Mr. Weingarden's take on this topic:

Hey folks, I've read many opinions about the simulation that we discussed in class. I would like to try to provide some clarification because there is some disagreement on why the simulation was performed and what the simulation tells us.

Why the simulation was performed

The simulation was performed because there is always a concern that when you get a particular result from an experiment that the result might have been obtained by sheer luck. So, a simulation was performed on the magnet and pain data described in the Chapter P Case Study.

How the simulation worked

The goal was to see whether we just got lucky that many people said they felt less pain after treatment with the magnet. To do this, we had the computer take all of the pain observations and mix them up. Then the computer randomly assigned 29 of the participants pain readings to the category called "active" and 21 of the pain readings to the category called "inactive." The mean of each group was calculated and one was subtracted from the other. This tells us the difference in average pain between the two groups.

The actual study found that there was a difference of about 4 between the mean pain observations in the two groups. We ran our simulation 10,000 times. Each time the pain readings were mixed up and the difference in mean pain readings was calculated.

What the results mean

After running the simulation 10,000 times, we discovered that we never (or almost never) ended up with a difference of 4 between the means of the randomly chosen active and inactive groups. This means that it is highly unlikely that the actual results of the study occurred by just random chance. In other words, it's likely that magnets do lower perceived pain.

I would like to provide an example to contrast this conclusion. The simulation showed us that out of 10,000 mixings of pain readings, we get a difference of zero between the means about 2,000 times. So, if in the actual study, we saw a difference of zero between the means, it's entirely possible that could have occurred by sheer random chance. This is because a difference of means with the set of data we have will be zero quite often if we just randomly shuffle.

Your thoughts on the Magnets and Pain Simulation

On this page, you can add your thoughts regarding the Chapter P Case Closed simulation.

- Click the EDIT tab above and on the far right to edit this page
- Tell me about the simulation here and make sure you add your name. Please do not delete or modify other students' entries.

In the original experiment doctors wanted to find out if the use of magnets reduced the pain of polio patients. The experiment included 50 patients and the outcome showed an improvement in those that had the active magnets. To prove that the results weren't just by chance, we had the software rearrange the results 10,000 times. It ended up showing that the range was near 0 and never 4. This proves that the use of magnets actually does help pain in polio patients.
Amber Escudero

In this case study the docots were using magnets to see if they could reduce pain. The porcess was done and the patients gave a number 1-10 to describe their pain. The result came in and we wanted to know how accurate the numbers were. So, the numbers were grouped together and the active and inactive catergories were thrown away. Then the first 29 numbers were randomly grouped and the remaining numbers were grouped together. Those two groups were averaged and then the averages were subtracted from eachother to get a final average. This process was repeated 10,000 times. When the process was done once by the authors in the book the number found was 4. However when we did this 10,000 times the majority of the numbers were around 0. This tells us that the use of actice or inactive magnets doesnt really have an impact on pain.
myles harvey

The original Case Study was done to see if an Active or Inactive magnet could reduce the pain of people with polio. The patients and doctors didn't know whether they were using an Active or Inactive magnet. The results were based solely on what the patients felt or thought they felt. The simulation was done to answer the question of did the magnets really have an impact? What if all the magnets were really just Active or Inactive, not both? Then what would be the chances of the patients giving the same numbers on the pain scale? The answer to that question is why the simulation was done. They didn't add or change any of the numbers, they just rearranged them by mixing them up and picking the first 29 followed by the last 21. This step was followed by taking the means of both sets of data and subtracting them. The simulation showed us that 2000 out of 10,000 times the mean difference would be 0, What this tells us about the results we actually got from the study is that there really isn't that big of a differnce in pain between using the Active of Inactive magnet.
Shayda Ansari

In the original study the researchers had the patients rate their initial pain and the pain after treatment with the magnents (Inactive or active) on a sacle of 1 to 10 (10 being worst). Since the patients all rated their initial pain highly the researchers decided to focus on the answers given after treatment. Since neither the doctors nor the patients knew which magnent they were getting, 29 out of 50 patients recieved active magnents, 21 of 50 recieved inactive magnents. Once they got their final results they found the average pain after treatment for the active group and then for the inactive group. Then they subtracted the difference between the two groups to get a difference of 4.05. But there is a chance that they just happened to get this number and that the magnents didn't help at all. They decided to see how common an answer of 4.05 was. If the magnents did absoblutly nothing (they had the patients sit for 45 minutes with the magnents, perhaps it wasn't the magnents that helped the pain just passed) then the patients should have still answered the same number. So they randomly mixed up the numbers calling the first group active (29 in this group) and the second group inactive (21 in this group). Averaged each of the results and subtracted atcive form inactive. They did this random grouping then averaging for a total of 10,000 times! And of these 10,000 times they never even got an answer of 4.05. This shows that the answer of 4.05 is not a common answer at all. There is a 0% chance of getting the answer the did just by random chance, therefore the magnents must have affected the patients answers proving that magnents actually do help to reduce pain.
Julie Schneider

In class we looked at a case study to see the effects that an active magnet would have on pain. The experiment took 29 pain patients and tried to sooth their pain with an active magnet. It then took another 21 pain patients and gave them an inactive magnet to try to sooth their pain. Neither of the groups were told whether or not their magnets were the active or inactive ones. After we recorded the pain of the two groups we got the mean of how the rated their pain and took the difference. In class our difference was around 4. Needing to make sure that our results weren't due to chance we rearranged the numbers that the pain patients used to describe their pain and put them into new groups and followed the earlier procedures of getting the mean of both groups and then finding the range. We then made a program to do this for us and had the program run the data 10,000 to see the results. We found that the range came out to be near zero the majority of the time and never came out to be 4, using this information we can safely assume that our results were not due to chance and the active magnets really reduced the pain of the patients.
Kai Bowden

In class we tested to see if active magnets would or would not help reduce chronic pain of patients. We read that the experimenters (doctors) took 50 patients and randomly selected 29 to receive the active magnet treatment and the rest to receive the inactive magnet treatment. To get rid of any answer swaying variables, neither the patients nor the doctors were aware of what magnets they were given. After careful testing, we graphed the results and saw that the active magnets seemed to have lessened the pain of the patients. In order to support the results, we took the mean of both the active and inactive magnet patients and subtracted them and got an average pain rating of about 4. To further suggest that our result was not by chance or luck, we punched the same initial pain ratings of all 50 patients into a simulation and told it to randomly select 29 patients to receive the active magnet test and 21 patients to receive the inactive magnet treatment, using those initial pain ratings. We then told the simulation to run the test 10,000 times and told it to come up with a graph of what the results were. To our surprise, the simulation showed that there was a 0% chance of getting our answer of about 4 which proves that the active magnets MUST have helped reduce the pain of the active magnet treated patients.
Alex Choi

In the Case Study for the Preliminary Chapter, an observational study was preformed to see if magnets could help relieve the pain in polio suffering patients. The procedure was that a doctor asked for a pain rating from the patient using numbers 0 through 10 with 0 being no pain and 10 being extreme pain. He then records the results and then proceeds to pull from 1 of 50 envelopes containing active and non-active medical grade magnets and applies it to the patient. He then asks the patient to give another pain rating and he records it. The final results imply that the magnets do in fact help with the pain. To prove that these results are genuine and not just the outcome of pure happenstance, it is necessary to test how likely it would be for this situation to just happen randomly. The relative variable that would be used to test this situation would be the mean difference between the active and inactive magnet results. To test this you take all the data, active and inactive, and randomize them into two groups with the amount of data points equal to the two amounts of actual active and inactive magnets used in the original study which is 29 active and 21 inactive. You then find the means of both groups, find the difference between them, and record the value. Repeat this process many times, in this case 10,000 times, and you will have successfully created close to a totally random happening of the observational study many times. If the mode of the simulation was close to the original mean difference if would prove the study’s results too likely to happen on a random basis. But, the result was not this, proving that the study had some ground to stand on and could be looked upon as a legitimate study with a clear result. This simulation proves that magnets could actually be used to relieve pain in polio suffering patients based on the original study, not just pure happenstance.
David Wolfe

At the beginning of the case study it seemed as though the manets were useful tool in helping reduce pain. However the question soon arises if this just a random chance of data. Too find out if this was a random chance we must first place all 50 results of te pain readoings from both groups into one grooup and randomize them into two groups active and inactive. OUr rationale for this is that what if some of te people in the original inactive group were in the active group. By taking the mean of both groups and subtracting you get the mean difference. Then the questio becomes, how often will we get the result of 4.05? After running the above steps 10,000 times, with each being a new randomly selected group with new means, it results in the histogram found in "Case Closed." A mean difference of two comes up 200 of those 10,000 times. Whereas it seems that a value of 4 comes up maybe 10 times. Zero comes up around 2,500 times. What does this tell us? It means that 4 is an incredibly random chance of data to get. Its nearly unreal how rare this set of data comes up. This tells us that there is really no difference in using or not using magnets as a result of zero comes up more often. The fact that the original study got a value of 4 was pure random chance, though it may have been small.
Austin Lasater

The motive for this case was to find out whether or not applying magnets to patients with polio would decrease the pain level. The doctors experimented on the group of patients by applying an active magnet on 29 of them and the in active magnet on 21 of them. However, the doctors did not tell the patient if the magnet was active for psychological factors that could contribute to inaccurate results. In the first study that we did in class, we found that the active magnet did appear to have a lower pain level than the inactive magnets. After taking the mean of both inactive and active groups and finding the difference between that amount, we came up with an answer of 4.05. The question that occurred was: was the results we gathered actually just a random chance of luck, or was it actually accurate? To test this question we randomized our initial numbers, without changing or adding any other numbers to it. We did this 10,000 times electronically, and never got a result of 4.05. By doing this, we could gather that magnets did not have as big of an affect as we presumed from our first stimulation. We gathered that 2000 of the 10000 results had a mean difference of 0, which means that magnets weren't that effective.
Busy Rylander

The Case Study we reviewed in the preliminary chapter examined the effect magnets had on physical pain, and in this case the subject group were polio patients. Before any experimenting the participants recorded their current level of pain which spanned between 1-10. Doctors tested active magnets with 29 patients while the other 21 were given inactive magnets, all the while both doctor and patient unknowing to what type of magnet they were given. After 45 minutes of the test, patients were once again asked to disclose their level of pain. In the end the overall report seemed scattered, although it was definitely evident that the people who were provided with active magnets appeared to have lower levels of pain. Using that data we calculated in class that the difference of means in each group was 4.05. However the scientists in this study wanted to ensure that the chance that the reports weren't just random, so they created a histogram which displayed 10,000 of the differences between the mean pain scores. In an attempt to understand this process, we re-created the simulation by taking all 50 data points and scrambling them. Following that we placed the first random 29 in the active group and the last 21 in the inactive, and then found the mean of each group and subtracted. After repeating this process an increasing number of times we continually discovered 0 to be the result. This procedure shows that the chance of us calculating 4.05 was close to 0%, therefore meaning that the active magnets most likely are effective in aiding pain.
Madison Richeson

The original case study we discussed in class was where doctors used magnets to see if they would help aid the pain of patients with polio. The doctors used both inactive and active magnets on 50 patients. Using those results they made the study larger to see its effect without having to test more patients. In this simulation, we were taking our results and rearranging them 10,000 times to stimulate a larger experiment to prove that the active magnets actually do make an impact.
Alison Dobner

The reason we did the simulation was because we were trying to find out if the magnets were able to reduce the pain of the patients. When we first looked at the charts, it looked like the magnets did help out the patients, but we did the simulation to make sure that those results were not just luck. in our results we saw that the chances of those results being luck is close to 0%.
Julio Rodriguez

The Case Study was done to see if magnets that were active or inactive could reduce the pain of Polio Patients. The doctors experimented on a group of 50 patients and split them into two groups. 29 of them recieved active magnets while the other 21 recieved inactive magnets. The ones who recieved an active magnet had reduced pain levels, and the others with inactive magnets still had a high pain level. However, the doctors nor the patients knew which magnet was active or inactive. The simulation in class, we didn't add or change the numbers, just rearrange a various amount of times up to 10,000. We took means of both sets of data and subtracted them to get the average. In the simulation, it showed that 2,000 out of 10,000 times the mean difference would be 0. The results told us that there wasn't that big of difference in pain of the patients between Active or Inactive magnets. By random chance, when we rearranged over 10,000 times, we never got the number 4.
Shea Craig

The purpose of the experiment was to find out if magnets helped reduce the pain of polio. Patients were unaware if they received an active or inactive magnet. The results were that the active magnets did indeed help. The mean difference of the experiment was 4.05. After a simulation that the class did 10,000 times electronically, we found that a mean difference of 4 hardly showed up. Meaning that the magnets did actually help and it wasn't just a random chance.
Lauren Shane

First, we decided that the magnets helped reduce pain in the polio patients because there were higher numbers in the inactive magnet group than in the active magenet group. The question of whether or not the magnets actually worked was brought up. We decided to assume that all the magnets don't work and to mix the numbers randomly and find the mean of each set and then average of it all. By mixing the active group (consisting of 29 numbers) and the inactive group (consisting of 21 numbers) and then repeating the process 10,000 times the results were basically around 0 almost every time. It would be very rare to get a range of 4 like we did in class. Simulation is important because it shows the likelyhood of something assuming what you found before that was not true. It helps determine how correct your answer actually is and its a great way for people to understand why what you find is correct or incorrect. This simulation showed us that our actual results in our study must mean that the magnets most likely work because of the low chance of getting 4, which we did.
Torrie McSkimming