In this video I explain the concept of internal validity and how it acts as a checklist for considering whether an independent variable has been effectively manipulated, a dependent variable has been measured without bias, and a clear pattern can be found in the data. I also mention the importance of avoiding simple yes/no answers and remembering that all of these features are open to scrutiny, criticism, and debate. Finally, I underscore the importance of limiting conclusions to the data that has actually been collected.
Don’t forget to subscribe to the channel to see future videos! Have questions or topics you’d like to see covered in a future video? Let me know by commenting or sending me an email!
Need more explanation? Check out my full psychology guide: Master Introductory Psychology: http://amzn.to/2eTqm5s
Video Transcript:
Hi, I’m Michael Corayer and this is Psych Exam Review and in this video I’m going to talk about internal validity. So internal validity the refers to “inside” the experimental setting, have we done everything correctly so that we have a valid experiment?
So you can think of it as a kind of checklist to go through to look at each of the steps of the experimental process and make sure that everything’s been done appropriately.
So the first thing that we want to look at is the independent variable. And the question we have about the independent variable is whether it has been manipulated effectively. So we want to see that we have a clear definition of what this variable is and that we have also a clear manipulation between our groups, between our experimental group and our control group. So for instance, if I did a study on sleep deprivation and I told you that half my participants slept for 4 hours and half slept for 8 hours that sounds like a good manipulation but if you look at it and you see that actually I just asked people to sleep for that amount, I said “please sleep for 4 hours tonight, please sleep for 8 hours tonight” but I didn’t have any way of ensuring that that actually happened then it wouldn’t be an effective manipulation.
We want to be sure that this is the difference between our two groups. Related to that, we want it to be the only difference between our two groups. The best way to ensure that is to use random assignment so we want to make sure that we’ve had random assignment so that the only difference between our experimental group and our control group is the independent variable and that some other variable hasn’t influenced which group they’re in.
We also want to know about the dependent variable. And what we want to know about the dependent variable is has it been measured in an unbiased way?
So we want to make sure that there is no bias in this measurement. This bias could come in the form of demand characteristics that have caused participants to respond in a particular way or it could be bias coming from the observer. So for instance if a doctor knew that a participant was receiving the medication or was receiving a placebo then this might influence his interpretation of the symptoms, therefore it would influence the results and it would be a biased measurement. So we want to avoid any potential for bias in how we’re measuring our dependent variable.
Lastly we want to make sure that we find a reliable pattern. This is between our manipulation of the independent variable and the effect on the dependent variable. If we don’t have a clear pattern in our data then we can’t draw clear conclusions about what this means.
Now an important point here is that when we consider these things, these are not simple questions. I said it’s a checklist but it’s not the kind of checklist where just go through and say “yes”, “yes”, “yes”, or “no”.
So it’s not going to be simple yes or no answers to these questions and researchers are going to disagree with one another about what should be considered an effective manipulation or what’s considered an unbiased measurement. Or even how they should interpret the results, which statistics should we use to determine whether there’s a clear pattern between these two variables?
These are areas for debate and argument and criticism and that’s a good thing. We don’t want to just settle for a simple yes or no when these are actually very complicated topics.
They’re complicated because there is not one way to do these things. There’s not one way to define a variable, there’s not one way to manipulate a variable, and there’s not one way to measure a variable.
In fact there’s not one way to interpret the findings that we have. There’s a number of statistical methods we might use and researchers will frequently disagree with one other about whether something has been done correctly. Even if we decide that we have internal validity that we’ve done all of these things correctly and we believe that we have an effective manipulation, an unbiased measurement, and a clear relationship we still have to be careful about the conclusions that we draw.
So even when we have internal validity our conclusions must be limited. They must be limited in two ways. The first is they must be limited to the actual variables and that means how we defined the variables, how we manipulated, and how we measured them. We have to keep our conclusions to those definition.
Secondly, we have to make sure that we limit our conclusions only to the sample that we studied. It might be tempting to try to draw sweeping conclusions about all people, or everyone in a particular population but we have to be sure that we limit our conclusions just to the sample that we actually studied.
I hope this helps you to understand internal validity, if so, please like the video and subscribe to the channel for more. Thanks for watching!
