Correlation Is Not Causation — Research Design, Reliability, Validity, and Ethics Explained

In 1974 a psychologist named Elizabeth Loftus showed participants a film of a car accident and asked one simple question. She used the word "smashed" with one group and the word "hit" with another. One week later the "smashed" group was significantly more likely to report seeing broken glass in the film.

There was no broken glass.

One word in a question changed what people remembered seeing with their own eyes.

That experiment didn't just change how courts think about eyewitness testimony. It demonstrated something that every clinician who reads research needs to understand: the way you design a study determines what the data can actually tell you. Measurement is not neutral. Every research design decision shapes the findings that come out.

This is why research methodology matters — not as abstract academic content, but as the framework that allows you to look at any study and ask the right questions about what it actually proves.

In this episode of Licensure Lifeline we make the most dreaded domain on every licensing exam genuinely learnable — no math required.

What we cover:

🧠 The history — Elizabeth Loftus, the misinformation effect, and why one word in a question can change everything a participant remembers

📋 Research design — experimental research and what random assignment actually allows you to conclude, quasi-experimental research and its limitations, and non-experimental designs including correlational, descriptive, and ex post facto research

🔬 Reliability — the four types you need to know: test-retest, internal consistency, inter-rater, and parallel forms — what each measures and how they differ

✅ Validity — content validity, face validity, concurrent validity, predictive validity, and construct validity — the hierarchy of evidence and the distinction that costs the most exam points

⚖️ Ethics in research — the Tuskegee Syphilis Study and why it matters, the Belmont Report's three foundational principles (Respect for Persons, Beneficence, Justice), informed consent requirements, and the role of the IRB

🎯 The four-question exam method — every research and statistics question on a licensing exam is asking about design and what it can conclude, reliability, validity, or research ethics. Identify the category. Apply the framework.

Five exam-style multiple choice questions covering experimental research design and causal conclusions, test-retest reliability, face validity versus construct validity, the Belmont Report's three principles, and ex post facto research and its limitations.

Also in this episode:

A rapid review commissioned by the National Institute for Health and Care Research examining interdisciplinary mental health research — and what it tells us about who gets included in research and why that matters methodologically. And the July 2026 APA journals featuring new research on the impact of redlining on mental health — a perfect real-world example of what correlational research can and cannot establish.

Want to go deeper? This week's Licensure Lifeline newsletter covers statistical concepts the exam tests most — measures of central tendency, standard deviation, Type I and Type II errors, effect size, statistical significance, and a research critique exercise where you evaluate a real study abstract using today's frameworks. Always free — link in the show notes.

The cheat sheet, 12-question deep dive quiz, and live study session are inside Licensure Lifeline Circle — home of the Licensure Lifeline 90-Day Study System. Fourteen-day free trial. And the founding member offer opens next week — stay tuned. Link in the show notes.

Have a story to share — an exam win, a clinical moment, something from supervision that changed how you see the work? Email us. You might appear in The Fifty-First Minute. Link in the show notes.

And before you close this page — if this podcast has helped you study, share it with one person who's preparing for their exam right now. A five-star rating on Apple Podcasts takes ten seconds and helps other pre-licensed therapists find the show.

Resources:

📥 Free Study Roadmap PDF — 12-domain NCE exam prep guide
 → licensurelifeline.com

📩 Free weekly newsletter
 → licensurelifeline.com

⚡ Licensure Lifeline Circle — 90-Day Study System
 → licensurelifeline.com

Transcript

Copy Transcript Copied

SPEAKER_00 0:09

In 1974, a psychologist named Elizabeth Loftus ran an experiment that would change how courts, clinicians, and researchers think about memory and about the reliability of the data they collect. She showed participants a short film of a car accident. Then she asked them questions about what they'd seen. But here's the critical part. She asked different groups slightly different questions. One group was asked, how fast was the car going when they hit each other? Another group was asked, how fast were the cars going when they smashed into each other? One word changed. Hit versus smashed. That's all. The group that heard smashed estimated the cars were traveling significantly faster than the group that heard hit. And when Loftus followed up a week later, asking whether there had been broken glass in the film, the smash group was significantly more likely to report seeing broken glass. There was no broken glass in the film. One word in a question changed what people remembered seeing with their own eyes. Loftus called this the misinformation effect. And what it demonstrated with elegant, uncomfortable clarity was the way you asked a question changes the answer you get. That measurement itself is not neutral. That every research design decision, how you word a question, who you include in your sample, what you choose to measure, shapes the data that comes out. This is why research methodology matters, not as abstract academic content, as the framework that allows you to look at a study, any study, including the ones that will influence your clinical practice, and ask the right question about what it actually proves. Today we're going to build that framework on that framework. And we're going to make it as concrete as exam and exam ready as possible. Welcome to the Lecture Lifeline podcast. This is your host, Matt Lawson. Today we're going to be looking at research and statistics. Yes, I know we do not get into this field to deal with numbers, but this is a nice and unsatisfying reality of the work that we do. Here's what I want you to hear before we start. Research and statistics is not a math test. The licensed exam is not going to ask you to calculate a standard deviation by hand or run a regression analysis in your head. What it is going to ask is whether you understand what different research designs can and cannot tell you, whether you know the difference between reliability and validity and why both matter. Whether you can look at a research finding and evaluate whether it means what it claim, what it means what it claims it means. That's critical thinking. That's clinical clinical literacy and it's genuinely learnable. Um, if you've been dreading this domain for months, um, I'm telling you right now that this is one of those things that you know we we have to know. Um, it really actually helps if you read white papers, if you read studies, and as part of your practice toward this exam, um, just reading over these pretty analytical, research-heavy papers actually really helps with getting this stuff down. Before we get into today's content, three quick things. If you found this podcast through a search or a share and this is your first episode, welcome to the show. Um go to leicturelifeline.com and download the free study roadmap PDF before doing anything else. It's great. Um, it's you know, it goes over all the different domains and how to kind of prep for them. Um, it makes a really good campaigning for these shows. Um, but it's a free instant download. Um, you just put in your email address. This will also get you the free newsletter, which is the second thing I want to mention today. Um, this comes out every week alongside each episode. It's a deep dive into the things we talk about here, along with some other things that go along with the things that we talk about here. Provide some clinical examples, case vignettes. So if you're not yet on this list, please take a look again at leicerlifeline.com. That's where you'll you can get access to both the roadmap and the newsletter. Lastly, and you know, I want to say this specifically if you're like three to six months out from your exam, take a look at Leicester Lifeline Circle Group. This is a place where you know I I put up more information, again, more deep dive-y type stuff. Also, you get community support there. Um, you know, it's really exciting to see the people that have joined already. Thank you so much if you have joined the group. You get a two-week free trial when you sign up. You don't have to put a credit card in, anything to do the two weeks. Then you can decide if this is something that is going to benefit you at the end of those two weeks. But in that within that group, I have quizzes, I have different study guides. Um, I even do a weekly um study hall where I'll go over the topic for the week and then just give people a place to kind of do some QA stuff around the different topics or just get some support around the work that you're doing. So, you know, this week's study, the study halls are 12 o'clock mountain time every Friday. Um, basically, I take my lunch break and do these study halls for people. They're really great. I'm really enjoying them right now. But it's really fun evolving this podcast into something that is really, I'm hoping, going to just support people to become clinicians. Um, I'm growing this stuff and putting more things in every day. I'm constantly working on things, and I appreciate everybody that listens to the podcast and everybody that's joined the different groups in the newsletter. Thank you so much, and I look forward to seeing more people there. Okay, let's get into the news. First story. Um, we're going to talk a little bit about how research gets done and why methodology matters. A rapid review of interdisciplinary research practices in mental health has been commissioned by the National Institute for Health and Care Research in the UK, examining how the interaction of different disciplines and the knowledge and perspectives of people with lived experience of mental health problems can create new solutions. The collaboration of medical and social sciences with the arts is one significant area, particularly when engage engaging people with experience of trauma in research. Why this matters to you? The story is about research design in real time. The question of who gets included in mental health research and in what role is a serious debate. When people with lived experience are included as researchers rather than just subjects, the research questions change, the measurements instruments change, and the validity of the findings often improve. That's the practical application of everything we're we're covering today. You know, for the longest time, a lot of the research was done on pretty specific groups, um, intended to be done like, you know, you look back in the 60s, 70s, um, intended to be done on white men, um, oftentimes white men that were in college, college-age men, because they're kind of like captured individuals on these college campuses. And, you know, for many researchers, it just kind of that's just kind of what they did. Um, so, you know, we're as I kind of said in the beginning, it does, it matters. Like lived experiences for everybody can be very different and change the results very much so, depending on who is in those studies and who's running those studies. So this is really neat that this is speaking to kind of that and putting um more diversity within these studies to just speak to a broader range of individuals, oftentimes individuals that could really benefit from mental health care. Second story. The July 2026 issue of APA journals features new research on the impact of redlining on mental health and research on divergent patterns of genetic overlap between severe mental disorders and metabolic markers. The redlining study is a perfect example of what correlation research can and cannot establish. It can show us that neighborhoods historically subject to redlining have higher rates of certain mental health conditions. It cannot, by design, establish the that redlining caused those outcomes. Because you cannot randomly assign people to neighborhoods. Um understanding this distinction between correlation and causation, between what a study found and what it proves, is one of the most important research literacy skills you can develop. And it is directly testable on your licensing exam. If you don't know what redlining is, really important that you take a look. Um, basically, this was a push kind of back in the day where um certain neighborhoods, specifically neighborhoods with black and brown individuals in them, um, were basically cut off from services. Um, things like transportation, trains, things like that. And these areas, in some cases, were actually um deemed like kind of disaster zones. And, you know, it did. If you you can imagine the impact that it might have had on a population, not just then, but for generations to come. Um, because you still see the impact of redlining in a lot of neighborhoods throughout the country. So very interesting topic to take a look at. Um, you know, you can start to see again where these uh different things kind of impact mental health, mental health in those areas. All right, let's get into research design. The first concept within this uh these topics that we're going to talk about today, research design is the architect of a study. The decisions a researcher makes about how to collect data and determine what the data can actually tell us. Three major categories worth knowing for your exam: experimental research, the gold standard. The research manipulates an independent variable and measures its effect on a dependent variable. Participants are randomly assigned to conditions. Because of random assignment, the researcher can establish cause and effect. The randomized control trial or the RCT is the most rigorous form of experimental research and the design that produces the strongest evidence for clinical intervention. Some key terms to know. First up, independent variable, the thing the researcher manipulates, the treatment, the intervention, um, the intervention, the condition. Dependent variable, the thing researchers measure, the outcome. Random assignment, the participants are randomly placed into experimental or control groups. This is what allows causal conclusions. Then control group, the comparison group, receives no treatment, a placebo or treatment as usual. And then finally, experimental group receives the intervention being tested. I know these this was one of the areas that really caught caught me, just being able to keep these different terms straight. And it is like this is one of those ones that I, you know, I had to come up with some type of mnemonic for it, um, just to kind of help me remember and keep these things straight. This is the one that gets a lot of people tripped up on the exam, is just remembering what these different um groups are and variables and things like that. So it's worth coming up with your own mnemonic for it. Random assignment is what distinguishes experimental from quasi-experimental research. And, you know, we live in a day and age where research is one of those, you know, somehow it became this political thing, and it's kind of sad. But, you know, we're studying, we're studying things here. We're, you know, it's like that's what we're doing. We're studying, we're researching things, and you know, we're using a lot of it's important that we use this kind of standard around the research. Um, but quasi-experimental research, this is this looks like an experimental research, but lacks true random assignment. So participants are placed into groups based on existing characteristics, their diagnosis, their school, their neighborhood, rather than by random chance. Um, you can establish relationships between variables, but cannot establish causation with the same confidence as true experimental research. Kind of under the similar umbrella as quasi-experimental research is non-experimental research. So there's no manipulation here, no assignment. The researcher observes and measures variables as they naturally exist. Three important subtypes exist under the non-experimental research. That's correlational research, exam, exams the relate examines the relationship between two variables, establishes whether variables are related and in what direction, but cannot establish causation. The redlining study from today's news segment is correlational research. Then you have descriptive research, describes a population or phenomena. Um surveys, case studies, naturalistic observations tells you what it what is, not why. Then ex post facto research examines existing conditions after the fact. The researcher identifies groups based on something that already happened and looks backwards at the potential causes. These cannot establish causation because the group was not randomly assigned. So as you see, the random assignment piece is significant. So anything within the exam where you don't see random assignment, you have to start thinking of these other non-experimental or quasi-experimental types of research as part of the answer. Next, let's take a look at reliability. Reliability is the consistency of a measurement. A reliable measure produces the same result when applied to the same thing under the same condition. Think of this, think of it this way: a reliable scale gives you the same weight every time you step on it, unfortunately. Um, an unreliable scale gives you a different number each time, even if you haven't gained or lost any weight. So, yeah, you know, you you get on a scale, you want to see, hopefully, you want to see some changes in the direction you're hoping for. Um, but you know, what this is saying is, you know, if you're supposed to weigh 230 and all of a sudden you're wearing 240, or the next year weighing 220, it doesn't make sense within 24 hour periods you have fluctuations like this. So that's where reliability of that scale comes into question. So there's four types of reliability on it that you're gonna need to know for your exam. Test, retest reliability, the same test administered to the same people at two different time points, produces consistent results, measures stability over time. I should mention that all of these are really taking a look at the test and its reproducibility, right? Um, so you know, this is kind of like how tests are looked at, different tests are looked at, and they you know, they undergo the scrutiny around like, can this test produce what it's supposed to produce? Um so they look at it through like a reliability lens. The other type here is internal consistency. The items within a test are measuring the same same construct. If a depression scale has 20 items, they should all correlate with each other because they're all supposed to be measuring the same thing. Um, Crombracht's um alpha is the most common statistical measure of internal consistency. Um, in a rater reliability, two different raters or observ or observers applying the same measure to the same thing arrive at the same result. Critical for any assessment that involves clinical judgment. So, you know, you can't just create an assessment and put it out there. Like there are some things that you kind of need to do. So a lot of these internet assessments or assessments that you get like out of magazines or whatever, um, you know, they don't they don't scrutinize them at all. So, you know, you always have to kind of take them with a grain of salt. Um, you know, that's that I'm sure everyone listening to this podcast knows this, but you know, I there there has been more than one occasion where I've been working with somebody and they bring in an assessment that they took through some, I don't know, through some magazine or something as part of you know what they're dealing with. And it's it's just one of those things that there's some entertainment value there for people, just in general. Um, but you know, some people do take these things really to heart and want to understand what it means. So if you see something like that, um keep that in mind. Lastly, you have parallel forms reliability. This is where two different versions of the same test produce consistent results. This is used when you need to retest someone and want to avoid practice effect. Reliability is necessary but not sufficient for validity. A test can be perfectly reliable, consistently giving the same result, and still be measuring the wrong thing entirely. So a broken thermometer that always reads 98.6 is perfectly reliable, but it's just not valid. Speaking of that, let's go ahead and get into validity. Validity is whether a measurement actually measures what it claims to measure. A valid depression scale measures depression, not anxiety, not social desirability, not something else that happens to correlate with depression. Validity is more complex than reliability because it's not a single property, it's a family of related questions about what an instrument is actually measuring. So let's look at some different aspects here. So content validity. Does the test cover a full range of constructs it's supposed to measure? A depression scale with only sleep and appetite items has poor content validity. It's missing most of what depression is. Next we have face validity. Does the test appear to measure what it claims to measure? The weakest form of validity here. Something can look valid without being valid, but worth knowing as a term. Criterion val validity does the test predict or correlate with the external criterion? So two subtypes with under the criterion validity, concurrent validity, the test correlates with another measure of the same construct administered at the same time. Um does this new depression scale score correlate with clin clinically rated depression severity right now? You also have predictive validity, the test predicts a future outcome. Does this depression scale depression scale score at intake predict treatment responses six months later? Um finally, construct validity, the broadest and most important type. Does the test measure the theoretical constructs it's supposed to measure? This requires evidence across multiple studies and methods, not just one correlation. Um the exam distinction that costs the most points, reliability versus validity. That's this is gonna be the big biggest one that you guys gotta get straight. Know that reliability is about consistency, validity is about accuracy. Know that you can have reliability without validity, but you cannot have validity without reliability. I'm so sorry, but this it is like I'm I'm getting flashbacks um from trying to remember these. But in an inconsistent measure can be valid because it gives different results each time. It can't be accurately measured, it can't accurately measure anything. All right, I'm glad we got through that. Um, I wish this got more exciting, but you know, these are the these are things we gotta know. Um, and it is, it's absolutely, I still I regularly still read studies um just to kind of keep up on all the things. Um, you know, it's gonna be one of those things that is gonna constantly play in the background of your head once you get it down. Um, but you do. Like, this is how we keep our ideas fresh. This is how we like knowing like reading these studies is how we keep our ideas fresh, how we keep our our understandings of things fresh. So we just have to know this stuff. So let's get into ethics and research. Research ethics became a formal field after a series of historical violations so serious that fundamentally that they fundamentally changed how research is conducted. The most in his important historical reference for your exam is the Tuskegee syphilis study in 1932 through 1972. The U.S. Public Health Service studied the natural progression of untreated syphilis in around 400 black men in Alabama. Without their informed consent, without treating them even after penicillin became available in 1947. The study ran for 40 years. It is one of the most significant ethical violations in the history of American research and directly led to the establishment of modern research ethics and requirement. The reason that I always do a historical kind of intro for this is it is so important as clinicians that we understand history as it associates with psychology. I am kind of a history nerd anyway, so that's another selfish reason for me, including the his historical stuff here. But, you know, the reason something like the Tuskegee Syphilis study still is still relevant today, if you look back at what happened during COVID and the high rates of people in black communities, black and brown folks, that just refused to take the vaccines, this is one of the reasons why. This is where a lot of distrust was sown in those communities. I mean, we're talking 1932 to 1972. I mean, I was born in 1974. Um, you know, this it wasn't a crazy long time ago. This is stuff that's pretty recent. And, you know, there's still a lot of people. Um, there are people that I work with, um clinicians that I work with that live through these moments. And this is, you know, it's it it shaped how they thought about being able to trust the government. But from these the syphilis studies came the Belmont Report, 1979. Um, it was published in response to the to the Tuskegee studies and and other violations up to this point. The Belmont Report established three foundational principles of ethical research that are directly testable on the licensed exam. Respect for peers. Individuals are autonomous agents capable of making their own decisions. Research participants must give informed consent. Those with diminished autonomy require special protection. Beneficence. Researchers have an obligation to maximize benefits and minimize harm. Do no harm. Um, and then maximize possible benefits. Then justice. The benefits and burdens of research should be distributed fairly. The people who beat the um bear the risk of the research should also be among those who benefit from it, not populations that are simply convenient to study. But now we have things like the Institutional Review Board, IRB. Um this committee that, you know, this is the committee that reviews research proposals to ensure that they meet ethical standards before the study even begins. Um, knowledge of the IRB requirement is testable. Also, informed consent. Participants must be given sufficient information about the study to make voluntary informed decisions about participation, key elements, um, the nature of the study, the risk and the benefits, the right to withdraw at any time without penalty, and how confidentiality will be protected. But again, you don't have to go too far back in the history of psychology to just see how wild these studies were, how just kind of wild, wild west, um, a lot of these studies that were conducted before these things were put in place to protect the people that were in the studies. Um, you know, there are a lot of things that were done that were just completely unethical to people. And, you know, these different things that we have in place now are really meant to protect. All right, let's get into some questions. A researcher wants to test whether a new CBT protocol reduces anxiety symptoms more effectively than standard treatment. She randomly assigns 60 participants to either the new CBT protocol or standard treatment as usual and measures anxiety symptoms before and after the eight-week intervention. Which of the following most accurately describes this research design and what conclusions can it support? Is it A, correlational research? This can establish a relationship between treatment type and anxiety reduction. B quasi-experimental research. This can suggest but not confirm a casual relationship. C. Experimental research, random assignment allows the researcher to draw casual conclusions about the effects of the CBT protocol on anxiety symptoms, or D. Descriptive research. This describes anxiety symptoms and how they change across two different groups. The answer to this is gonna be C. The defining feature of experimental research is random assignment. Because participants were randomly assigned to conditions, systemic differences between the groups are controlled for, meaning any differences in outcomes can be attributed to the intervention rather than the pre-existing group differences. Memory strategy for this one random assignment equals experimental equals causation possible. No random assignment equals quasi-experimental or non-experimental equals correlation only. The word randomly in a research description is an exam signal for experimental design. Question number two. A school counselor administers the same anxiety screening measure to the same group of students three weeks apart. The scores were highly consistent across the two administrations. Students who score high the first time score high the second time, and students who score low the first time score low the second time. This finding most directly supports which psychometric property of the measure. Is it a content validity? The measure covers the full range of anxiety symptoms, B construct validity, the measure assesses the theoretical construct of anxiety, Cest relia test retest reliability, the measure produces consistent results across time, or D. Concurrent validity. The measure correlates with another anxiety measure administered at the same time. The answer to this one is gonna be C. Test retest reliability. Test retest reliability specifically refers to the consistency of a measure across two administrations at different time points. The scenario describes exactly this the same measure, the same group, three weeks apart, consistent results. Content validity would require evaluating the breadth of the content covered. Construct validity would require evidence across multiple studies. Concurrent validity would require comparison to another measure at the same time. Memory strategy for this one. Test retest equals same measure, same people, different times, consistent results. The re in retest is your signal. It is being administered again. Stability over time equals a retest reliability. Question number three. A researcher develops a new measure of therapeutic alliance. The items on the scale are reviewed by 10 experienced clinicians who all agree that the items look like they assess therapeutic alliance. However, subsequent research reveals that scores on the new measure do not correlate with treatment outcomes, client retention, or any other measure theoretically related to therapeutic alliance. Which of the following most accurately describes the psychometric status of this measure? Is it A, the measure has a strong construct validity? Experts agreement this is expert agreement confirms it measures through therapeutic alliance. B the measure has face validity but poor construct validity. It appears to measure therapeutic alliance but does not have but does not behave as a valid measure of construct as a construct should. C, measure has strong content validity. The items cover the relevant domain as confirmed by expert review, or D. The measure has strong concurrent validity. The expert, clinician, um validated the items against their clinical experience. The answer to this one is gonna be B. Face validity. The degree to which measures uh the measure a measure appears to assess what it claims is the weakest form of validity. Expert agreement, expert agreement that items look right establish face validity only. Construct validity requires the measure to behave as theory predicts, correlating with related constructs, predicting relevant outcomes, and discriminating from unrelated constructs. A measure that experts agree look valid but doesn't correlate with anything theoretically related as face validity without construct validity. This is precisely why face validity is considered insufficient on its own. Memory strategy for this one. Face validity equals looks valid looks valid. Face validity equals looks valid. Construct validity equals behaves valid validly. You can have face validity without construct validity. Man, this stuff is really hard to say over and over again. Question number four. In 1932, the Tuskegee syphilis study research studied the natural progression of untreated syphilis in black men in Alabama without their informed consent and continued the study for 40 years without treating participants even after an effective treatment became available. The ethical framework developed in response to violations like the Tuskegee, like Tuskegee identified three core principles of ethical research. Which of the following correctly names all three principles from the Belmont report? Autonomy, non malfeasance, um fidelity, B respect for person, beneficence, justice, c informed consent, confidentiality, volunteeriness, or D beneficence, non malfeasance, or autonomy. The answer is gonna be B. Respect for persons, beneficence, and justice. The Belmont Report in 1979 established three foundational principles. Respect for persons, individuals are autonomous agents who must give informed consent, beneficence, maximize benefits, minimize harm, and justice, fair distribution of research burdens, and beneficent benefits. Option A and D contain principles from the bioethics more broadly, but not the specific Belmont framework. Option C list elements of informed consent, which is a mechanism of respect for a person, not a separate principle. Memory strategy here is going to be Belmont equals RBJ, respect for persons, beneficence and justice. RBJ for the Belmont. Alright, last question. A researcher conducts a study examining whether childhood trauma history is associated with adult depression. She recruits two groups, adults with depression and adults without depression, and asks them about childhood trauma experiences. She finds that the group with depression reports significantly higher rates of childhood trauma. Which of the following most accurately describes what this study can and cannot conclude? Is it A, the study establishes that childhood trauma causes adult depression? The groups were clearly different on the outcome variables. B, the study establishes a correlational relationship between childhood trauma history and adult depression, but cannot be established causation. There was no random assignment and the design was retrospective. C, the study establishes predictive validity, childhood drama predicts adult depression. Or D, the study is not valid because it relied on self-report data about childhood experiences. The answer is going to be B. So this is going to be exo um ex post facto research. The researcher identifies groups based on an existing condition, depression versus no depression, and looked backwards at potential causes, childhood trauma, because participants were not randomly assigned, and because the trauma history preceded the study retrospective retrospectively, the design cannot establish causation. Alright, memory strategy for this one. No random assignments plus looking backwards equals cannot establish causation. Let's say that again. No random assignment plus looking backwards cannot establish causation. Alright, folks, that is gonna be it. Thank you for sticking with me through this one. I know this is not people's favorite topics to get into, but is it a necessary one for the test? It's just gonna give you points if you can get this stuff down straight. Um, so a couple of things I want you to take away. Number one, the design determines what you can conclude. Experimental with random assignment means causation is possible. Everything else means correlation only. That distinction is the foundation of research literacy and will appear on the exam. Number two, reliability and validity are not related, but distinct. Reliability is consistency, validity is accuracy. You can have reliability without validity. You cannot have validity without reliability. Remember, the broken thermometer that always reads 98.6 is reliable but not valid. Lastly, the Belmont Report's three principles, respect for persons, beneficence, justice, are the ethical foundations of research. Just know them. Know that the Tuskegee experiment is the historical violation that prompted this. That the combination appears on the licensing exam regularly. Before we close the episode for today, I want to make sure you're leaving with more than you came in and what we covered today. Um, you know, don't forget about the roadmap. Again, you can go to licentiellifeline.com, download that there. This is a free PDF that just gives kind of a roadmap and covers the main topics of like the NCE. Um and just gives you an idea of how to space things out with your studying. Um, you'll also be able to get signed up for the newsletter there as well. Um, this is a free um newsletter that comes out weekly, delivered to your inbox, just goes deeper, um, highlights everything we talked about in the podcast, but then goes deeper just with more information for you when you're studying. And then don't forget about the Lester Lifeline Circle Group. I want you to take a look at this. This is a community that we are building that I'm hoping is going to just provide just an extra level of support for individuals to give them a place to come to study, to get support, to ask questions, to take quizzes, things like that. Give it a look. Um, again, you can find that through the LibsterLifeline.com website. Until next time, everybody, thank you for being here. Thank you for being on this journey and never stop learning.