Introduction to Psychological Research Methods
Psychological research methods form the bedrock of psychology, distinguishing facts from myths through systematic study. Psychology follows the scientific method, emphasizing an evidence-based approach rather than intuition.
The Scientific Method in Psychology
- Hypothesis Formation: An educated guess based on prior research.
- Deductive Reasoning: Using hypotheses to make predictions.
- Observation and Experimentation: Collecting empirical data to test predictions.
- Theory Development: Repeatedly supported hypotheses strengthen into scientific theories or laws.
- Ongoing Process: Science is continuous, with no definitive end.
Types of Psychological Research
Descriptive Research
- Purpose: To describe phenomena without explaining causes.
- Methods:
- Observation (e.g., coding aggressive behavior in playgrounds)
- Surveys and Interviews (gathering large-scale data quickly)
- Case Studies (in-depth analysis of a single subject or phenomenon)
- Strengths: Provides detailed descriptions and naturalistic data.
- Limitations: Cannot explain why phenomena occur.
Correlational Research
- Purpose: To identify relationships between variables.
- Correlation Coefficient (r): Ranges from -1 to 1 indicating strength and direction.
- Positive correlation: Both variables increase together.
- Negative correlation: One variable increases as the other decreases.
- Important Note: Correlation does not imply causation.
- Applications: Identifying risk factors (e.g., parental dieting and eating disorders).
Experimental Research
- Purpose: To determine causation by manipulating variables.
- Design: Random assignment to experimental and control groups.
- Independent Variable: The factor manipulated (e.g., stress level).
- Dependent Variable: The outcome measured (e.g., parental harshness, child rebellion).
- Control Group: Provides baseline for comparison.
- Ethical Considerations: Experiments must avoid harm and respect participant rights.
Validity and Reliability in Research
- External Validity: Generalizability of findings to real-world settings.
- Internal Validity: Confidence that the independent variable caused the observed effect.
- Bias and Logical Errors: Must be minimized through random assignment and careful design.
Common Research Biases
- Experimental Bias: Researcher’s expectations influencing outcomes.
- Demand Characteristics: Participants guessing the study’s purpose and altering behavior.
- Research Participant Bias: Social desirability affecting honesty in responses.
- Placebo Effect: Participants’ expectations causing perceived effects.
- Mitigation: Double-blind experiments where neither participants nor experimenters know group assignments.
Sampling and Generalization
- Population vs. Sample: Research is conducted on a sample representing the larger population.
- Representative Sampling: Ensures sample demographics match the population to avoid bias.
- Random Sampling: Increases likelihood of representativeness, especially with large samples.
Research Settings
- Laboratory: Controlled environment allowing precise manipulation.
- Naturalistic: Observing behavior in real-world settings for ecological validity.
Data Analysis in Psychology
- Descriptive Statistics: Summarize data using mean, median, mode, range, and standard deviation.
- Inferential Statistics: Determine if results are statistically significant and not due to chance.
- Statistical Significance: Typically set at p < 0.05, indicating 95% confidence in results.
Ethical Considerations in Psychological Research
- Participant Rights: Informed consent, confidentiality, and debriefing.
- Historical Context: Past abuses (e.g., Tuskegee syphilis study) highlight the need for ethical oversight.
- Institutional Review Boards (IRB): Ensure research complies with ethical standards.
- Deception: Allowed only with strong justification and followed by debriefing.
Applying Research Methods: Case Study on Conspiracy Theories
- Descriptive: Observing online communities to understand behavior.
- Correlational: Examining relationships between belief in conspiracies and factors like isolation or political affiliation.
- Experimental: Ethically testing interventions to reduce conspiracy beliefs.
Conclusion
Psychological research methods are essential for building reliable knowledge about human behavior and mental processes. Understanding these methods, their strengths, limitations, and ethical requirements empowers critical evaluation of psychological findings and supports ongoing scientific progress.
For a deeper understanding of the importance of research in psychology, check out Why Research is Crucial in Psychology: Understanding Scientific Inquiry.
To explore the ethical dimensions of psychological research, refer to Ethics in Research: Deception, Animal Studies, and Institutional Oversight.
For insights into common misconceptions in psychology, see Understanding Psychology: Key Concepts and Common Misconceptions Explained.
If you're interested in the various research approaches, visit Comprehensive Guide to Research Approaches in Psychology.
Lastly, to understand correlation, sampling, and experimental bias in research, check out Understanding Correlation, Sampling, and Experimental Bias in Research.
Hey folks! Welcome to lecture two. Today, we're talking about uh as it says on the tin here: uh psychological research methods. This is one of the most important lectures that we're going to have over the course of the semester because this represents
how we lay the Bedrock for what we're going to consider fact and Theory uh in in Psychology. um Whenever we talk about uh sort of a fact of psychology it was arrived at using one of these methods or more of these methods one or more of these methods um and and so it's important to
understand that uh when we talk about psychology being the systematic study of mind and behavior these are the systems. This is the systematic uh study part um which is again the Bedrock of Psych right um this is how we distinguish uh facts about
people from myth from folklore from folk wisdom uh things that might be useful sometimes right but uh this is this this is where we get closer to uh being able to uh to call what we are saying fact. So let's Dive In
So uh psychology follows the scientific method. um It's not uh what determines whether something is a science is not uh the subject matter right it is the approach that you use right um this is this comes up in my clinical work all of the time because I I'll direct clients to do
experiments on on their own emotions on their own experience on their own experiences right and so that brings that brings it closer to a science as opposed to uh sort of going with their gut right that kind of thing but to be even more specific about what we're focused on here today
um the scientific method is is the thing that many of you learned really early in your uh in your education um which is uh this this process where we uh have a this sort of circular relationship between we have a hypothesis or an educated guess right
it's important educated guess I have done research based on This research it seems like the answer is blank right and if the answer is blank right this is the deductive part here if the answer is blank then right uh rather if the answer is you know x y z right we can use deductive reasoning
right to uh to make predictions right so if the Earth is uh orbiting around the sun right um ooh better idea the if the if the Earth is turning right that's our uh that's our hypothesis the Earth rotates right then uh we can uh we we can assume right that the sun will appear
to move uh through the sky throughout the day and will come up the next day at a pretty you know uh predictable time period so what do we do we do we we kind of do an experiment really what we do is we stop and observe right based on our hypothesis right so if this is true about how the Earth moves
right then these you know then this will happen right the sun will come back up the next day right we make that prediction we see what happens with our empirical observations and we go oh yeah the sun went down and then the sun came up again right um and that happens over and over and over
again and that strengthens our hypothesis right and this and this and this goes on right notice there's not an end to this this keeps going right as our observations continue to support our hypothesis our general premise this strengthens into a scientific theory and in
some cases sign like a scientific law right um Theory and law both very strong Concepts right I think uh there's a tendency rhetorically to say well that's just a theory you have a theory um something doesn't become a scientific theory until we've done many many experiments and the
results all support it right and then it becomes a theory that we might teach right and so this and this goes on and on and on right so if that's true then this then you know if that's true then we use deduction right we use a rule to draw a conclusion we make observations
many observations many individual observations and based on those many observations we decide if our rule works or not right and this goes on and on and on forever science is never going to be finished right um there's always more Nuance there's always new data there's always
uh new ideas and uh we can always become more precise we can always become more accurate right um a thing that I want to to say here is for instance somebody who goes and gets their PHD um you write something called the dissertation right uh and what a dissertation is is it's a
really long paper that you've done a lot of hard work on on a lot of research on and what that dissertation really is right to put it in the simplest terms is in order to get your PhD you write this big paper and the result of that paper is you you sort of create a new fact
that's a very simple way of describing it an overly simplified way of describing it but the fact of the matter is is that's what every single person who gets a PhD does is they make a new fact foreign and they have a uh like a sort of a a committee of other instructors who also have
phds who look at this and go yeah all right this is this is this is good right and and to give you uh that should give you an idea of how much there is for us to figure out right if every person who gets a doctor in front of their name has generated a new fact you can
think about how small those facts must be right um and think about how uh the the process of science will go on forever we will always we will always have a psychology right um and so there is no finish line there's just always more to learn and so we have our general theory from that theory we
generate a hypothesis right if this is true then I bet when I do a b c x y z will happen right um if we take something like oh well uh you know a person with with depression has uh is really not so much sad they are that is a disease of motivation somebody might say right so that's our
Theory right actually depression is a disease of motivation right not a disease of sadness okay so then if that's the case right then uh individuals with depression their reports of their mood will be pretty diverse but uh they'll for the most part report issues with school and work related
to like you know completing work that's our hypothesis right so what do we do right if that's my hypothesis right so if it's true that depression is really more about motivation than it is about uh uh you know just feeling bummed right um then uh then these things will be true
so we could we go and we collect data we we grab a uh and we'll talk more about this we grab 500 people with depression and we ask them a bunch of questions or um We Gather a bunch of data about their lives or we observe them or we put them in a a special situation right we got 500 people and we
do an experiment on them each individually right there's a lot of different things we can do we'll talk about many of those different things but we collect data and we collect a lot of it and then we analyze that data that is what statistics is for that's why statistics is incredibly important
and it's a class that you should take very seriously and you should dive into with boldness uh because statistics are everywhere they're used constantly to convince you of things and understanding what gen where those statistics come from will make you more literate about what
is what is accurate versus what is propaganda and um statistics really comes alive when it helps you answer your own questions about the world right and so I I really uh those of you who maybe are a little nervous about numbers I encourage you to uh to dive in feet first to this process of
analysis because it's very important so we analyze that data and we make sure that what we found right what we saw when we collected data we can trust it to make conclusions and then whenever we write up a report where we summarize our data and Report our findings and either we go yeah
it seems to support the theory that I based my hypothesis on or you go well not quite right this actually seems to you know refute my prediction or refute the underlying Theory and then we we all as a scientific Community got to talk about what this means right so that's what things
like conferences are for uh that's what things like going out to lunch with your colleagues is for it's like did you hear what so-and-so found what is it like oh my gosh we gotta you know start from scratch or we got to go back or we got to rethink this right and those are very exciting
and scary times right um both at the same time often okay and that's the and that's the process you should recognize this right because you uh generally speaking if you if you had your if you had your primary education here in the United States particularly in California you very likely
um had to do like a science fair kind of thing where you had a question and then you had a you did some reading in the library you checked out three books right and then you generated a hypothesis right and then you did an experiment or you made
an observation right I watered you know like I watered three different seeds I watered one seed with water one with salt water and one with Coca-Cola which one will grow more right and then you saw you measured which plant grew more right and then you wrote a little report
about it right it was great that's it that's it you did it you've done it already right um are the procedures a little bigger and a little more strict when you do it for real absolutely but most of you again those of you educated in California have done a version of this already
so uh the first kind of research we want to talk about is descriptive research where we describe a phenomenon the goals of science there are four goals of science not everything we do in science does all four but ultimately we're trying to accomplish four things
we want to describe explain predict and control right we want to describe what's going on this is what I'm seeing this is how bright it is this is how loud it is this is how like long or short or you know how long it takes that kind of thing right we want to explain it the reason that this
is happening is but likely because of blank right likely because of ABC likely because of XYZ right we want to predict and because you know because of this explanation because of this description I bet this is going to happen in the future or when you do that
it's going to explode right or it's going to grow right or it's not going to grow right and then control is okay so knowing all of these things being able to make these predictions means that if I build this particular apparatus I build this structure I do this procedure this is going to
happen right if for instance I know the positive reinforcement is a really good way to motivate a person to do what I want then if I want to get a person to do what I want I need to praise them when I see them doing that thing right I need to reward them I need to catch them being good right
it's like a basic child rearing thing catch your kids being good and they'll be good for you right um and so that's about prediction and control right I bet if I spend a lot of energy highlighting the things my kid does that they like a lot they'll do more of those things right and
then from there you get to this you know this idea of control right well you know I can raise a kid who is motivated to do well by doing XYZ right can I totally control the kid's Behavior no absolutely not but we can exert some influence right which still counts as a kind of control at the end of
the day so those are the four goals of science descriptive research is about describing right um and so uh there's a couple ways we do this the first is observation this is where we walk into a room and we watch we go to a place and we watch we uh there are a lot of
ways we can do this really systematically um there's a procedure called coding in which your observers are given like really careful instructions about how to take notes about the presence of certain things uh an example of this is you're going to go to a playground and
you're going to take data on aggression That You observe right on the playground and so then you know if you're thinking carefully about this you have to ask yourself what counts as aggression right and in a good observation The Observers have been provided with very
clear instructions about what observation about what aggression is right you know aggression is you know anytime a a kid comes into contact with another kid um you know with with Force right and that's that's already I've
already done a bad job of uh of what that coding might look like um but the idea here is that ultimately you know uh you get observers who are all um all looking for the same thing right and doing it in a way where two different observers right
um would agree on what they were what they were looking at so over time you get this this information about like oh yeah so kids are way more like we see more aggressive aggression before launch right for during first recess than we do during uh during lunch recess or the final like
that 210 recess I don't know kids still get that but the later in the day recess it's interesting it seems like maybe kids these kids are hungry I don't know right um that's uh that guessing why that is is a different thing the observation part is just when do we see more aggression
oh morning recess uh and then after that it kind of drops off all right that's that's interesting another kind of descriptive research is any kind of survey or interview right and so what we're talking about here is you know uh many of you have have taken surveys already
um if you've taken a class at Golden West before I I'm fairly certain you're given um uh like student satisfaction service they're called they all have every institution has different names for them um but you're given a you know a questionnaire about you know what was you know what ways did
this teacher promote a good learning environment right that kind of thing um and that data gets collected and it gives you a sort of description of how the teacher is doing when you look at all the results of their students right
um interviews same deal you sit across from somebody and you you ask them a bunch of questions sometimes those those questions are set in advance and other times it can flow a little more conversationally ultimately after that happens you take a look at what
was said you probably recorded that interview and you are again looking for particular themes right but always again based on Theory right so you have your theory about what you're looking for and so you start looking for those things
finally case studies so uh what a case study is where we take one particular example one person one phenomenon one thing that's happening right and we ex exhaustively research it we look at it very very very carefully right and get as much information as we can about it right and
what that does right and I'm going to go back I'm going to go back through each of these um uh but what that does is that uh gives us really really really detailed information about something right a really good description about something uh the weakness
of a case study is that it's not really what we we would call it it's not generalizable we can't take what we learned from a case study and apply it to other people right uh a good example of this again kind of leaning on my clinical background here a case
study of an individual schizophrenia doesn't actually tell us a ton about schizophrenia because the presentation of that disorder is radically different from person to person right um that said we could have lots of you very useful information about one person's experience
that might apply to some people maybe uh but at the very least it's going to give us lots to go on when we start treating that person right surveys and interviews are a great way to gather lots of information from lots of people very quickly but we get weird things some weaknesses are we get
weird things like the person's interpretation of the question can vary from participant to participant uh and ultimately we can only really ask very narrow questions a lot of the time right um and so uh it's difficult to sometimes capture information about more complex phenomena
with observation uh what's cool about that is we're not interfering right we the researcher don't touch what's going on and so we get this sort of unblemished naturalistic um representation of what we're trying to study the bad thing is because we're not doing any manipulation
or exerting no control over it um it's very easy to miss the factors uh that influence what we're trying to understand right so uh I can observe kids are more aggressive during the morning recess but I I get no closer to understanding why that might be
right because I haven't done anything I haven't manipulated anything I haven't introduced anything I haven't tried to change anything or I'm not really like looking for any other variables right and because of that any other any of the other things that could be going
on so because of that all I can do is say this is how aggression changes throughout the day and now there's much more work to do before we can really use that information right so ultimately descriptive research does not answer the questions about why things are
the way they are right we don't get to that explanation piece that explanation goal of science right we can do things like oh yeah well then as you can see here you know we can go oh well the Northern Lights appear at this time of year in these places right but we get no closer
to understanding well why does this why does this happen right so one of the ways uh another uh one of the ways we can start to get a little closer to this is through something called correlational research which is all about identifying relationships
um and we measure a correlation from on a scale from negative one to one right and what that does is that tells us how strong the relationship is right so for example a good uh um a really strong positive correlation right
is the relationship between hours spent studying right and your grade in the class right as you spend more time studying we can predict right that your grade in the class will go up now this tells us that there's a strong relationship between those things right
um a good example of a negative correlation and this is one that I had to reckon with when I was uh when I was in school is the more time you spent a negative correlation is the more time you spend watching TV or playing video games right as that goes up your grade in the class is going
to come down right so a positive relationship is when one goes up the other goes up and a negative relationship is when one goes up the other goes down right now what I haven't said is that they cause one another right so we'll uh we'll get to that right so this negative one to
one category right gives us the magnitude a one is a perfect relationship or with with pretty good certainty whenever right for every hour you study your GP go GPA goes up by 0.1 right if I can say that right we might have close to a an R right a correlation coefficient an R value of one
right um and that can get lower and lower and lower it can get really really low right a correlation of zero means there's almost no relationship right so the plus or minus gives us the direction of the relationship here some uh uh kind of to give
you an idea right so an R of 0.25 right we really need a period here right an R of 0.25 right is the very top of kind of weak between 0.25 and 0.5 is moderate 0.5 to 0.75 is strong right and then 0.75 to 1 is very strong right in Psychology we love strong we we will take strong we will break out
the break out the champagne and the streamers for for strong right um a lot of other than that of a lot of other Natural Sciences don't really use correlation because they can focus on experiments um which we'll get to in a little bit but for social sciences we like correlation because it
um well we'll talk about why um but uh in Psych we're very very happy about a strong and we're over the moon super lucky if we see a very strong and then one an R of one is a perfect relationship we chart correlation on Scatter Plots to visualize them right so this is an example of belong to the
lecture the more yawns students have right so here's a we got a long lecture here and we're going to see many yawns right medium electromedium ions if we take a bunch of different lectures over time right so this was a lecture that was in you know this was lecture
in September a lecture in November a lecture in December you know a lecture in the second week of December third week of December right and they're all of varying lengths and they'll have varying numbers of yawns right if we if we plot those on a graph we get the scatter
plot and you can see the line what is called the line of best fit is already drawn in here but even if this line wasn't here you can see these kind of go in a line right and what that tells us that then the factors vary in the same direction right
when one goes up the other goes up so an example of a negative correlation right the longer the lecture the lower student attentiveness so we have a long lecture with low attention a short lecture with high attention a medium lecture with medium attention you're getting
this you're getting the idea here and then when we chart many different lectures and we get this this line and that's a negative correlation right and here's our line of best fit right which tells us oh yeah this is definitely a negative right because this is this Clump sort of linear
thank you and so these factors vary in the opposite direction what you need to know is that correlation does not equal causation right let me I'm going to go back here and that's annoying but here we go
length of time listening to a lecture does not cause students to stop paying attention right being in a long lecture does not cause yawns right there's some there's some stuff that that happens as a lecture gets longer right that's really causing these changes correlation does
not equal causation for instance hot weather is strongly correlated with physical violence right there are more violent crimes when the weather is high does being hot cause people to be violent no right no but there's a relationship between those things right when it's hot there's more people out
more people interacting with each other means more opportunity for those interactions to go badly right we can start to explain those relationships right but we have to but the correlation just tells us about a relationship and so when you see that there's a a strong
relationship between two things you can't assume there's a causal relationship right well you have to assume is that it's possible there's some other variable causing those things to be connected right so for instance why is Parental harshness correlated with child Rebellion right parents
who tend to be harsher tend to have more rebellious children are harsh parents driving their kids to Rebellion are rebellious kids driving their parents to harshness in stressed out families it's just if you're stressed out it's just everybody that's just
everybody doing badly right so parents have to be harsh kids after our kids are rebelling right um our families displayed genetically right our ordinary families just so they just genetically predispose to you know well if you've got parents who happen to naturally be very harsh they're
going to get together and have a kid who's going to be who's also going to be harsh but when you're a child you don't have any power so instead you just are rebellious right um and more right there's a million in perhaps infinite right if you get creative enough explanations for why
these might be related to each other right and so all correlation can really do ultimately is point you in the direction things to experiment on right that's all it can really do in the end and so excuse me I had to adjust my C here um and so
you have to remember right I want to say it again it's written not it's written up here for you correlation does not equal causation so when you see a correlational result which are everywhere it's all the time right such and such is linked
to blank whenever you hear that right there's a relationship between blank and blank right it's correlational thing which doesn't mean you can dismiss it entirely right because correlation is great for identifying so-called risk factors right which is like oh a person with x y and z
we can predict pretty strongly that they're going to have you know problem a right um so an example an example of this uh where we use risk factors a lot is thinking about things like eating disorders again I'm leaning on clinical work here um uh families uh kids who have parents who are
preoccupied with weight and appearance who do a lot of dieting themselves right so kids with parents who diet right that increases the risk of that kid getting an eating disorder right observing a parent who diets period right and you might go well there's good diets and there's bad
diets it's like no it doesn't matter right just a parent who is vocal and open about the fact that they are restricting their food intake increases the risk does it cause anorexia no right but a family who has that a family that's very critical right there's two risk factors
right a family with a history of mood disorders there's three risk factors right um and so uh again those things do not cause anorexia but if I see those things right it's worthwhile too in the back of my head be like okay I need to be on the lookout for maybe these concerns right
right and again we listed these these possibilities here it was one you know one through five it could be any combination right does not settle why a behavior occurs right again because we get this what I've been describing is called the third variable problem we talk
about things like confounding variables it's the same thing it's a variable that um or a factor right that we're not considering right so here in this example we have parental harshness and child Rebellion right those are the variables right the level of how harsh a parent is
and the level of childhood Rebellion those are our variables right because they can vary if that's all a variable is it's a thing that can vary so don't get too uh caught up on that terminology and so there might be a third variable here you know um these things could relate to each other but the
third variable could be stress the third variable could be genetic neuroticism right just like how kind of unpleasant of a person are you genetically and have you passed that to your kids right another example happy mood predicts sociability people who tend to be happy tend to be more social
just being with others lift our spirits to happy people seek out company right uh is it that working on on your own or being more isolated just based on life circumstance make you sadder right um do the same mechanisms that
produce the feeling of good mood also produce the desire of the desire to be social right again could be more rain so again when you see a relationship between two things just remember one correlation design equal causation people say it over and over again it's to the to the
point where it begins to lose its meaning here's what I'll say uh for you all when you walk out of this class at the end of the semester correlation just means you're not done you have more work to do before you can draw strong conclusions right
all right when we want to find out the why we do an experiment we have to do an experiment if you didn't do an experiment you don't know why something is happening right this helps us determine causation
so in an experiment right be randomly assign people to groups we have an experimental group and a control group so in an experimental group we are going to do something to that group right uh we look at the thing we think might be causing a difference right that's
our hypothesized cause and we manipulate it that's our independent variable so for instance so harsh parents rebellious kids I want to know what causes that to happen why is that relationship there that's interesting and my hypothesis is stress
so my in my experiment group I take a bunch of families and they are in my experimental group and if I hypothesize that stress causes parents to be harsh in kids to Rebel right stress becomes my independent variable and stress is what I'm going to manipulate what is that what on Earth does that
mean it means I'm going to stress these families out I'm going to subject them to an experimental condition that is stressful I got to be gentle about this we'll talk about ethics toward the end here but I have to be a gentle right in the way that I stress these families out for instance I
might I might give them a really difficult puzzle to solve with a timer or I might annoy them with uh with loud music for a long period of time um uh I could do something real extra and I'm thinking of all the things that could be annoying here I could um you know I could say hey family come to
room 101 right and then I go oh Family actually I'm so sorry we're going to be in room 102. oh oh 102 is full I think we got to go up to 201 right up some stairs right oh really this is full let's go to 205 right move them around a bunch and and try to get them kind of keyed up of like
the kid get you know kid gets settled down then we gotta get the kid up and moving again right I'm gonna annoy them I'm gonna stress them out right um oh I could do this is the last one of six and now my now I'm getting creative uh
um which is gonna be bad for for time here another thing you might you might do is you might tell the family when they arrive um goodness you're a half hour late because you told them you told them to get here at two o'clock and then you're gonna pretend that the time was actually 1 30 right
so now the family feels bad or they're just like ah or you're miscomicated with and now we've done something wrong that's one group they're getting they're getting the experimental manipulation the control group are treated equally but they don't get any manipulation so this
family is going to come in right to the lab still and we're going to put them in a room with an activity a real like a neutral activity here's a puzzle here's some toys um you're just going to be here for a half hour we don't do anything to them right
and then we look at the difference between those rooms right so we're going to look at the difference between those groups so we've got one group who doesn't get stressed out another group who does get stressed and then from there we're going to
measure parental harshness and Rebellion Maybe it's easier to pick one of those but that's not terribly important right now um so we're gonna so we're gonna measure parental harshness and childhood child Rebellion how rebellious this kid being right how much are they acting out and how
harsh are the parents being in response to that and we're gonna guess our hypothesis is that the stressed out group will have more of those things right and the group the control group who didn't get stressed out we didn't make them stressed out uh those levels are going to be lower right
and so what's important to remember right is the reason we have a control group The reason we have to have a control group is so that we can get a baseline for how families tend to be right for how these families tend to be um so that way we can say look this is a group without stress
when we add stress this shoots up otherwise we just if we don't have a control group we just have a number where it's like this group experienced a level seven stress and without the control group we have nothing to compare that to what does that number mean right so
out of this we get to make some statements about causation right if child rebellion and parental harshness go up when we stress the family out we can say look parental harshness doesn't cause kids to Rebel stress pushes both parts of this family to do this negative behavior so stress is
the cause of this right and the reason they vary together is because stress affects them both right all right so when we do an experiment or we make any uh draw any conclusion from research we've done we need to be concerned with two things uh validity and reliability
so with validity we have a couple different types there's external validity which is does our experiment translate to the real world right so for my example where we bring the family and we stress them out um did I stress them out in a way that they're likely to experience in real
life personally I'm pretty proud of my two ideas of bringing them from room to room over and over again right letting the kid get settled and then stressing everybody out by moving the whole family again doing that a couple times or lying to them basically and being like hey you're late you know
um because that kind of stuff happens all the time especially to families who are disorganized right and so would we say that experiment has external validity which is to say um does this actually tell us anything about how families are in real life outside the experiment
right do the results generalized to the real world right internal validity right um our department are the dependent variable changes the result of the independent variable or manipulation so we have this uh you know we have this mischievous
mischievous cook here who perhaps has snuck in a secret ingredient or other variable into the mix and the idea here is our dependent variable right our dependent variable changes the result of the independent variable manipulation so we did the experiment we saw this change am I allowed to
say that it was my stress condition right that my experimental manipulation actually produced the change in harshness and Rebellion or did something sneak into my experiment and I'm not talking about like a like an animal snuck in and did something what I'm talking about
is is there something about the way I'm doing this experiment that allowed some other factor to influence what's going on right it's hard to ask these two things ultimately does our experiment address the constructs or ideas that I was trying to address
right and we can think about that question externally and internally right uh also the thing to keep in mind is are there bias or logical errors right in the approach we take right um did I randomly assign people to different groups or did I
accidentally with my you know internal bias put really disorganized families in one category and the nice looking family so I just put them in the control group so like oh you all deserve a nice quiet time right um and that would obviously mess up our experiment
so there are there are a couple of biases uh that can uh that can sneak into our um that can sneak into our process here that we need to make sure that we are aware of um and these are uh these are everywhere um everybody has biases um
that uh that influence uh how we see the world right nobody is nobody's unbiased right um no individual can be objective right um there are many people who really Pride themselves on the kind of super reasonable and logical and I'm I see things objectively um and that is a um uh uh
I don't want to give too much life advice here but you're making people around you miserable by pretending that that's true um is the gist of it because um the only way to eliminate that kind of bias is one it's impossible can't do it but you can reduce it and the only way to do
that is through consensus and cooperation with others right and having many many observations in many many instances and right so your own personal perspective is never objective um it is it is uh how you perceive the world as a product of all of your experiences up
to that point which are like as a matter of course radically different from anyone else so um the re the way that we can we try to deal with that is by uh doing experiments on uh many people right um and trying to again like get consensus is the idea so
um one you can talk about these biases everybody can fall victim to these no individuals is you know again without careful uh careful systemic process no individual can really overcome these um uh again without like doing a very careful experimental design for instance so experimental
bias is really straightforward um this is when you the experimenter behave in a way that influences what's going on in the experiment for instance easy one of these my favorite example of these is um you were doing research on a drug for depression right one group is getting Placebo
pills fake pills sugar pills they take the pill nothing's gonna happen they might they get like a small dose of sugar very very small to the point where no one would notice it other group gets the the actual you know antidepressant right if you are aware of which pill is which you are going to
be much just whether you like it or not you're non-verbals right your demeanor right is going to be different when you give a person the placebo or the fake versus when you give them the one that you expect as the experimenter you want this to work because you're probably trying to prove my
idea for a drug works so you're going to give the you're going to as you're handing over the uh the antidepressant that works you're going to probably be a little more enthusiastic you're going to communicate non-verbally very subtly uh hope optimism right whereas the person who's
getting the sugar pill you're going to hand it to them and you're going to be like well I'm wasting your time right which is in turn going to influence the mood or the expectations about that medication of the experimenter right so the experimenter can again influence things in
subtle ways um really straightforward ones are things like you you the experimenter can be tricked into like looking with your eyes at like a correct answer or the expected response right and so that can influence what's going on
um so demand characteristics this one oh this one's rough this is it's just so demand characteristics are basically like do the conditions of the experiment give away what's being measured right the conditions of the experiment give away what's being measured
right where then participants are being like Oh you're trying to figure out this about me right so um you know uh again um thinking about the stressing families out thing um uh a family that's stepping in knows they're in an experiment you know the you know the the experiment comes out and
says oh you're a half hour late and then the mom goes like I think you're trying to stress us out right because I know I've carefully read the instructions I got here at two when I was supposed to I think this is oh this is pretty experiment right and then all of a sudden boom
it's blown up right and so uh yeah the structure of the experiment can give things away research participant bias is a little bit different it's very similar but a little bit different by virtue of knowing you're in an experiment that your behavior is being observed you change your
behavior right there's a concept for instance called social desirability and when especially when we're doing stuff with surveys isn't just experiments but when we're doing stuff with surveys we have to watch out for this um if you're researching stuff that is socially undesirable
right or researching stuff about like activities that generally are looked down on by Society or people experience experience discrimination because of these activities um you'll have a harder time getting accurate data because many people won't be open or honest right about those
things right because there's shame or stigma around them right so for instance an example of this is trying to do research on the um the prevalence of various like sexual fetishes right um sexual fetishes that Society kind of goes out of its way to kind of tease or make fun of right
um fewer people are going to report having those right um and they're gonna hide they're gonna hide that right and so beyond this though um uh when we bring somebody into a laboratory right or you know a room that's relatively featureless you know fluorescent lights plain Furniture
um people get nervous people feel uncomfortable and people are going to act differently than they would in real life which is going to hurt our external validity and then the placebo effect this is a fun one right when you ex when you expect medicine to work
you start to feel it working before it's actually doing anything chemically right so you take Advil you feel a little bit better um but uh you know it hasn't started working yet right um your your body is is pretty powerful most of you are feeling all of your feelings and thoughts and experiences are
the product of right your body and so it stands to reason that when you have an expectation right A Feeling a different way there would be some mechanisms that would start to kick in throughout the rest of your body right so you you take a sugar pill be after
being told that's a it's a you know it's a pain pill you're going to feel a little bit better all right what's fun about the placebo effect is um the the manner in which the placebo is delivered it changes the effect so for instance if I give you an injection of like a saline solution
right so I give you I take a needle and I stick it in you um you'll feel the placebo effect stronger you'll feel better than if I just give you a pill right because uh because your body anticipates really strong medicine comes from Needles right the color of the pill can affect this as well
um and then my favorite example of this is and I I see fewer and fewer students drinking these all the time which is warms my heart because they're gross and bad for you but um uh energy drinks for instance uh do not have to taste the way they do energy drinks could taste really good
um they could taste like any other soda for instance um however people feel like they work more when they are made to taste more like medicine right and so that weird kind of bitter medicine taste that energy drinks have
is there to engage your placebo effect and trick you into being like Oh well this is gonna work right and then when you have that expectation it feels like it works right um so that's a fun example of how placebo effect is just used in product design
so the way that we address this right remember I talked about like well you can't you like an individual can't really be objective um the way that we address this is we and I talked about how do we get past that right how do we do science if an individual can't be objective well the way to
get closer to this um is a double-blind experiment in which the groups are randomly assigned right and neither the participant nor the experimentor knows which group is which so let's go back to the um the antidepressant the depression pill example um the way that works is um the participants don't
know what group they're in right and the person who is handing out the medication each week or whatever isn't given uh is given an unlabeled bottle right or rather on the bottle it just says this is for participant ABC and then the participant shows up and says hi I'm ADC and
you just hand them that bottle of pills and you have no idea if that's the experimental group or the control group but the person who's going to analyze the data later does know right oh ABC got the placebo cool let me make sure they're in that category right and that's how that works
um it works really well and it IT addresses a lot of these things right um now uh bias uh sneaks into research in all kinds of ways for instance um you know who decides what research gets done for instance what questions get asked um what sorts of uh um
we'll get to this later I suppose but um uh what sorts of sampling procedures are used all of those things kind of get dictated by people within the power structure of um uh of a university and so um bias can sneak in that way too it's not just how
the experiment is conducted but which questions are being asked right uh it also matters right because it's a question of like you know will this get us more funding um are people generally interested in this are we going to get featured you know on CNN or you know
um or you know are we gonna get put on the news right we're gonna get covered in the newspaper if we do this experiment or it's just nobody gonna care right other than us in this room we're really excited about this topic um so buy it that you know a double-blind experiment is not
sufficient to remove all bias because the Again by virtue of being a system made of people uh the like the scientific Community is also riddled with bias right um and so it's much like I talked about the scientific method is ongoing there
is no end to science there's also no end to addressing the impact of the human factor right so that's not a Finish Line we can reach I think a lot of people um feel exasperated when it's like okay we fixed this issue but now there's this other issue and the idea is it's not good it's
not going to end because there's always going to be um we're always going to have biases and stuff that need to be addressed right and so tackling bias needs to be part of this ongoing process all right
so um we can take different approaches to uh um uh we can take different approaches um to understanding a phenomena right there's not one best approach right um because you might say well the experiment gives us well it gives us the answers as to
why so we should just always do an experiment but figuring out what the heck we should do an experiment on takes a lot of research first right so let's start with this the apparent increase in belief in conspiracy theories right like what
there seems to be a rise right with the rise of like juvenon um I feel like uh I feel like this peaked and it's finally going down but like I think more like more people than ever more people than like since the 1650s right believe the Earth is flat right just a absurd stuff right um and so
the rise of Rise of conspiracy theory um we might have questions about where that comes from right um and so uh a thing to to ask ourselves is what are some different ways we could come to understand this right so one is observation a way we could do this is an experimenter could
uh join a web form or web Community right like a Facebook group or a subreddit or a Discord Channel probably most likely these days right of people who support this conspiracy let's just use Q Anon right because there's you know there's a fun documentary on HBO you can watch about that
um uh and so I'm not going to get into what that conspiracy theory covers but um you join one of those Discord servers and just observe right make notes about what conversations you're seeing who's most active who's in charge who posts the most right where is you know is
there one individual who sort of like you know feeds the ideas into this group or is it sort of a much more Community more diffuse kind of thing right also with observation this counts as well um you know if you got the cooperation of the members or leadership of this community you might be able
to just be like hey can I just like get a a just dump of your whole website for me to like that I can search through and you know code like hey this word appears a bunch or like oh look this this you know this community also like is really like super racist or this this community also like is
really into you know it was weirdly in the video games or right like there's you know what are the other you know threads moving through this right and we that would be up that would be observation that would help us understand the shape of this conspiracy theory right again doesn't tell us why
we're getting more conspiracy theories but it could give us a hint that we could then research surveys interviews was as we reach out we contact people who are part of these groups people who espouse these theories we ask them questions about things like well how did you get into
this right like would be an interview question like how did you learn about this um surveys of things like how much do you believe this how much do you believe this other conspiracy theory right so we could start to get at things like hey you know a person who believes one conspiracy
theory very quickly will endorse a bunch of other ones like we went from Q Anon to Bigfoot right um and so we could be asking questions right of these groups studies we could take one person right and this is this is like a 48 you know four hours uh like
60 Minutes sort of um uh part of their approach usually when they're talking about something is they will hone in on one or two people to be like this is the face of this right and so um with the conspiracy theory thing I think the thing we see a lot is it'll usually be a younger
person somebody in their 20s or 30s right like an adult like a young not young adult but like a not yet middle-aged adult let's say that being like yeah I lost my parents to this right they fell down the hole and now they're you know deep into this you know this stuff um and that would be
closer to a case study where we're really going to learn everything about this person's grandparents like what's their background how are they raised like okay Grandpa was in the military right so he um and he fell in with like this particular group and uh you know been over time his politics
shifted right what caused that shift uh you know oh you know the uh the impeachment of Nixon or something right um you know did XYZ and so what we're doing is we're we're working through the steps sort of historically of this person's life that led them to be somebody who was vulnerable
to falling for conspiracy theories or um maybe we don't talk about it as vulnerability but more like uh cultivate an interest in these kinds of ideas right and we try to use one person's journey and really understand it super keenly super well so that we can um start to wonder right about
um how it you know we can then start to ask questions is how similar is this journey to everybody else's right and also too a useful thing about case studies where I see this a lot is um uh the idea of like I was able to save my parents from this I got my parents to like not
do it right and uh you know then we look at a case study of that process of like oh so how do we like heal a person who has then Fallen like out of touch with reality right because again the thing about conspiracy theories is um they bring you out of touch with reality and then also
because of that they make you unable to actually solve the problems you're so concerned about right if you're so concerned about a single like cabal running the world you are then actually powerless to solve the problems because um there is no such cabal right um that's fiction and so if you're
worried about this fake thing you can't actually solve problems right um which is really useful really useful to kind of maintaining the current like power structure isn't it um and so uh uh it's very helpful to understand these these things correlational research right so we look
at relationships between like one variable might be depth of a person's Devotion to like I am 100 sure this is true right and I want to devote you know 100 of my life to it right and then looking at what variables predict that right the level of endorsement right so we might do things like
you know how isolated does this person feel or did this person feel before they found this group right um what else for conspiracy theories um I could be mean and say we could look at intelligence right um that's kind of a dubious one you should be careful about but um we might
be able we might find relations between yeah it turns out you know the lower your right the lower your IQ or there's a particular range of IQ that really strongly predicts endorsing these right or um it could just be like political affiliation right like super strongly left leaning versus
super strongly right leaning um and like which end of the spectrum predicts you know uh um uh you know endorsement of these things right so we could look for or make guesses about what factors predict an interest in conspiracy theory right
finally an experiment and so this is this is where this is this is where this example I've chosen gets a little rough because I believe personally it would be unethical to do anything that might encourage a person to support a conspiracy theory because again it brings
you out of touch with reality and then stops you from actually being able to make your life better right because again you're focused on this fictional thing um right you're so like concerned about like we have to show the world the World is Flat right we have to show the world but the
Illuminati wrote everything right um you know we have to we have to show the world that um you know this pizza shop is involved in child trafficking right which brings us that's closer like kind of a q Anon thing right and we get these people get so preoccupied with these with with these
with these things it's like it's bad for them it isolates them from you know people who you know from their families from their their friends who you know uh want what's best for them and they get stuck in these online communities that are again wrapped up around this fiction so um I think it
would be unethical to do an experiment that was like Hey if we take this group of people and we you know manipulate their Facebook or Instagram feed to only show them conspiracy theory stuff right I bet you know exposure to these ideas will cause more people to subscribe to them right
hopefully that makes sense right this idea of you know okay well let's manipulate their feet which by the way fun fact uh Facebook um and Instagram I know it's the same company but um they uh were caught recently doing experiments like some people right where they looked and they were like hey I
bet if we increase the amount of negative stuff we show this person their posts will become more negative and they did these kinds of experiments and manipulations and they got caught and had to stop it's a mean thing to do to somebody without their consent right um we'll talk about that when
we get to ethics so I think it would be unethical to incur to do anything that would make a person more likely to join a conspiracy theory so then for me the experiment that I would then design is I would take a group of people who are part of a uh a conspiracy theorist community and examine
which interventions can reduce their belief in it and their support of it right and then we start to get an idea of maybe what causes these ties it's um and that is um uh you know I'm kind of like one degree removed from what we're actually interested in which is why are more people
subscribing to these and so then do an experiment on what gets people to stop believing in them is not quite the same thing but it's important to consider again the ethics of how you're doing research right and so the experiment again here is I want to uh try a particular intervention of like
oh well we're gonna we're gonna um something along the lines of like we're gonna sit these people down with an actual person they believe is part of the Illuminati or we're going to sit these people down like we're gonna take these flat earthers and we're gonna sit them down with
um uh you know a pilot who has seen the curvature of the Earth or an astronaut or that kind of thing right maybe it direct human contact with the so-called enemy right we'll fix this right so that then we do an experiment where we take a control group and we have the meat with just like a normal
person and then we have an experimental group where we have them meet with like an astronaut and we see who still thinks about who thinks the Earth is flat after those meetings right and so you can see that yeah I mean this we've been on the slide for a while I realize
um you know there's there's tons of different ways and and I hope my hope is that you heard each of these ideas and were kind of like oh well each of those it'd be really interesting to see the results of each of those right even though the experiment might get us closer to why we get a
lot of really valuable information from each of these right um that could Inspire other research so when we do research we need that we need a sample we cannot do an experiment on the whole world right um and so we have to think about okay so we're doing an experiment and after the
experiment we're going to draw some conclusions who are we drawing conclusions about is a question that's important to ask um and the answer to that question is your population right so if I want to know okay well if I'm interested in college success right my population is
probably college students right what factors ensure that students graduate on time right so my population is college students if I'm going to do an experiment or I'm going to learn I send out surveys I'm going to do anything I have to choose who I'm going
to look at so I'm going to take my population of college students right and then I'm going to draw a sample which is the portion of the population we actually look at right ideally we want a representative sample so we take a group of people out of this so we take let's say
all of Golden West right and the population that I take should be representative of the population of Golden West so for instance this is an online class that you're in right now so um if I am going if I'm interested in college success and like let's just say success transferring right
um I need to be looking I need to make sure that my sample isn't just taken from in-person people right I can't just walk around the quad or whatever and ask people questions because then I'm leaving out everybody who's taking classes online which is a large part
of gwc right so if let's just say 50 of the population of 50 of Golden West students are in person and the other 50 are online my sample should also be 50 50 in person and online so in this example here right uh where we see these people if I've got you know again 50 50 right
and I'd get a sample of four which is really bad ha it's a you need a big sample but half of them should be in person half of them should be online this does also true demographically this is also true based on gender sexual orientation country of origin all of these things we should try and think
of as many demographic factors as possible and try to build a sample that's representative right um and so representative sample is the opposite of a bias sample so for instance if I sit out on the quad and give out surveys about transfer success right I'm creating a bias sample because I'm only
talking to people who won are on who are walking on the quad which is like not in the I don't think let's just say for sake of argument it's not in the center of Campus so it's biased toward people taking particular classes who parked at particular parking lots right who happen to be there at a
specific time of day I don't think Golden West offers a lot of night night timing classes but um uh but you know biases toward a particular time of day right and all of these things might be sneaky factors here right um for instance I would bet students who are you know taking that
8 30 or whatever like super first thing in the morning class those students who are like get up and go I bet they're trying I bet I bet they're I bet they're more successful right I could be totally wrong about that um but a student who's willing to you know keep a uh keep a you know a
decent sleep schedule who's willing to wake up early for their education probably has a bunch of other factors that are pushing them towards success right do this is not to say that students who aren't mourning people right um uh I think at least because I'm including myself here at least
half of your psych Department we're not morning people right and so I'm I you know don't feel if you're like an afternoon person don't worry about it um the great philosopher Descartes not a morning person and um uh the queen of Sweden made him wake up super early and he died because of it
so don't feel bad if you're not a morning person um and don't look up that story because then but it's more complicated than that um it was also really cold in the morning in Sweden so I might have not might have done it poor guy um and so uh but I would I would guess generally speaking that
um you know if you're interviewing students mostly first thing in the morning you're going to get a bunch of students who are you know a little maybe a little bit more proactive I don't know um and that could be my own personal bias I'll really admit that right you know sorry
for discriminating against uh afternoon people and night owls um uh I respect all of you so with that type with that terrible digression out of the way I apologize um uh hopefully you start to understand that once we fail to get a representative sample
all kinds of other things could start sneaking into what we end up measuring right so usually we try to get a representative sample by sampling randomly and you might go well wait how does that how does that work well if you have a big enough sample drawn randomly you will eventually
get a representative sample and the best way to think about this is to really shrink the numbers most of you can imagine flipping a coin four times and getting heads three times right so you flip a coin four times and it comes up heads three times
you and I would predict right if we flip a coin four times we're gonna get heads twice and Tails twice right we're gonna go 50 50 because the coin has two sides but we all can imagine flipping a coin four times and getting those three heads in a row right
however and most most students y'all can you'll get this kind of intuitively if I flip that coin a million times I'm gonna be very very very close to 500 a million times yeah 500 000 heads and 500
000 Tails because over time right that 50 50 probability is going to even out right and the random sampling is the same thing if we're truly drawing randomly from the population like you know putting everybody's student ID in a big hat and drawing it randomly right
if we do that enough right if instead of taking four people we instead take a thousand students right we are much more likely to get a sample that represents the population right and then we don't have to worry about you know somebody going through
and being like okay well we need another person with like a Vietnamese background in here where we need another person with an African-American background we need another um we need another you know uh uh we need another like we need another gay person we need three more
straight people we need to you know we have to like build we have to build the sample and when you start really putting in that human factor you let in lots of room to kind of make mistakes right and so if you can get enough people randomly you are more likely to have a sample that is uh
um accurate right and so that's and so Random large random samples are what you want that's a big plus so uh there are a couple different places where we can then do research one is in the laboratory which is sort of an artificial world uh the advantage of that is
it's controlled right we can uh for instance we can have do cool things like because we're in a building with very few features in a room with very few features we don't have to worry if it's sunny or rainy right affecting somebody's mood right you just come in and you're in a room
no windows and it's just boring fluorescent lighting and that's what everybody gets right in a natural setting however we get to do more naturalistic observation and so the advantage of a natural setting is we're doing more naturalistic observation which means we get
to see people kind of out in the wild right we are going to be able to generalize our findings more right um possibly generalizer findings more there's actually arguments either way um but with this naturalistic setting uh we're again we're seeing people in the wild
we're seeing people as they are right and so all of the little factors that might change a person's Behavior are still present which could either which make it so that it's hard for us to make strong conclusions because it could be some other sneaky factor in the environment
but we can be a little bit more sure that when this person goes somewhere else in the real world right that will get similar results whereas with a laboratory setting we have much more control so we really isolate that variable we can really say look everybody went through
the exact same thing except for this one thing I manipulated so I'm sure the changes we saw are from this exact thing but then we have to wonder can we take those results outside the lab right an example of this is wow this person really did this did this I did this wrote this
report really well in school but then they had to go and write a report for a job and all of a sudden without a rubric all of a sudden without um without knowing the sort of Industry standards right that report actually doesn't hold up well right so School trained you to write you know
five paragraph essay right but when you got to the Working World it turns out if you write more than a paragraph your boss is annoyed with you for writing so much and so really the skill you needed outside of the quote the laboratory of the classroom right was to be able to write
really concisely and effectively right instead of expanding your idea right into five paragraphs that's kind of an example of how that controlled environment kind of backfire when you step outside of it all right so that's how we gather our data right and so then we have to analyze and interpret
our data we do that with Statistics right which is the math mathematical methods used to report data so we have a couple different types we have descriptive and inferential descriptive statistics help us describe or summarize the mean median and mode those are descriptive statistics right so the
average number you've got is descriptive right so let's see how let's start here so the measures of central tendency are descriptive that's mean median and mode right so let's go back here so mean is like we have our average so we have three instances three zeros 110 113 and two twenties
when we total this and do the division we get an average of of nine right we take the median we line them all up and it's the one that sits in the middle at the midpoint so that gives us 10 right and then the mode is the thing that occurs the most which is zero right and you can
kind of see like oh these different Central Tendencies actually tell us different things about the data right the mean and the median give us pretty similar things in nine and ten um but the mode gives us this zero right so for instance um a useful thing to think about is um
housing prices it's you more useful to use the median for housing prices um because averages are pushed up by incredibly expensive and Incredibly cheap houses there's a lot more incredibly expensive houses right and so that messes with the average a little
bit but the median will put you right in where the actual middle is generally like so there's also measures of dispersion which is how variable the data is how different is one piece of data from another so we get the range right which is 0 to 20 here for this group right
and then the standard deviation which which is I'll read this to here the square root of the of the uh of the average squared deviation from the mean namely 9.147 so I'm not necessarily going to walk you through that equation necessarily but what the standard
deviation tells us right is you know so we have our average right and so then how uh in general how much is one how far is one piece of data away from any other piece right um and so that will uh that will that will help us get a sense of uh how
far from the average any particular data point is the easiest way to think about this is thinking about IQ the average IQ is 100 and the standard deviation is 15. right so what that means is if you have an 85 to a 115 IQ you are we can we can we can make really accurate predictions um about
63 of the population Falls between 85 and 115 IQ and we know this because standard deviation um uh that's the standard deviation works right and as we get more standard deviations away from the from the mean from the middle right from the the average as we get more standard deviations away
we know we're talking about fewer and fewer people so if you have a for instance if you have a 145 IQ you are a vanishingly rare intellect right uh and then if somebody tries to tell me that they have 160 IQ we're talking about fractions of a percentage right of these existing
right and I know all of this information because I know that uh because I know the average and I know the standard deviation of the data I can go wow that is an extraordinarily unusual number right that's what the measures of dispersion are useful for we can go oh
that's right at the end like with range we know that's all the way at the end or all the way up kind of the beginning it's the lowest possible the most standard deviation we can go wow that's a pretty like that number's above average but not by much all right so I have 110 IQ
cool sick that's you that's you got a good you got a good old brain there well done somebody tells me that 160 I'm gonna first be like well you're probably lying or you took a test on the internet and then then you know if if they like well no here's my you know here's my score on the
uh the weschler adult intelligence scale that was done by a psychologist I'm gonna go oh my goodness right that's incredibly rare so that's why it's useful to have these measures of dispersion because they're going to tell us uh we can then derive a lot of information from a single point of
data right we can go oh that's very very far above or not that not that far above the average right inferential helps us draw conclusions they are the bridge between the sample and the population right so um uh they help us ask or help us answer the question does the data confirm the hypothesis so
let's go back to the coin example that I gave right where you flip a coin four times and you've got three heads and one tails if you flip a coin four times and get three heads and one tails the question I have for you and I want you to take a second to answer this for yourself
is it reasonable to assume if you flip a coin four times and get three heads and one tails is it reasonable to assume you have a trick coin right there's something special about the coin that the coin's been manipulated in some way right the answer well you're still thinking
is no it's not a reasonable assumption right flip four coins I could grab a quarter right now and I'd probably get three tails and one heads right or three heads and one tail right um it's not it's not reasonable it's you have so few instances of flipping the coin right that it's
it you might get it might come out a little weird and you might get an extra right you might get an extra you know an extra heads when you know the probability suggests that over time you'll have 50 50 of each right now if you flip the coin if you if you flipped the coin a million times
or let's say a hundred times so the numbers are always there if you flip a coin a hundred times and you get 75 heads and 25 Tails right the same ratio as before but now you flipped the coin a lot it is reasonable to assume you've got a trick coin right and inferential statistics help you
look at more like so a coin is not complicated it's a 50 52-sided thing inferential statistics let you look at more complicated things and then say was it your experiment that made the difference appear right was it the when you when you stressed that family out and they became more
um rebellious and harsher right was that probably because of your experiment or was it just a chance right where sometimes the coins comes the coin comes up with three heads sometimes the family is just harsher right it has nothing to do with what you did right and so inferential statistics help
us answer this question and they help us identify something called statistical significance right um yeah so basically uh you are allowed to say your experiment observed something right you owe it what I did caused this difference
right you're allowed to say that if your statistics tell you that there's a five percent or less chance that it was all random chance rather that you can say with 95 certainty it was your experiment that caused the difference to appear
right and so we call this the alpha level or the confidence level right um and so you and your statistics and research methods classes you will you will go over how to obtain how to get these numbers right and uh we'll go into a lot more detail about how to interpret
your results right because it gets a little more complicated than this but ultimately the thing that you know the the thing that people are obsessed with is this Alpha level right um 0.05 right it's it's almost like a jokey value among statisticians 0.05 that's it you know and so
um what this does what most inferential statistics will tell you right is at amongst many other things right inferential statistics the different kinds of analyzes like t-tests Anova and kova mancova you know um uh you know uh regression right that kind of thing all of those will
tell you lots and lots of different things but the thing that you generally tend to look for first is you know did the differences I find are they likely right is there a 95 chance that they're up they're because of my manipulation right or is there or is it you know right is there a greater
than five percent chance that it was just my sample was weird right or some other Factor foreign so your participants have rights [Music] um hopefully that doesn't surprise you or hopefully that doesn't make you angry rights what are you talking about um uh your participants have
uh have rights um that should be respected um the medical and psychological Community the medical community in particular just because there's worse stuff that they can do but the medical community has a long history of um not really respecting the rights of various kinds of patients
um I just I think time wise we're running along here and so I I'm not going to necessarily get into too many specifics because they're really Dreadful stories but um uh the Tuskegee syphilis experiment in particular is one example um uh that is worth is worth investigating
um and demonstrates the need for strong ethical oversight for all experiments um and so I hope um here in 2022 you can kind of accept on face value that ethical oversight for all experiments involving humans um and maybe even animals we can go there too um is uh is worthwhile
um to ensure that everybody is protected right so uh the American Psychological Association has some guidelines for how to do experiments that generally need to be followed any University that conducts experiments on people has
um uh an internal review board an IRB that's meant to make sure that all all experimenters are educated and are following those guidelines so that one component is informed consent you have to to some extent tell people ahead of time here's what's going to happen right now there's room to
in some cases you're not going to tell people like here's what we're trying to learn right but instead you're going to say like oh well you're going to do um you know you're going to be asked to do an activity together as a family and here's what that's going to entail
and here's what kind of feelings you might have right that kind of thing confidentiality you have to protect people's identities right um this is I it was wild I saw somebody that somebody was obviously not a scientist or not part of Academia was very much like I can't
trust this data because the participants haven't been identified it's like no you can't do that um because if people think that their names are going to get out they're going to behave differently and act differently and so we have to protect people's identity one because it's
worth researching sensitive topics or topics that maybe that there's societal stigma around and two um when people know that their names are going to be published next to their data right um they get nervous right imagine being like oh we're going to give an IQ test to a thousand
people and we're going to print everybody's IQ score right and their name right people are going to not want to do that right only people who are confident in their intelligence are going to sign up or be willing to do that and so we protect people's identities to go hey come as
you are right and no one will find out it was you because all your data your your data is just going to be one de-identified line amongst thousands of participants right so that protects people and it encourages people to you know behave realistically there's debriefing afterward where
after they're done with an experiment we tell participants hey here's here's actually what we were looking for right here's what we're looking at you know any questions um here's an email you can reach out to if you like feel sad about what happened right and that doesn't that you know
nope basically nobody calls the number but um you know you give it to people anyway right so if you feel upset because they're you know I have some big feelings about what what happened you know um call us here's what we were looking at here's why this part of the experiment happened
um here here let us let's reveal to you here's an element of deception right um we you know for instance with the the stressed out families like we totally knew we told you to come here at two we told you 1 30 to you know increase the stress level of the
situation right we might go over those pieces right debriefing I think uh um comes a lot from the fact that there are some experiments you can do in Psychology where people will learn things about themselves that they maybe weren't quite prepared to learn about themselves
um well and we'll we will probably talk about some of those experiments his some those historical experiments over the course of the class deception you have to be very you have to make a good case for why you're deceiving people if you're going to lie to people you have to there has to be a
good reason for it at the end of the day good reasons include things like look if we were open and upfront about exactly what we wanted this would totally disrupt the process and it's pretty easy to make that case but um you need to at least have good reason to deceive folks
and so yeah institutional review boards um uh are super important I'm speaking of somebody my thesis uh got super tangled up in my in my IRB at my institution at Cal State Fullerton um because I was working with kids my my thesis was on an efficacy study about a therapeutic
summer camp and because I was working with kids um there was a lot of really intense scrutiny paid in my project and it really delayed me and kind of messed me up a little bit um and uh I was very sad at the time but even then and especially now
um that was hugely necessary super important that we're making sure that these again these kids who are suffering from very psychological disorders and social emotional behavioral issues are being protected and not being subjected to something that's going to like hinder their um their healing
right like he didn't hinder their progress um and so again I'm speaking from somebody who was massively inconvenienced by their IRB to the point where it almost delayed my graduation um and I I'm I'm here telling you super necessary super super important cooperate with your IRB and recognize
that they are they're doing something to ensure that our discipline broadly remains respected and remains a force for like good and progress right and actually improving people's lives so um yeah uh so I I kind of wanted to mention that again you know coming from a perspective
where you might expect me to be a little bitter about it but now um super important um and I'm glad that those kids um I I was certainly uh advocating for and uh wanting very badly to take care of the the kids I worked with on this experiment but also I'm glad that they had a
um uh even more even more Advocates um yeah so that is those are sort of the basics of how we approach research and psychology right this is how we the what we've gone over is is how we um is how we we build knowledge in Psychology every fact you're going to read in your textbook
for the rest of the semester is the product of an experiment or the product of really intense correlational investigation right using using many like you know lots and lots of um carefully designed correlational studies and that have been repeated ideally not necessarily
all of them have been successfully repeated but the idea is to replicate as many of these studies as possible so that we can confirm that they're accurate and so everything we're going to talk about is the product of this process that we just went over right and so
um whenever we talk about an experiment in the findings of those experiments you can assume right that the methods I've covered were used right and so when you hear a fact or something that we talk about that you feel skeptical about it's okay or it's like that's really matched my experience what
I want to encourage you to remember is that um uh your individual experience is unique and distinct right and the sort of uh the concepts that have been identified right that we're describing uh are usually the sum of the the experiences of thousands of people
right hundreds of people dozens of people right and I hope that you'll then take your own personal experience and understand that maybe your experience is the exception right or an exception or that perhaps the way that you're thinking about your experience maybe is is
subject to your your kind of your personal bias or expectations right whereas these things that we'll be talking about this semester are the product of processes designed to iron those things out right and so that basically what I'm saying is it is safe in this class to approach some of
the stuff we're talking about with an open mind um because uh nothing I tell you is made up right so um I gotta go to Great pains to make sure that nothing I tell you is just made up and if I do tell you something that's just made up or is my best guess I will tell you ahead of
time I will I will say something like I'm taking my professor hat off for a second to say that sort of make this yesterday answer this question so you know carefully read this chapter there's much more in the chapter uh well not much more but there's there's there's
more in the chapter so make sure you take a look at that that's always going to be true um and um yeah I think carefully about these these processes and think carefully about the questions you have and let's see if we find you some answers over the course of the semester um yeah so I will
uh you will you'll hear more from me uh next week and that's going to be it for now take care folks
Heads up!
This summary and transcript were automatically generated using AI with the Free YouTube Transcript Summary Tool by LunaNotes.
Generate a summary for freeRelated Summaries

Comprehensive Guide to Research Approaches in Psychology
Explore the scientific approach to psychological research, including theory development, experimental and correlational methods, and key concepts like variables, operational definitions, and random assignment. Learn how psychologists design studies to describe, explain, predict, and control behavior.

Why Research is Crucial in Psychology: Understanding Scientific Inquiry
This lecture explores the vital role of research in psychology, emphasizing empirical evidence, scientific methods, and critical thinking. It highlights how research validates psychological theories, debunks myths, and shapes our understanding of human behavior.

Understanding Psychology: Key Concepts and Common Misconceptions Explained
This lecture explores the scientific study of psychology, focusing on behavior and mental processes. It debunks common myths, highlights the goals of psychology, and explains the importance of research over common sense assumptions.

Ethics in Research: Deception, Animal Studies, and Institutional Oversight
This lecture explores key ethical considerations in psychological research, focusing on the use of deception, animal research, and the role of oversight committees like IRBs and IACUCs. It highlights the importance of informed consent, participant dignity, and minimizing harm while discussing historical examples and current standards.

Understanding Correlation, Sampling, and Experimental Bias in Research
This lecture explores key concepts in research analysis including the difference between correlation and causation, the importance of representative sampling, and how experimental biases like placebo effects and experimenter bias can impact study results. Learn practical examples and strategies such as double-blind procedures to ensure research validity.
Most Viewed Summaries

Kolonyalismo at Imperyalismo: Ang Kasaysayan ng Pagsakop sa Pilipinas
Tuklasin ang kasaysayan ng kolonyalismo at imperyalismo sa Pilipinas sa pamamagitan ni Ferdinand Magellan.

A Comprehensive Guide to Using Stable Diffusion Forge UI
Explore the Stable Diffusion Forge UI, customizable settings, models, and more to enhance your image generation experience.

Mastering Inpainting with Stable Diffusion: Fix Mistakes and Enhance Your Images
Learn to fix mistakes and enhance images with Stable Diffusion's inpainting features effectively.

Pamamaraan at Patakarang Kolonyal ng mga Espanyol sa Pilipinas
Tuklasin ang mga pamamaraan at patakaran ng mga Espanyol sa Pilipinas, at ang epekto nito sa mga Pilipino.

Pag-unawa sa Denotasyon at Konotasyon sa Filipino 4
Alamin ang kahulugan ng denotasyon at konotasyon sa Filipino 4 kasama ang mga halimbawa at pagsasanay.