People in control groups are usually given a sugar pill and are observed to assess if any changes occur to their health during the experiments. How do we know that these changes, if any, were the result of taking the sugar pill and not just random changes that might have happened even without taking the placebo? Are there any studies that prove that the placebo effect really exist?
In: 0
There have been quite a number of studies on the placebo effect. We are still trying to figure out how it works and how we can make it better or worse. The idea of using a sugar pill for the placebo effect was the first idea we had. People eat a lot of sugar throughout the day so giving some patients just a tiny pill extra would not be significant. Even the pill under test likely have just as much sugar in it as the placebo. The sugar will therefore have no impact on the research. And some of the first research into placebo had one group take placebo drugs and the other take nothing, and the placebo group consistently did better.
We have now a lot more research into placebo. There are huge industries trying to come up with placebo drugs that match the real ones as closely as possible. And we have research comparing different placebo treatments to each other. For example if the patient is told they are on the placebo it does not work as well as if they are not told, however it still works better then if they get no treatment. And the worse the placebo is, such as bad tasting, having side effects, etc. the better it works. There are even research into placebo surgeries where they put the patient under narcosis, cut their skin, and suture it back together. And these surgeries sometimes works just as well as the real ones.
That’s where sample size and statistical significance come into play. If you only tested 5 people and 3 of them reported effects with the placebo then that could easily be a coincidence. If you test 5,000 people and 3,000 consistently reported effects with the placebo then that’s much less likely to be a coincidence. You can never absolutely prove things like the placebo effect because it’s not something we can actually observe. The “proof” comes from multiple studies showing that there is a correlation and that the correlation is probably not due to random chance.
There have been quite a number of studies on the placebo effect. We are still trying to figure out how it works and how we can make it better or worse. The idea of using a sugar pill for the placebo effect was the first idea we had. People eat a lot of sugar throughout the day so giving some patients just a tiny pill extra would not be significant. Even the pill under test likely have just as much sugar in it as the placebo. The sugar will therefore have no impact on the research. And some of the first research into placebo had one group take placebo drugs and the other take nothing, and the placebo group consistently did better.
We have now a lot more research into placebo. There are huge industries trying to come up with placebo drugs that match the real ones as closely as possible. And we have research comparing different placebo treatments to each other. For example if the patient is told they are on the placebo it does not work as well as if they are not told, however it still works better then if they get no treatment. And the worse the placebo is, such as bad tasting, having side effects, etc. the better it works. There are even research into placebo surgeries where they put the patient under narcosis, cut their skin, and suture it back together. And these surgeries sometimes works just as well as the real ones.
I can’t explain li5 but the answer is „statistical significance“. You can check if the result you’re getting from a test has a chance to come by randomly / by chance / by non related effects – and how strong the chance is. The scientific community talks of significant results when this chance is as low as 5%.
But this value is only the start. Depending on your field and test you make, one should strive for higher values to have more confidence in the result. You can increase this by having a larger sample size for example:
If you have more test cases the chance of getting a specific result randomly or by non related effects is lower.
There have been quite a number of studies on the placebo effect. We are still trying to figure out how it works and how we can make it better or worse. The idea of using a sugar pill for the placebo effect was the first idea we had. People eat a lot of sugar throughout the day so giving some patients just a tiny pill extra would not be significant. Even the pill under test likely have just as much sugar in it as the placebo. The sugar will therefore have no impact on the research. And some of the first research into placebo had one group take placebo drugs and the other take nothing, and the placebo group consistently did better.
We have now a lot more research into placebo. There are huge industries trying to come up with placebo drugs that match the real ones as closely as possible. And we have research comparing different placebo treatments to each other. For example if the patient is told they are on the placebo it does not work as well as if they are not told, however it still works better then if they get no treatment. And the worse the placebo is, such as bad tasting, having side effects, etc. the better it works. There are even research into placebo surgeries where they put the patient under narcosis, cut their skin, and suture it back together. And these surgeries sometimes works just as well as the real ones.
The most direct placebo effect studies involve giving people the same sugar pill and telling one half of the group that the pill does X and the other half that the pill does Y. For example, telling half the group the pill will give them energy (and that group sees a rise in blood pressure and heart rate, and report feeling more energetic) while telling the other half the pill will help them relax (and that group sees a drop in blood pressure and heart rate, and report feeling more calm). Same fake pill, same conditions with the only difference being the “fake” intended effect, which turns out to really happen.
And every good experiment can tease out the random events from statistically significant events.
Also, I think you may misunderstand exactly what the placebo effect is because you asked “How do we know that these changes, if any, were the result of taking the sugar pill and not just random changes that might have happened even without taking the placebo?” While we have ample evidence to establish the placebo effect exists, we do not ascribe any real power to sugar pill itself – it’s the brain itself making the changes due to the suggestion that the sugar pill does something. But the pill itself does nothing – it’s just a convincing charade that “tricks” the patient’s brain into doing the actual work. Remember that every part of your body is connected to your brain, and nearly everything your brain does is inaccessible to you consciously. If your brain wants to slow down or speed up some metabolic function or turn on or off pain or other receptors, etc., it has the tools to do so.
That’s where sample size and statistical significance come into play. If you only tested 5 people and 3 of them reported effects with the placebo then that could easily be a coincidence. If you test 5,000 people and 3,000 consistently reported effects with the placebo then that’s much less likely to be a coincidence. You can never absolutely prove things like the placebo effect because it’s not something we can actually observe. The “proof” comes from multiple studies showing that there is a correlation and that the correlation is probably not due to random chance.
I can’t explain li5 but the answer is „statistical significance“. You can check if the result you’re getting from a test has a chance to come by randomly / by chance / by non related effects – and how strong the chance is. The scientific community talks of significant results when this chance is as low as 5%.
But this value is only the start. Depending on your field and test you make, one should strive for higher values to have more confidence in the result. You can increase this by having a larger sample size for example:
If you have more test cases the chance of getting a specific result randomly or by non related effects is lower.
That’s where sample size and statistical significance come into play. If you only tested 5 people and 3 of them reported effects with the placebo then that could easily be a coincidence. If you test 5,000 people and 3,000 consistently reported effects with the placebo then that’s much less likely to be a coincidence. You can never absolutely prove things like the placebo effect because it’s not something we can actually observe. The “proof” comes from multiple studies showing that there is a correlation and that the correlation is probably not due to random chance.
I can’t explain li5 but the answer is „statistical significance“. You can check if the result you’re getting from a test has a chance to come by randomly / by chance / by non related effects – and how strong the chance is. The scientific community talks of significant results when this chance is as low as 5%.
But this value is only the start. Depending on your field and test you make, one should strive for higher values to have more confidence in the result. You can increase this by having a larger sample size for example:
If you have more test cases the chance of getting a specific result randomly or by non related effects is lower.
Latest Answers