Chapter 7 Homework Solutions – 5 Steps to Setting up an Experiment
Answers on these could obviously vary, but I just wanted to give you a few more examples of the type of thing I’ll be looking for on the test.
18.Showing video lectures with and without humor.
Step 1 Identify Population and Variable
The group of people we are trying to make a claim about is students, more specifically college students, even more specifically, college students taking a statistics course (or something related to it). So that’s the population.
The thing about students we want to try to measure is how well they learn. Again, being more specific, we are going to try to measure how well they remember/learn material presented in a video lecture. So the variable is whether the learn/remember the material.
Step 2 Get a Sample
This is sort of described for us in the problem – they randomly selected 117 students to participate in the study. It’d be a good idea to try to insure some variety in terms of grade, major, etc., but we’ll assume they’ve selected their sample responsibly.
Step 3 Divide Sample Into Groups
We’re trying to test if making the videos funny increases how well students remember/learn, so we’ll want two groups – one which views the funny videos and the other which views the standard (presumably boring) videos. Again, it’d be best if the group assignments were random, but we’ll assume they were.
Students were not told that they were participating in some sort of experiment, but simply that they should watch the videos. As such, this experiment is basically single blind because no student knew which characteristics of the video were being studied. In order to make the study double blind, we could make sure the people who grade the quizzes later didn’t know which group the quiz-taker was in.
Step 4 Impose Treatment
Sit the kids down and show them the videos. It’d be best if it was in some sort of controlled environment, so that we have some assurance that they’re actually watching the video and not updating their Facebook.
Step 5 Gather Results
In order to test students’ understanding of the material in the video, we would give them a standard quiz and compare results between the two groups.
Some possible confounding variables could include: general study habits/academic performance of the students, how much sleep the students had the night before, past comprehension of the material presented, natural serious attitude toward studies, etc.
22.Offering internet-based telephone services with different pricing schemes.
Step 1 Identify Population and Variable
The group of people targeted in the study described was college students, but the company may wish to generalize their findings to all potential customers – i.e. all adult Americans or something like that.
If we had one of these participants sitting in front of us now, the question we’d want to ask them would be “how much do you use this internet-based phone service?” So our variable is use of internet-based phone system.
Step 2 Get a Sample
As mentioned, this group decided to target residents in a college dorm, but it may be wiser to select a more diverse sample. They could possibly contact a random group of customers who receive broadband internet and ask them to participate in the trial.
Step 3 Divide Sample Into Groups
The company is trying to compare the affect two different pricing schemes will have on how much their service is used, so they should break their sample into two groups – one group who pays a flat rate and one group who pays a non-flat rate.
Step 4 Impose Treatment
Hook’em up to the internet and chat away.
Step 5 Gather Results
Since this service is through the internet, it is probably very convenient to collect data on when and how much the service was provided and compare this data between the two groups.
Possible confounding variables could include: people who use cell phones exclusively, local students who make fewer long distance calls than non-local students, students who don’t pay their own bills, so they would be unaware of possible savings by calling during non-peak hours, etc.