Psychology week three assignment

there we go all righty all right you all so welcome to week three as always once i start screen sharing i can't see the chat or hear or i'm sorry see any comments or anything like that any hands raised so please feel free to just unmute yourself if you have questions comments you want me to pause for a minute i i want you to to stop me and ask your question please Okay, so let's start off a little silly today. How are we feeling? Beginning of week three, we came into this course hot and heavy with a lot of big data work. So some of my favorite staff memes, me in the top left corner, my staff teacher explaining to me P less than 0.05 and P, me still confused with which one is significant and which is not significant. Not sure if I should use quantitative research or qualitative research. I love the Patrick Starr one, critically evaluating someone else's research article versus writing a paper explaining your own research. It feels like that. I'm sure so many of you are like, why am I studying numbers so I can understand feelings? And all of us just smile to hide completely how overwhelmed we are half the time. So a little silliness to start. You're doing great. Message going into this week is don't miss the forest for the trees. Those of you that were at residency heard me give a version of this before, but you all are headed towards the end. So the whole last year of your program is going to be like, wake up and finish your PhD. so don't get lost okay i want to take a few minutes to remind you what you already know so let's do that you've at least had some previous research methods courses so you've learned essentially about the scientific method the design of research fundamentals different sampling methods different types of variables like independent and dependent variables basic measurement scales like nominal ordinal interval ratio, how we're measuring the things we're interested in. So this gives you a good foundation. And then we need to build from that onto hypothesis testing and choosing the appropriate statistical test. Okay. So many folks stay confused about null hypothesis testing. And this is the time where I hope if you haven't gotten it, that you're really going to get it now. So the way that things work in the scientific method and using null hypothesis significance testing is you first formulate a research question in which you operationalize both your independent and dependent variables saying how they will be measured or manipulated. You state both your null and your alternative hypothesis. So the null hypothesis always says there is no effect. There's nothing to see here. and the alternative hypothesis says there is an effect. Now, we are statistically testing the null hypothesis that says there's no effect. So, we're looking for the probability of no effect, okay? So, step three, based on your research question, your variables, and your hypotheses, then you choose the appropriate statistical test. So as I said, it depends on your question and the types of variables. You set your significance level or your alpha level. So the conventional in behavioral science is an alpha level of 0.05, meaning that we're looking at a 0.05 or 5% probability as our cutoff range. Then we would go out and collect the data and input the data into whatever software we choose. So SPSS, R, Jmovi, Excel, whatever you prefer. And then you'll run the analysis and interpret the output. So there are many, many kinds of statistical tests, you all, but anytime we are looking at the output, once you've run the test, whatever it is, the first thing you're going to go do is look at the p-value. So you're going to look at the probability and then the test statistic and the confidence intervals and effect size. So the p-value is going to tell you whether or not you can reject the null hypothesis of no effect. It just simply tells us, is the probability that we would get that test statistic greater than or less than 0 .05? We just say it's either significant or it's not. And then you look at the test statistic and the confidence intervals and the effect size to say if it's a meaningful effect. Because you can get significant results that don't have any practical meaning. So we look at all of those things. Step seven, you make a decision, so you either reject or fail to reject the null hypothesis, and then draw conclusions back to the context of your research question. Ultimately, you will report your results and limitations, and then make suggestions for future research or to address limitations, and all of those things. All right, so then I want to go through some of the most common types of statistical tests and kind of walk you through when you would want to use them with examples and structure. So let's start with a one-sample t-test. You want to use a one -sample t-test when you are comparing one group to a known benchmark value. So, for example, do professional musicians practice more than 20 hours per week? There's evidence out there that gives us an idea of how many hours per week professional musicians practice, and we could take a sample of professional musicians and see how they compare to that known benchmark. So, the structure here is there's no actual independent variable. It's just that benchmark value. So, what the average or the benchmark is. And then there's one continuous. So, on a numerical scale, one continuous dependent variable. So in this case, it would be practice hours. And it's only one single group of subjects. So you look at your single group of subjects and you compare their results with the known benchmark. That's a one sample t-test. Now, there are other kinds of t-tests. So, the next I want to talk about is the independent samples t-test, independent, independent samples. So, this is what we use when we want to compare two unrelated groups. So, for example, if we wanted to compare grip strength between rut climbers and swimmers. Those are two completely unrelated separate groups and we're going to measure them on the same outcome. So for an independent samples t-test, that structure is a categorical IV with two groups. So in this case, it's climbers and swimmers. The continuous dependent variable would be grip strength. And the observations need to be independent of one another. So, the groups are separate. They're not related. Okay. So, we've done one sample t-test, independent samples t-test, and now there's the paired or, well, we'll just call it a paired samples t-test and have multiple names or repeated measures t-test. So you want to use a paired samples t-test when you are measuring the same subjects twice. So anytime I hear of a pre-post type design, I know they're most likely using a paired samples t-test to measure scores before and after some type of intervention. So, for example, memory scores before meditation training versus memory scores after meditation training. So, the structure, it's a within-subjects design. In this case, time or condition, you know, before or after, before and after meditation training is your independent variable. the continuous dependent variable would be the memory scores, and the the structure here involves matched observations. So we need to compare the same individuals pre and post to see if that group changed. Okay, kind of stepping it up here. So if you have two groups, like one or two groups, you're most likely going to use a t -test to make comparisons. But if you have three or more independent groups, so a categorical independent variable with three groups and a continuous outcome variable, you're going to want to use a one-way ANOVA. So you all have dug into a little ANOVA work already. But for example, if you wanted to compare reaction times between folks who worked morning shift, afternoon shift, and night shift. So, you can understand that structure. It's one categorical independent variable. So, shift worked with three levels, morning, afternoon, and night, and one continuous dependent variable. So, reaction times. And those would be independent observations across all three groups. now a two-way ANOVA is when we're looking at two independent variables okay so two factors affecting one outcome so for example looking at the effects of lighting so natural versus artificial lighting and music classical versus jazz versus none on productivity levels See how we just get a little more complex and layered as we move along here? So the structure of a two -way ANOVA is two categorical independent variables, one continuous dependent variable, and independent observations across categories. Okay, the repeated measures ANOVA, you want to use this. Um, so it's, it's essentially like the ANOVA version of a paired samples t-test, but we've got three or more measurements per subject rather than a simple pre -post. We have three or more measurements per subject. So for example, blood pressure checkups across, uh, for quarterly visits. So, if we were trying to look at differences in the same people across four time points, we would need to consider a repeated measures and OPA. So, here the time or condition was the independent variable with three levels. I'm sorry, it was actually four levels. And a continuous dependent variable. Importantly, these are the same subjects throughout. Mixed ANOVA, we got a little familiar here, when combining between subjects and within subjects factors. So, for example, if you wanted to compare language learning progress with using weekly tests between intensive and standard courses. So, the structure of a mixed ANOVA, you must have at least one between subjects factor where you're measuring two separate groups and you must have at least one within subject factor where you're measuring within groups as well and then you need a continuous outcome measure and this involves a mix of both independent and repeated measurements All right, you all. So there's your little quick and dirty review for you. I want to make sure that you're prepared for the week three quiz. So you're going to need to be able to identify the structure of studies. So is it between groups? Are they two completely separate groups? or is it within subjects? Are we looking at changes over time within the same subjects? How many groups were there? How many measurements? What's being compared? These are the questions to ask yourself to identify the study structure. Some of the key phrases and tests that you need to be familiar with are t-tests of course so one sample you're comparing um to a standard or benchmark if you've got between two groups think independent samples if you see before and after or pre -post think paired samples if you've got more than two groups you need or a variety of categorical, multiple independent variables, you're going to need to consider something from the ANOVA family. So, three or more groups, one-way ANOVA. Multiple measurements over time would be repeated measures, ANOVA. Groups measured multiple times would be considered a mixed ANOVA because we'd be looking at differences between and within groups. And then if we're looking at two or more factors affecting an outcome you could have a two-way where all the there are very complicated and open designs for each factor that you add okay then i want to make sure that you're prepared for the assignment this week so for your descriptive statistics you all please pay careful attention to what you do with categorical and continuous variables. So remember, it doesn't make sense to calculate the mean or the average of a categorical variable. So think of what city you were born in for me. I was born in a little place called Chaffee, Missouri. We all have different cities of birth and there is no average city of birth. You can't add together and divide categories meaningfully. So the mean and standard deviation should never be reported for a categorical variable. For those, you want to give frequencies and percentages. So however many people belonged in that category and what percentage of the overall sample they made up. Now, continuous variables that are numerical in nature and indicate an actual scale value, that's where we want those descriptive statistics like means and standard deviations and those things. So, please be careful attention to that. Practice writing integrated summaries combining multiple data points. You'll need to review You'll need to create and review and interpret the histogram overlaying and normal bell curve onto it. And please continue working to master proper APA table and figure formatting, just all the APA stuff. For your outlier analysis, I want to make sure that you understand multiple detection methods. And I just want you to practice justifying outlier decisions and know how to discuss the characteristics of your distribution and to explain any associated practical limitations based on any flaws in the data. So for statistical tests, you need to know how to select the appropriate test type. Practice writing complete hypotheses, review the checking of your assumptions, make sure you check all of the associated assumptions and report that, and more APA stuff. Work on mastering reporting of your APA results. all right so now i'm just going to quickly walk you through the assignment and i want to point out how many points these are worth you all so pay attention to the point value of questions um and and make sure that your answers are uh appropriate for those number of points so for question one you're going to describe the variables sound environment and test score You will generate frequency tables and bar charts for all categorical, so nominal or ordinal variables. For continuous scale variables, you need to generate and interpret those descriptive statistics. Make sure you include your histograms with the normal curve, the density curve, superimposed over each variable. and then summarize the data in a paragraph, highlighting the key findings, so the shape of the distribution, central tendency, variability, all of that stuff. For question two, you will analyze whether the test score variable has any significant outliers. So, you will identify potential outliers, including your output, and writing a brief justification for whether or not these are true outliers. So think Z scores and box plots are the way that we identify outliers. And then you need to justify whether any potential outliers should be kept, removed, or transformed. So I want you to tell me what you would do, but do not remove or transform any of the outliers. Just simply report what you would do based on your analysis. Okay, question three. You want to determine whether participants in your sample perform significantly differently from a benchmark average score of 80 on the math exam. So that's the variable test score in the data set. So you'll answer these questions. What statistical test would you conduct and why? Run the test in your software of choice and paste your output into the document. Then write up your findings as if you were reporting them in a journal article perfectly following APA guidelines. Moving on, the next question is test -taking environment, so the variable environment, associated with math test scores. So are they associated? You will need to identify the appropriate statistical test for this analysis and explain your choice. You need to identify the independent variable, if there is one, and dependent variable. It might not be a true independent variable in this situation where we're just looking for an association. Conduct the test, paste your output in, and then write up the findings as if you're reporting them in a journal. Question five. The scenario is that participants rated the attractiveness of a potential romantic partner while under the influence of alcohol. So there's the variable name, attract underscore ALK. And again, when completely sober, so underscore SOB. Ratings were on a scale from 1 to 14 where higher values indicate greater attractiveness. These variables should be treated as scale or continuous variables. And you are asked to analyze whether alcohol influenced ratings of attractiveness. So, what test would you conduct and why? Conduct the test, paste in your output, and then give me a little formal write-up of the results. Alrighty, so then question six, you will see this output here. One group of children had 30 minutes of unstructured playtime, which is labeled unstructured, while another had 30 minutes of an observer directing playtime. So, that was structured. So, all children were then asked to draw a picture of something they had imagined, and then independent observers rated the creativity of drawings. So, higher scores meant more creativity. The researcher conducted a t-test for independent means on the data the spss output is given here and what i'm asking you to do on this one is explain what levine's test for equality of variances tells us why it matters and how it affects interpretation use the um use the information provided under levine's test for equality of variances so you'll see that there and then describe the results of the t-test in a way that would be clear to a reader who is unfamiliar with the data set so use proper construct names rather than using the variable name so you would say those in structured play versus unstructured play um i guess they're named that way but i was thinking really these others the attractiveness and all of those. Rather than use the codebook name, use the actual construct name when you talk about that stuff to make it easier for the reader. Alrighty, then you will get a look at these estimated marginal means. So, a psychologist studied the effects of both location and noise and how those things impacted math level scores. So, the results are displayed in that figure, and you are asked to indicate what statistical test was performed. Make sure you base your answer on the study description as well as the figure and explain your reasoning. You will need to properly label and number the figure according to APA guidelines, so don't forget to do that. Write the hypothesis for the impact location and noise level on exam scores. And then based on the figure, describe the overall results of the study based on this figure. Okay. Question eight. A psychologist is studying the effect of sleep deprivation on reaction times. Participants are asked to complete a reaction time tasks after three conditions. One, a full night's sleep, two, 24 hours of sleep deprivation, and three, 48 hours of sleep deprivation. Reaction times in milliseconds for each condition were recorded as continuous variables. The researcher hypothesized that sleep deprivation will significantly impact reaction times. So based on this scenario, report the statistical test that you would use and explain why it's appropriate. Write out the hypotheses for the test you selected and clearly state both the null and alternative hypotheses. And you don't have to run this analysis but if you were to run this how would you check assumptions so for example sphericity and what would you do if they are violated so just that's a hypothetical there all right you all i know it's a lot your head's probably swimming i want to take a minute to point out all of the additional resources that are in blackboard for you this week. So those things are there. Feel free to dig through those, reach out to me if you have questions. If you need a little more guidance on your analyses, there are so many good resources out there. So I've only been using Jamovi about a year myself. And anytime I'm doing a new analysis, I go spend a little time on Google. Jamovi has a lot of resources directly themselves. So this is just a little screenshot of a search where I found information quickly on descriptives and mixed ANOVA and Jim Obey. So don't be afraid to look for additional resources or again, just ask me. A lot of folks struggle with APA writing and I want to recommend Purdue OWL to you all. This is one of the best free writing sources outside of our campus resources, obviously. But this is a really great source of APA style formatting guidelines, all of that stuff. So if you still struggle with APA and just writing up your results in that expected format. Go spend a little time digging around on Purdue Owl. All right. And then try to keep it light here at the end. So the hedgehog, that cute little hedgehog is cheering for you because you can do everything. I believe in you all. You can do it. And also this is me every morning too, like this poor little bunny. I'm like, listen, you can do it we can all do it i believe in you okay come out of here stop my screen share and check in with my people how we doing oh that was that was great yeah i um watched the uh the pieces to that i was just wondering like how to do it on jamoby i know we had um dr tedes last uh and it was great because he would show us the like what he how he would do it and we'd go back and do that so that was fun but like i i'm not quite there yet to where i know how to do everything on jamoby so that's why i'm like okay i'm gonna give it another shot and yeah so just remember this week steve you're not asked to do, like the things that you actually need to analyze in Jamovi are, I promise it's not going to be hard or bad. They're some of the most basics. So if you get stuck, reach out to me. Okay. But like that page I showed with, like as you, Jamovi resources are so great because I do screenshots and I can, like last week, I walked you all through the whole assignment last week. But the more practice you get, the better. And going out and finding some of your own additional resources like that is actually going to help you more. So I highly recommend doing a little, if it's time to work on the descriptives part, Google descriptive statistics in Jamovi. so once figuring out which test you need to use for the question i think is going to be the most challenging part and then actually running the analysis is easy interpreting the results might be a little tricky but you've got this you're like if you say so huh yeah i feel like the hamster but okay i was a guinea pig the guinea pig in the mirror i get it but but you can do it we are all that we are all that poor little thing we we all get down but just start early reach out if you have questions i mean i'm not i can't like give away answers or anything but i will tell you if you're going in the right direction if you email me before friday at like noon or one i will make sure i answer you and give you as much clarity as I possibly can. Fair enough. Yeah, that's fair. And I'll get on it. Okay. Yeah, sounds good. Yeah, just start early, practice, and email me if you need anything. All righty. Anyone else have questions, you all? No, thank you very much. Thank you, Dr. Ratliff. You're welcome, you all. Like I said, I know it's challenging. You're all doing the right things. I'm just, I'm a 24-hour response away if you get stuck. So send it to me sooner rather than later if you have questions, and I will help you move right through. Okay? Much appreciated. Yeah. All right. You all have a good week. I'll see you soon. You too. Thank you. Bye-bye. Bye. You too. Thank you.