dependent variables in a study. In this case, it is possible that there is an association between smoking and caffeine as well as smoking and lung cancer.
Example Question #1 : How To Identify Confounding Factors In An Experiment
A study finds that caffeine intake has a strong positive correlation with grades for college students. In other words, on average, the more caffeine intake a student has, the higher a grade the student gets.
Which of the following could potentially be a confounding variable in this experiment?
The amount of soda that each student consumes
Amount of sleep a student gets each night
The caffeine intake of students
The grade a student receives
The amount of coffee that each student drinks
Amount of sleep a student gets each night
The only confounding variable in this experiment is the amount of sleep that each student gets. A confounding variable is one that has an impact on both the dependent and independent variable. It is possible that the amount of sleep a student gets is related to caffeine intake, which in turn affects the grade a student receives on a test or assignment.
Example Question #93 : Data
An experiment testing the effects of caffeine on endurance performance in athletes assigns caffeine to a randomly selected group of athletes and has them exercise. Another trial was conducted in which the same group exercised without anything given to them to take. The results did not match the expected results. What should be done to improve this experiment?
Caffeine may affect different people in different ways, so varying amounts of caffeine should be administered.
A different group should be used for each trial because the athletes' first trial may have influenced their second trial.
Nothing, the experiment is sound
There may a placebo effect with the caffeine, so an identical application without caffeine should be given to the control group.
The group should be randomly selected from a population of athletes and non-athletes, not just athletes.
There may a placebo effect with the caffeine, so an identical application without caffeine should be given to the control group.
The placebo effect can potentially be a confounding variable. By knowingly taking a substance, participants may feel more energenized. By administering the same substance both trials, with the only thing changed being caffeine content, this corrects for this possible confounding variable.
Example Question #1 : How To Identify Confounding Factors In An Experiment
A small local umbrella company is trying to test the effectiveness of their umbrellas by looking at how many umbrellas they sell each year.
In 2014, the company sold 2,000 umbrellas.
In 2015, they sold 1,500 umbrellas.
They assume that their umbrellas are less effective which is why sales decreased.
However, there could be many confounding factors. Which of the following is NOT a possible confounding factor?
There may have been less rain in 2015 than in 2014, therefore decreasing the need for umbrellas.
An increase in prices may have led to decreased sales.
There are no confounding factors.
If the same people live in the city during 2014 and 2015, people may already have umbrellas in 2015 and might not need to buy them.
Three of these could be confounding factors.
Three of these could be confounding factors.
Any of these answers could explain why umbrella sales dropped. You cannot assume any specific cause explains a change in data like this-- further experimentation should be done rather than assuming cause and effect.
Example Question #11 : Data Collection
A study is trying to determine if a particular medication (Y) is effective in weight loss. Patients participating in the study were randomly assigned to groups A, B, C, D, or E. Group A will receive one dose of Y, Group B will receive two doses of Y, Group C will receive three doses of Y, Group D will receive four doses of Y, and Group E will serve as the control group.
Which group will be receiving the placebo (a sugar pill)?
Group E
Group D
Group B
Group C
Group A
Group E
The control group in an experiment typically receives placebo treatments (in this case - Group E). Since all of the other groups are receiving at least one dose of the medication, they are considered to be experimental groups.
Example Question #1 : How To Identify The Placebo Effect In An Experiment
To test if vitamin C actually makes people feel better, a vitamin company decides to run a 5-day study where they give one group of 100 sick participants vitamin C pills and another group of 100 sick people placebo pills, and monitored another group of 100 sick people who took no pills.
At the end of the 5-day experiment, 90 participants in the vitamin C group reported feeling better. 30 participants in the no-pill group felt better after the 5-day period. Interestingly, 50 participants in the placebo group felt better after the 5-day period.
What could explain these numbers?
5 days is too long for the experiment.
The placebo effect: some participants in the placebo group began to feel better because they thought they were taking something that would help them.
Vitamin C actually makes people feel sicker.
5 days is too short for the experiment.
There is no logical explanation.
The placebo effect: some participants in the placebo group began to feel better because they thought they were taking something that would help them.
The placebo effect is when effects are seen in a group of people who did not actually receive a treatment.
In the vitamin C group, 90 participants felt better.
Naturally (no-pill), 30 participants felt better.
With the placebo, 50 participants felt better. Since more people felt better with the placebo than with no treatment at all, it appears that some percentage of people believed that they would feel better with a pill and actually began to feel better due to the placebo effect.
Example Question #1 : How To Establish Blind Experiments
A study is attempting to test various brands of cola against each other. Which of these would not be a measure that could be used to help create a double blind test?
Removing or covering up the labels on the cans or bottles.
Making sure the test administrator also does not know the brand of each sample until after the test.
Randomizing the placement of the cola samples.
Filling the sample cups the same way.
Serving the samples to the taster in their usual containers.
Serving the samples to the taster in their usual containers.
Remember that a double blind experiment means that neither the subject nor the person conducting the experiment knows the identity of the various treatments until the test is over, so as to minimize bias.
Therefore of the options,
serving the samples to the taster in their usual containers, is the one which would not help create a double blind test.
Example Question #1 : How To Establish Blind Experiments
Which of the following is an example of a double blind experiment?
An experiment in which researchers know which therapy test subjects are receiving, but test subjects are unaware
None of these are an example of a double blind experiment
An experiment in which neither the test subjects nor the researchers know who is receiving the treatment and who is receiving the placebo
An experiment in which both test subjects and researchers are aware of who is receiving the treatment and who is receiving the placebo
An experiment in which test subjects know which therapy they are receiving, but researchers are unaware
An experiment in which neither the test subjects nor the researchers know who is receiving the treatment and who is receiving the placebo
A double blind experiment requires that both researchers and test subjects are unaware of who is receiving the treatment and who is receiving the placebo. If only one group is unaware, it is a single blind experiment. If both groups are aware, the experiment is not blinded.
Certified Tutor
All AP Statistics Resources
