AP Statistics : How to identify confounding factors in an experiment
Study concepts, example questions & explanations for AP Statistics
All AP Statistics Resources
Example Questions
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.
Certified Tutor
All AP Statistics Resources
