Intro to Research Methods: Study with Video Lessons, Practice Problems & Examples

Roadmap of the Lesson

Hey, everyone. So this chapter is a bit long. So we've built a roadmap for you just to help you refer back and know where you are in the chapter at all times. So we're going to be, kind of kicking off with a little introduction to research methods. We're going to have separate videos for case studies, observational studies, things like naturalistic and laboratory observation will get covered there.

We'll then talk about psychometric assessments, things like surveys, interviews, and psychological tests. We're then going to jump into reliability and validity, basically how do we know if those psychometric assessments are actually working or not. And then we're going to talk a bit about experiments and then we'll talk about validity when it comes to things like experimental design. From there we're going to jump into developmental designs talking about cross-sectional and longitudinal studies. Basically, study designs that help psychologists understand how humans grow and change throughout the lifespan.

And then we're going to wrap up by talking about how psychologists actually evaluate their research findings. So we're going to have some videos covering descriptive statistics that will go over measures of central tendency, things like mean, median, and mode. And then measures of variability, things like range and standard deviation. We're not going to be doing much math, don't worry it's more just how psychologists are going to be using these things. We're then going to talk about correlations which are a very important statistical tool in psychology. But as important as they are, it's very good to know that correlations do not tell us anything about causation.

We're going to have a whole separate video talking about that. And then we'll wrap up by talking about inferential statistics, basically, things like p-values. Alright. So feel free to refer back to this roadmap at any point in the chapter if you're kind of wondering where you are, and I will see you guys in our next video. Bye bye.

Case Studies

This video, we're going to be talking about case studies. A case study is a very in-depth examination of one person or sometimes a small group's experiences, abilities, and behaviors. With a case study, you're going to be gathering as much information as you can about that one person. So their history, their relationships, their symptoms, behaviors, skills, you name it. Researchers can be following the same person for months, even years at times.

So you really get a lot of information with these types of studies. These are typically going to be done on a very unique person or a person in a very unique situation. To give you an example of that, one of the most famous case studies in psychology is the case of Henry Molaison, better known as H.M., who suffered from a very debilitating seizure condition. In 1953, at the age of 27, he had a surgical procedure called a bilateral medial temporal lobectomy. That means his medial temporal lobe was removed on both sides of his brain.

This surgery did cure his epilepsy, however, from this point on, he was unable to form new long-term memories. His memory span was now about 15 minutes long, and he lived with this condition right up until his eighties. Psychologists studied him for decades, and we learned so much about memory from him. He really made an incredible contribution to science, and we're very grateful. Just to give you some examples, despite the fact that he could no longer form new memories, his long-term memory was almost fully intact. He could remember his childhood, his adolescence.

He knew his family. He had skills from before his surgery. He could do crossword puzzles and play bingo. His personality was reportedly largely unaffected. He was a very kind and cheerful person for his entire life.

He could even occasionally learn a new skill, but he couldn't remember learning it. It was very interesting. He taught us that memory is a distinct function that is separate from other cognitive skills. He also taught us that memory is not just one thing; there are all different types of memory and that they seem to interact in different ways. And, of course, that the medial temporal lobe is essential for memory, particularly the formation of long-term memories.

So, that is a very famous case study that happened in psychology. Now, just to give you a sense of some of the common strengths and limitations that we see with case studies, one of the most common strengths is that they really have a high level of detail. Again, a team of researchers could be looking at the same person and putting all their time and resources into one person for months, if not years. So, you normally get a wealth of information. One other thing about case studies that can be really useful is that we can use them to research topics that practical or ethical considerations might normally prevent us from studying.

Obviously, we could never have a randomized control trial where we give 50 participants bilateral medial temporal lobectomies just to see what happens to their memory. Ethically, you can't do that. But if this happens to a person who is then able to consent, which he was, and he also had family involved in the consent process, well then we can study that one person and maybe learn some new things. So they can be useful in those situations. Now, in terms of limitations with case studies, sometimes information can be missing or hard to interpret.

For example, you could be doing a case study on a child. Maybe their history is not fully known. Maybe the person or child doesn't have the verbal skills to fully articulate what is happening to them or to accurately self-report. It can be anything. So, information can be hard to interpret at times.

One of the most important limitations to be aware of with case studies is that you cannot use them to make assumptions or generalizations about behavior. So a moment ago, I was gushing about how he taught us a ton about memory, and he did. But it was not the case that we studied and then said, "Alright. That's how memory works. Let's go home."

Instead, we studied him, got a bunch of ideas, and then scientists took those ideas and then did all kinds of controlled experiments in a laboratory with animals and humans to verify and validate those findings. And so we can get a lot of ideas from case studies, but we cannot just assume that what we learn from them will generalize to other populations. One other note that ties into that point is that they are very difficult to replicate, of course. These are unique people or people in unique situations. Odds are it's only going to happen one or two times.

You can't really replicate your findings. And so because of that, we do not want to be making generalizations about behavior. Alright. So those are case studies in a nutshell, and I'll see you guys in our next video. Bye bye.

Intro to Research Methods Example 1

Okay. So what is the main advantage of using a case study in psychological research? Now, without reading through all of these verbatim, we know that the idea behind a case study is to get really rich, in-depth information on a single person or small group of people who are in a unique situation or are sort of a rare case based on their life circumstances or their biology. Based on that, it looks like our answer is going to be b. Now, looking quickly at these other ones as a reminder, case studies, you know, we typically don't have a lot of control over the variables, especially the variables that might have caused that unique situation to happen to the person in the first place.

So that's why a is incorrect. We are typically not looking at large samples with case studies. To get that really rich, in-depth amount of information, we're usually only going to be studying one person or a very small group of people. And then finally, the data from a case study itself cannot necessarily be generalized to the broader population. The information that we learn in a case study can then be tested in different ways in laboratories and verified that way, but we don't want to just take the data from a case study and then automatically assume that it can be generalized to other people.

We want to verify that through other methods. So based on that, our answer here is b.

Which of the following statements best describes a case study?

• A. A psychological research method that involves manipulating one variable to discover its effect on another.
• B. A statistical technique that examines the direction and strength of a relationship between two variables.
• C. A research method that involves utilizing archived data.
• D. A research method that involves an in-depth examination of a single individual or small group of people.

Observational Studies

In this video, we're going to be talking about observational research. This is a method in which a researcher systematically observes and records behavior. This can be very useful because people do not always accurately self-report on their behavior. Either they are not able to, not self-aware enough to, or sometimes people just lie. If you observe the behavior directly and can make objective judgments about it, you don't have to worry about participants not self-reporting it very accurately.

Typically, observational research is done without or with minimal interference from the researcher. The idea here is to get the most natural behavior possible, and of course, people usually don't behave naturally if they know that they're being watched. There are two main types of observational research. We have naturalistic observation, which is an observational method where psychologists observe the behavior of their subjects, either humans or animals, in their natural environment. If you're studying chimpanzees, you might be out in the jungle of Tanzania.

If you're studying humans, you might be going out into a university, into a grocery store, into a bar, to actually observe the behavior. Typically with naturalistic observation, you as a researcher would want to really blend in. You don't want people to know that they're being observed. Again, the idea here is to get the most natural behavior within the natural environment as possible. In contrast, we also have laboratory observations.

This is a method where psychologists observe the behavior of their subjects in a controlled environment. You would bring your participants into your lab, give them a certain task to complete, and then you would usually leave the room so they're alone in the room and you're either watching them through a two-way mirror or perhaps you have a camera set up in the room to record, the person or the interaction. Sometimes there are even hidden cameras to kind of take that pressure off the person. The benefit of laboratory observations is that you can control your subject pool and the tasks that they are doing.

For example, if you want to study marital conflict, you can have a married couple come into your lab and have a disagreement right there in the room. These are very useful in that sense. Now, we're going to get into the strengths and limitations of observational research and these strengths will apply to both naturalistic and laboratory observations. One general strength of this method is that it typically provides very rich, detailed data that is very useful for describing behavior. To give you an example of what I mean by that, back in grad school, I did a ton of laboratory observations and my favorite one was when we would give 3 to 4-year-old children a very challenging puzzle box and we would tell them if you can solve it, you can have a prize at the end.

We'd leave them alone in the room for 5 minutes and just record their behavior. With that task, we could assess things like children's problem-solving skills, their logical thinking skills, their creative thinking, how they use tools in their environment to help them solve a problem. But we could also use that task to assess how young children respond to frustration because most preschoolers get very angry when they can't solve a problem or they're not going to get a prize. So we'd observe how they respond to frustration, how they self-soothe, how they self-regulate, etc.

Just with that one task, we could assess 5 or 6 different aspects of a child's development or skill level. That's what I mean when I say that a well-designed observational study is going to get you very rich data that is useful for describing behavior in multiple ways potentially. These, of course, are also a really good way to study more natural behavior. You don't have the chance of a person altering their responses on a questionnaire to look more socially desirable. This is especially true of naturalistic observations.

In terms of limitations for naturalistic observations, you are going to have minimal control over your environment and your subject pool. For example, if you're doing a naturalistic observation in a grocery store, there's a chance that maybe the customers coming into the store that day don't demographically represent your area or your county particularly well. If you're studying chimpanzees, maybe there's a predator in the area that day, and the chimps just don't want to show up and you have no control over that. With laboratory observations, one thing to be aware of is that subjects can alter their behavior when they know that they're being watched, and this is especially true for adults.

Although, just to give you my personal experience, from what I've seen, adults will act a little bit differently for the first couple of minutes. But once they get comfortable in the lab and kind of get into the flow of whatever task you've given them, you'd be surprised how naturally people will act even in a laboratory setting. Alright, those are observational studies, and I will see you guys in the next video. Bye-bye.

Intro to Research Methods Example 2

Okay. So which of the following is a key characteristic of naturalistic observation? Based on looking through these answer choices, I think we're going to be focusing on options b or c because those discuss observing the behavior of participants. Option b is about observing behavior in a natural environment, and option c involves observing behavior in a controlled environment. Observing behavior in a controlled environment is much more characteristic of a laboratory observation. What option b describes, observing the behavior of participants in their natural environment, is very characteristic of that naturalistic observation.

Ideally, you go out into the person's environment, whether that is a grocery store, a hospital, a university, or could be out in the forest if you're studying animals, for example. You want to get into that natural environment so that they do not know they're being studied and can behave in the most natural ways. So, our answer here is going to be b.

Dr. Anderson studies how married couples deal with conflicts. She plans to bring couples into her lab and leave them alone while they discuss 3 topics that they commonly disagree about. Dr. Anderson will be filming these discussions and then coding the behaviors of each couple. What research method is being described here?

• A. Naturalistic observation.
• B. Case study.
• C. Laboratory observation.
• D. Experiment.

Psychometric Assessments

This video we're going to be talking about some different types of psychometric assessments. So these are measurement tools that we use to quantify psychological constructs and these are used in a variety of settings. We use them in research, in clinical settings, educational settings, you name it. So there are 3 main types of psychometric assessments. The first is probably the most common measurement tool that we use in psychology period, which are surveys and questionnaires.

So we use these to gather self-report data. And what's great about them is that you can use them for pretty much any topic. You can ask people about their behaviors, their skills, experiences, attitudes, opinions. They are incredibly versatile. Next up we have interviews.

Interviews are conversations that are designed to gather in-depth information on a specific topic and interviews can be structured or unstructured. So in a structured interview, the interviewer comes in with a specific set of questions really probing specific topics. Depending on the person's response, they may have a specific probing question to get more information about that. Whereas in an unstructured interview, it's going to be much more like a normal conversation, very kind of loose and flowy. So interviews can kind of fall on like a spectrum of very structured to very unstructured.

And then finally we have psychological tests. So these are standardized tests and standardized just means that they have a uniform set of procedures for administering the test and scoring the test. So all of that is done in a very standard way and typically people have to be trained in order to use these types of assessments. And so we can use standardized tests to measure all different types of constructs, cognitive skills, language skills, personality traits. We use them a lot when we are diagnosing psychopathologies, like depression, schizophrenia, etc.

So they are very useful for those types of settings. Alright. So those are our 3 types. Now to kind of go over some of their strengths and limitations, starting with surveys, the biggest strength of the survey is that they are efficient and cheap which is really the main reason that they are the most popular tool in psychology. You know, you can hand out a 1,000 surveys with the click of a button.

They are wonderful in that way. They're also very good for sensitive information. So surveys can feel very nice and anonymous for participants. You know, you don't have to talk to a researcher face to face, so people may be more willing to admit to more undesirable or embarrassing behaviors or thoughts in a survey versus like in an interview, for example. With interviews, the main strength is going to be the depth of information that you are able to gain which is obviously going to be much more than what you could get with a survey alone.

So, you know, with a survey, you might get that a person feels anxious all the time. That's kind of all you get. Right? Whereas with an interview, you could say, like, wow, all the time? Like, do you feel anxious in the morning, in the afternoon, in the evening?

What's making you feel anxious at each time of day? Is it the same? Is it different? Etcetera. So, you know, much more in-depth information from those interviews.

And then with tests, obviously, that standardization and all that administrator training typically and ideally is going to provide you with more consistent and reliable results. Now in terms of limitations, and these limitations kind of apply to all 3 of these, just in varying degrees, one of the biggest limitations is going to be social desirability, which is when people are basically trying to alter their responses in a way that makes them look more socially desirable or makes them look like a quote-unquote better person. And that's something that you'll see very commonly with surveys and interviews, especially. One other thing to note is that these are going to be dependent on things like reading level and comprehension. You know, we often kind of assume that adults have good reading comprehension but that's honestly not always the case.

So it's something to just be aware of when you're giving out a lot of surveys. This can also be an issue with more standardized tests because again you often have to, like, give the instructions in a very specific way, and you can't go off script. So if the participant hears that and then doesn't understand the instructions, sometimes there's not much flexibility in how you can explain it to them. So that can be a problem as well. And then one last thing to be aware of is that sometimes the wording of questions can impact responses.

So, you know, if you had a question that was like, do you support increasing taxes for public spending? People are probably going to hear increasing taxes and be like, no thank you. Right? But if you said something along the lines of, you know, do you support investing in our local education, hospitals, recreation centers, you might get some more, you know, more positive responses on that. Right?

You know, if you ask a person, have you engaged in binge drinking in the past week? A lot of people don't actually know the definition of binge drinking, so you might get kind of low responses. But if you said, have you consumed 5 or more alcoholic beverages in 1 night in the last week? You might get, you know, a little bit more responses from that. So just be aware that the way that a question is worded can definitely impact the answers that you get, so you want to be making sure that you are asking the question in a way that will get you the most accurate answer.

One last thing to note is that all of these different types of assessments should be tested for something called reliability as well as validity. And we're actually going to talk about that in our next concept video, so I will see you there. Bye-bye.

Intro to Research Methods Example 3

Okay, so which of the following is an advantage of using surveys in psychological research? So a reads that they allow researchers to examine phenomena in a highly controlled environment. And while you certainly could use a survey in a highly controlled environment, they're actually useful because you don't need to be in a controlled environment. You can administer a survey to anybody, anywhere, which is really, really useful. So a is not really characteristic of surveys themselves.

B reads that they are not subject to social desirability but that's not true. We know that surveys can be subject to social desirability. People can and do alter their answer choices in order to appear more desirable on surveys all the time. So b can be a potential issue with survey research. C is that they provide in-depth detailed information on unique or rare phenomena.

Now surveys can give us a lot of very useful information, I don't want to downplay that, but they are not the method that is best known for giving us in-depth detailed information. You can only put so many questions on a survey, you can only have so many answer choices, and you can lose some nuance and detail in people's answers because of that format. So c is not really characteristic of surveys either. And then d reads that they allow researchers to collect data from a large number of participants relatively quickly. And that is a wonderful thing about survey research.

You can collect data from thousands of participants in a day, like, with the click of a button. It's extremely useful. So d is a huge advantage of using surveys in psychological research. And there you go.

Which of the following is true about psychological tests?

• A. They are standardized, meaning they are administered and scored using specific procedures.
• B. They only measure cognitive skills (e.g., IQ, memory, executive function).
• C. They tend to be highly subjective.
• D. They allow the researcher to improvise and be flexible with each participant.

Reliability and Variability: Psychometric Assessments

In this video, we're going to be talking about reliability and validity and how these concepts apply to our psychometric assessments. So first up, reliability is basically just the degree to which an assessment provides consistent scores across time and place. So, if we have a survey and we give it to participants in January and then again in December, are we going to get consistent scores? If people take our survey in the lab versus at school versus at home, are we going to get consistent scores? That's the idea behind reliability.

Now one important type of reliability that you may need to know is something called test-retest reliability, which, like I just alluded to, is the measure of the consistency of a test over time. For example, if we were to develop a new survey, we may have people come into our lab on a Monday, take the survey, come back into our lab on Friday, take the survey again, and what we are looking for, ideally, is that their score is the same at both time points. This would be evidence of good test-retest reliability. So that is reliability. Now, validity is essentially the overall appropriateness or accuracy of a conclusion.

And I do apologize for how vague that definition is, but there are many different types of validity and so the overarching definition tends to be a bit vague because we're trying to encapsulate so many different things. You don't have to know many different types of validity at this level, but one that I do recommend you are aware of is something called construct validity. So, construct validity is basically just the degree to which an assessment accurately measures what it is intending to measure. So for example, if you were trying to create an assessment for depression, you want to make sure that it actually measures depression and not anxiety or psychosis symptoms. Right?

That's very important. So one way that you can assess construct validity of a measure is by basically thinking about some other variables that should be theoretically related to your construct and seeing if they actually are. So for example, if we were developing a new survey to test extroversion, how much a person feels energized by being in social situations, theoretically, that should be related to how many hours people spend socializing. So what we would hope to see is that our extroversion survey is positively correlated. It is related to how many hours people spend socializing. And this would give us evidence that we have good construct validity.

We do seem to be measuring what we are intending to measure. You can also do the complete opposite; you can think of a theoretically unrelated construct and hope that it does not relate to your measure. So for example, if we're doing our extroversion survey, that should not be related to something like reading speed. Right? That has nothing to do with extroversion. So what you would hope is that there is no correlation or no relationship between your survey and this variable, and that would also give you evidence of good construct validity. So that is just one way to assess the construct validity of a measure. So, broad strokes here: reliability is basically telling you, you know, the degree to which an assessment is going to give you consistent scores.

Right? And then validity is getting more at, you know, are we assessing this variable in an accurate way? Can we really trust this measure to be assessing this variable is kind of the idea there. Alright. So that is how reliability and validity apply to psychometric assessments, and I will see you guys in the next one.

Bye bye.

Intro to Research Methods Example 4

Okay. So a psychologist administers a personality test to a group of participants and then administers the same test to the same group exactly 2 weeks later. She finds that there is a strong correlation between the two tests. This is an assessment of the test's what? So just kind of stepping back and thinking about that for a moment, what we're basically looking at here is whether this test can provide consistent and reliable results across time, which is exactly what reliability is.

So, our answer here is going to be reliability. Keep in mind that validity is more concerned with the accuracy of the conclusion of the measure, like, can we really trust that this measure is giving us accurate results? Which is not really what we're looking at here. Here, we are literally just looking at whether we can get consistent results across time, which is very much what reliability entails.

Keep in mind that both response bias and social desirability is when people answer questions in a biased way for some type of reason either they're trying to look more socially desirable or for another more general reason so that's not really what's being assessed here. So, here we are looking at the test's reliability.

What is construct validity?

• A. The degree to which the test produces consistent results over time.
• B. The degree to which the test is intuitive for researchers to use and score.
• C. The degree to which the test accurately measures the construct it is intending to measure.
• D. The ability of the test to precisely predict behavior.

In which of the following situations might test-retest reliability be particularly important?

• A. A workplace is giving out a survey to examine employee happiness in the summer of 2024.
• B. A research team is developing a new test to measure a stable personality trait.
• C. A research team is developing a new test to measure current mood in patients with mood disorders.
• D. A research team is developing an assessment to measure acute stress response during major life events.

Experiments: Part 1

This video, we're going to begin talking about experiments. An experiment is a controlled test where one variable is manipulated to discover its effect on another variable. The variable that is manipulated is our independent variable, which we theorize will have an effect on our dependent variable. The independent variable is manipulated, and the dependent variable is measured. In an experiment, we place our participants into groups, and we do that through a process called random assignment.

In random assignment, we place our participants into their condition or group at random. The idea behind this is that we're trying to increase the equivalency between groups. We want our groups to be as equal as possible on any potential confounding variables because we want our groups to be the same so that the only difference between them is the manipulation of the independent variable. When we get our result, we can very confidently say, "The only difference between these two groups was the independent variable, so that must be what caused our result to happen." If there is no equivalency between the groups, we can't make that conclusion as definitively. For example, if the average age of your first group is 50 and the average age of your second group is 20, when you get your result, you could say, "Maybe it was an age effect. Maybe they interpreted the instructions differently." If one group is mostly male and one group is mostly female, you could say, "Maybe it was the independent variable, but maybe it was a gender effect that we found." This can be true of any variable.

You don't want one group to be more motivated or more tired, or have a higher IQ on average. By randomly placing our participants into groups, we're going to be increasing the equivalency, ensuring equal numbers of males and females, even distribution of age, etc. When our groups are equivalent, we can be more confident that what we are finding is the effect of our independent variable on our dependent variable. With an experiment, you are going to have an experimental group or potentially multiple experimental groups. This is going to be the group or condition in which the independent variable is manipulated or is present.

The experimental group is where we are actually going to be observing the effect of the independent variable on the dependent variable. We're also going to want to have a control group, which is a group or condition in which the independent variable is absent. The idea here is that we want to be measuring our dependent variable as it would occur naturally without the independent variable. At the end of our study, we can compare these groups, and if they are different, we can assume that the independent variable must have had an effect on the dependent variable. To give you an example, let's just say we have a hypothesis that listening to classical music while studying is going to increase test performance.

The classical music, or the auditory environment of your studying, is our independent variable because we believe it's going to have an effect on test performance, which is our dependent variable. For our method, we're going to have students come into our lab. They will study a brand new topic that they've never learned before for 1 hour, then they're going to take a test, and we will be assessing, of course, their test performance. So we are going to have 3 experimental groups to make this a very scientifically rigorous study. In group 1, we're going to be directly testing our independent variable and will have that group listening to classical music.

Okay, but then someone might say, "Is it classical music, or is it any kind of music?" That's a great point. Let's address that with group 2, which will listen to pop music. Now we're kind of getting at, "Is there something really special about classical music as I'm hypothesizing, or is it just music in general?" But then someone could say, "Maybe it's not music. Maybe it's just noise." That's a good point. Let's have group 3 listen to ambient cafe noises. There's no music present, but we have a sound environment at an equal decibel level to our musical conditions. Then we'll have our control group, of course.

Group 4 here is going to have some noise-cancelling headphones, and they will hear no sound. So there is absolutely no auditory environment happening for group 4. We're just going to see test performance under naturally occurring conditions. Ultimately, we will randomly assign people into these groups, and then we're basically just statistically comparing the test performance of these groups. What we're hoping to see is that group 1 has a statistically significant higher test score than the other three groups, which would give us evidence for our hypothesis being correct. Alright. So that is our little example of an experiment. We're going to keep talking about experimental design in our next video, so I will see you guys there. Bye bye.

Intro to Research Methods Example 5

Okay. So Doctor Sullivan is studying the effects of a new drug on anxiety symptoms. She plans on having one experimental group who will receive the new drug and one control group who will receive a placebo or a sugar pill. She decides to put her high anxiety participants into the experimental group and her low anxiety participants into the control group.

What is the main problem with this experimental design? Let's analyze this more thoroughly. She is testing the effects of a new drug on anxiety symptoms, which makes sense. She plans on having one experimental group who is going to receive the drug and then one control group who is getting a placebo. All of that makes sense so far.

She decides to put her high anxiety participants into the experimental group and her low anxiety participants into the control group, and that is concerning. We now know for a fact that the experimental group and the control group are fundamentally different on a very important variable, and that's not good science. Random assignment should have determined the grouping, which is completely lacking here. Therefore, the issue with the experimental design is that it completely lacks random assignment, which compromises the scientific validity of the findings.

Reviewing the options quickly, option a mentions the lack of manipulation of the independent variable, but this is not the case, as she has a clear independent variable: the drug. The presence of an experimental group and a control group to test the independent variable is part of a good experimental design, so option a is incorrect. Option d, which concerns identifying the independent and dependent variables, is also not a problem. It is clear which variable is the dependent variable and which is the independent variable.

Option c suggests that placebo conditions are generally considered ineffective; however, this is untrue. Placebos can have very strong effects which are useful in control conditions, so option c is also not applicable.

The main issue with the design is the complete lack of random assignment, a fundamental component in conducting scientifically valid research. Therein lies the primary flaw in Doctor Sullivan's experimental design.

Why is random assignment an essential part of good experimental design?

• A. Random assignment ensures that the sample accurately reflects the population.
• B. It ensures that the experimental group will show the intended effects.
• C. It increases equivalency between groups, reducing the odds of confounding variables affecting the results.
• D. Random assignment ensures that the control group will not receive the independent variable.

Experiments: Part 2

Okay, so let's finish up talking about experiments. Now in order to make our experimental design as strong as possible, ideally it should be double blind or at the very least single blind. So double blind refers to situations where both the participant and the researcher who is in contact with them are unaware of what condition the participant is in. So if both the experimenter and participant are blind to the condition then that is called a double blind study. Now that is not always possible and so we can also have a single blind study which is when either the participant or the researcher is unaware of the condition.

So, for example, if we had a situation where the experimenter knew the condition but the participant was blind to it, that would be called a single blind study. Now the reason that we like to add this layer of blindness whenever possible is because it can reduce both experimenter effects and demand characteristics. And these terms refer to situations where basically people are altering their behavior in some way because they think they know what condition that participant has been placed in. So if an experimenter does this it's called an experimenter effect and if the participant does it it's called a demand characteristic, but these terms refer to a very similar phenomenon where basically just knowing the condition makes the person behave in a different way even if it's more just like kind of subconscious. It just kind of alters their behavior and we don't want that.

We want the independent variable to be driving the participant's behavior, not anything else, right? So that is blindness. Now one last very important thing to know about experiments is that experiments can be used to establish a cause and effect relationship among our variables and they are the only method in psychology that can do this. Okay? And there are two main reasons why experiments can do this.

So the first is that experiments establish something called temporal precedence which sounds very fancy but basically temporal just means time and precedence is basically saying, like, what comes first in time? So in experiments, the manipulation of the independent variable always comes before the measurement of the dependent variable. So basic logic just dictates that we can confidently say it is possible for this variable to cause this variable. You don't always get temporal precedence with like observational studies or survey research or case studies but you do always get it with experiments. Now experiments also isolate the independent variable by eliminating confounding variables more so than any other psychological method.

So by doing things like having random assignment, making your groups equivalent, by adding blindness, having these controlled laboratory settings, all of those things eliminate confounding variables and isolate the IV so we can confidently say only the independent variable could be causing changes in the dependent variable. Again, other methods in psychology just do not do that as well as experiments. So in terms of the strength of experiments, like we just talked about, that level of control that you're going to get with an experiment is basically unmatched among any other psychological method. So they have a great level of control, and again, they are the only psychological method that truly allows us to establish causality among our variables. K.

So with an experiment, you can say the independent variable caused the dependent variable to happen. With other types of methods, observational research, you know, survey research, you can establish a relationship. You can say this IV is related to this DV, but you cannot say that it caused that dependent variable. So that is the difference there. Now in terms of the limitations of experiments, the situation that you have your participants in can feel a little bit artificial sometimes.

And because of that, sometimes the results don't necessarily generalize to real world behavior as much as you might want them to. So, you know, for example, if you were studying competitiveness and you have people coming into your lab and doing, like, a little click and point computer game, you know, well, the amount of competitive behavior that they display in your lab probably isn't the same as the amount that they would display when they're, like, playing against their rival soccer team or when they're in the middle of a tennis game or whatever. Right? Though because that situation is kind of artificial, the result might not generalize. This is not always an issue but it can be so just kind of keep an eye out for it.

And I also have a note here that poor experimental design can compromise data. So it is true that experiments in general can establish causality. However, if an experiment is not well designed, then it cannot establish causality. So if we don't have random assignment, if there is no level of blindness, if we're not eliminating confounding variables, then it would lose the ability to really establish that cause and effect relationship. So when you are evaluating experiments, be critical of them.

They need, you know, they need harsh criticism because if they're really going to make a causal claim, we have to make sure that they have really good design. Alright. So those are experiments, and I'll see you guys in our next one. Bye bye.

Intro to Research Methods Example 6

Okay, so what is the purpose of making an experiment single blind or double blind? When we make an experiment blinded in some way, the participant and/or the researcher does not know what condition the participant is in, and basically that reduces the odds of them behaving in a biased way. Remember, if our experimenter behaves in a biased way, that's called an experimenter effect, and if our participant is behaving in a biased way, that could be the result of demand characteristics, basically them guessing what condition they're in and what the independent variable and dependent variable are, and then acting in a way that they think the experimenter is going to want them to act. If they don't know what condition they're in, the odds of that are greatly reduced. And so, based on that, our answer here is going to be b.

By making an experiment either single blind or double blind, it can reduce potential bias from the participants and/or researchers, which could influence the results, and we don't want that happening. So, our answer here is b.

Out of all the research methods used in psychology, why are experiments particularly useful at making causal claims?

• A. Experiments involve the use of large samples and utilize reliable and valid measures.
• B. Experiments isolate the independent variable and control for confounding variables.
• C. Experiments are designed so that the independent variable always precedes the dependent variable.
• D. Both b and c are correct.

What is a key feature of a double-blind experiment

• A. The researcher knows which group the participant is in, but the participant does not know.
• B. The participant knows which group they are in, but the researcher does not know.
• C. Neither the researcher nor the participant knows which group the participant is in.
• D. Both the researcher and the participant know which group the participant is in.

Validity: Experimental Design

In this video, we're going to circle back to the concept of validity. So remember, previously we had talked about validity as it applies to things like psychometric assessments. But when we are thinking about the validity of a study's design, or for example, an experimental design, we consider two different types of validity, and those are internal and external validity. Internal validity is essentially an indication of how well a study is controlling potentially confounding factors. So basically, you know, how well is this experiment designed, and how confident can we be that the result that we're seeing is the effect of the independent variable on the dependent variable.

Some things that can, kind of boost or increase internal validity in a study are things like random assignment, which we know is going to increase equivalency among groups, give us, you know, more control there. Using things like reliable psychometric assessments or, you know, assessments with good construct validity. So if we're just using really well established tools in the field, as well as things like making an experiment double-blind. All of these things are going to be basically reducing confounding variables and that can make us feel more confident that what we're seeing again is the effect of the IV on the DV.

Whenever I think of internal validity, I always just think, like, internal as in, like, inside of the experiment. So is everything internal to, or everything inside of the experiment working the way that it should? Is there a nice strong experimental design? And if so, you have stronger internal validity. Now, in contrast, we have external validity, which is basically the extent to which the findings can be generalized to the population.

So basically, you know, can we take the result from our sample and confidently say that it applies to the entire population? And external validity can be strengthened by things like random sampling. Remember with random sampling we're going to end up with a sample that better reflects the population typically, as well as creating, you know, study conditions that reflect the real world as much as possible. So to give you an example of that, like, I've worked with kids for most of my career, and I worked in one lab where the lab looked like a doctor's office almost. It had, like, tile floors and fluorescent lights and all that, and kids would be really nervous when they came in.

Sometimes they'd act a little bit weird. But in a different lab I worked at, we made our lab look like someone's living room. So it had, like, a plush carpet. It had big comfy couches and chairs and potted plants and pictures on the walls, and kids would walk in and be like, whose house is this? Like, where are we?

But they'd be really comfortable. You know, they'd get up on the couch and just be chit chatting, and they act much more naturally because they were so much more comfortable. And because they were acting naturally, we could kind of be a little bit more confident that our results were going to generalize to how they would act in the real world. Right? So doing these types of things are going to boost your ability to really be confident that your findings can be applied to the broader population.

Now I do want to be really clear that not every single study necessarily has to have good internal and external validity all the time. Some studies are a little bit stronger in their internal validity, some studies are stronger in their external validity. There are many different reasons, you know, why you might be stronger in one versus the other. The research might have different purposes, different topics. Sometimes it's easier to have internal versus external.

I just want to be clear that, you know, you don't have to be judging every study and expecting them to have, like, incredible internal and external validity all of the time. There are plenty of situations where it might be a little bit more imbalanced and that's totally okay. Alright. So that is validity as it applies to experimental design, and I'll see you guys in the next one. Bye bye.

Intro to Research Methods Example 7

Alright. Dr. Lau studies the relationship between sleep and cognitive performance. She recruits undergraduates from her university to participate in an experiment where they are randomly placed into four groups that get 8, 6, 4, and 2 hours of sleep respectively, and then they are given a set of cognitive tasks to complete. She finds that there is no difference in cognitive performance between participants who received 4, 6, or 8 hours of sleep.

What might be a concern about the external validity of Dr. Lau's study? So when we're thinking about external validity, we're thinking about whether these findings can be applied to the broader population. Can we generalize these findings? In Dr. Lau's study, it sounds like she had a perfectly fine experimental design, but one thing that might make it hard to generalize to the population is that her sample consisted of undergraduates.

Undergraduates are a demographic who notoriously are fairly young, fairly healthy, and are very accustomed to having very little sleep, and usually engaging in cognitive tasks on very little sleep. For that reason, the results might not generalize to people of all ages and all stages of life.

Therefore, the answer here is going to be that, because her sample consisted mainly of undergraduates, the findings may not generalize to the broader population.

Which of the following practices might improve the internal validity of an experiment?

• A. Random assignment.
• B. Providing participants with feedback about their performance.
• C. Using a representative sample.
• D. Increasing the sample size.