If you remember from our talk about the design of experiments in experimental research, after a researcher designs the experiment, and then tests it out with their different subjects, their next job is to gather the data and the resulting information from that experiment, and to analyze it. That's what we're going to talk about today.
The goal in analyzing the information is to see if the information either supports or refutes their hypothesis. Remember a hypothesis is their prediction about what they thought would occur. They also want to see if there's a relationship between the variables that are being studied in their experiment. This is all something that we're doing when we're analyzing.
To determine if there's a relationship within the experiment mathematically, scientists examine whether the results are statistically significant. Statistical significance means the likelihood that the variables being studied are related to one another. In other words, one causes the other one to happen. And that the results of the experiment are not the result of chance. In other words, the experimenter wants to know, would these results happen normally, without the changes within the experiment. Would it be something that you would see without whatever independent variables they're looking at. This is also why we use control groups within an experiment. To make sure that the conditions of the experiment are the actual cause of what's happening.
The actual number for statistical significance can vary, depending on the methods being used, and the way they're analyzing the data. But the general rule of thumb is that if the probability of the statistical significance is less than 5%, or p is less than 0.05, then the results are probably due to chance. They just normally occur. It doesn't have anything to do with the experiment. You can look into the reasons and the math behind this result of 5% if you'd like, on your own time. But it's important to remember , if the results are more than 5% then they are statistically significant. And that means that the results show the variables are related in some way. In other words, the experiment was successful.
When analyzing the data, it's also important to realize that there are other influences or results that can affect what conclusions we can draw. So researchers need to consider, first, what effects the participants can have in their own beliefs and ideas about the experiment on the results. This is where we talk about what's called the placebo effect. The placebo effect is when the expectation that a person's behavior will be affected by a treatment within an experiment, leads to the results, and not actually the treatment itself. So it's belief leads to results, not treatment leads to results.
This is the idea that underlies the concept of mind over body, because a person's mind can create physical changes. It's been shown in many different studies that this can actually lead to the results that we see from an experiment, and not from any actual efficacy in the treatment.
This is also why studies use control groups. When they give the control group what's called a placebo. That comes from something that resembles the treatment, and makes the person think that they're receiving the treatment, but actually has no effect in and of itself. A lot of times they use what's called a sugar pill, which looks like an actual pill that they're going to be taking, but it's actually full of something that doesn't have any effect on their bodies. This is able to test if the placebo effect is actually taking place here.
The researcher themselves can also influence the results. This is what's called a self-fulfilling prophecy. A self-fulfilling prophecy is when the expectation of a result leads to, or causes, the result to occur. So just because somebody thinks there's going to be a result-- for example, if you expect a child to be bad-- then you actually treat them badly. You treat them differently than you would normally, and as a result, the child will behave badly.
Or, in terms of an experiment, if you think that there's going to be some kind of results, or some kind of effect on a person, than you might give them small hints that encourage them to respond in the way that you hope For example, you might give them more attention when they're supposed to respond one way. You might be more positive, or your voice might change in a way that influences them to act more in the way that you want. These two things can also affect how we analyze, and they should definitely be in the back of the mind of the experimenters when they're doing their analysis.
Because of the effect that the placebo effect and a self-fulfilling prophecy can have on the results of an experiment, experimenters use different methods to control for these. One of those is what's called the double blind study, or the double blind experiment. This is when subjects are assigned to groups that neither they, nor the experimenter, know who's in each group.
So in the example of testing out a new drug, the participants will be placed into one of two different rooms, and neither they nor the experimenters know which of the rooms has the actual drug that's being tested. It's been kept from them. Either somebody else was used to place it in there that's not related to the study, or they use some kind of random method to determine who would do what. And in this way, the experimenter doesn't know, and so they can't collect the data in a way that leads to a self-fulfilling prophecy. And also, the participants don't know which one of them is actually going to be taking the drug that's being tested. So they can't have a placebo effect necessarily occur.
All studies, when they're doing their analysis, consider previous research and information. They build upon each other's work to create a larger body of knowledge about the subject that they're researching. That's why some experimenters use what's called a meta-analysis. This is a special kind of study or analysis where they look at the results of many different studies on the same subject, and they come to broader conclusions. And this is where we get a lot of our more broader or far reaching theories about different subjects within psychology.
The likelihood that the variables being studied are related, and that the results of the experiment are not the result of chance.
Improvement that is not attributed to the experimental condition, that comes from the subject’s mental state toward that experimental condition.
A fake pill (such as a sugar pill) or injection (such as a saline injection).
When the expectation of a result leads to or causes the result to occur.
An experiment where subjects are assigned to groups and neither they nor the experimenter knows who is in experimental and control groups.
A particular type of study that looks at the results from many different studies on the same subject to come to broader conclusions.