WEBVTT 00:00:00.000 --> 00:00:02.950 Okay, open your notebook. I've got the large 00:00:02.950 --> 00:00:05.549 coffee black, no sugar, because we are in crisis 00:00:05.549 --> 00:00:07.990 mode. And I've got the highlighters. We have 00:00:07.990 --> 00:00:11.189 exactly one night to cram all of this into our 00:00:11.189 --> 00:00:13.689 brains before Professor Pervitura's final exam. 00:00:14.089 --> 00:00:16.390 Welcome to the psychology undergrad. Breathe. 00:00:16.670 --> 00:00:19.149 Just breathe. We are going to be fine. It feels 00:00:19.149 --> 00:00:20.870 like, you know, a mountain of information, but 00:00:20.870 --> 00:00:23.449 it's really just a system. Think of this less 00:00:23.449 --> 00:00:26.690 as a panic attack and more as a structured review 00:00:26.690 --> 00:00:28.469 session. We've got the syllabus, we've got the 00:00:28.469 --> 00:00:31.370 notes, and we have the exam must know terms sheet. 00:00:31.469 --> 00:00:34.030 I'm trying to channel that calm. I really am. 00:00:34.229 --> 00:00:36.869 But look at this sheet. It's huge. It's like 00:00:36.869 --> 00:00:39.070 he took the entire dictionary, circled the most 00:00:39.070 --> 00:00:41.869 boring words and said, memorize this or fail. 00:00:41.969 --> 00:00:44.649 I mean, epistemology, operational definition, 00:00:45.530 --> 00:00:47.850 homelessness. Wait, is that even on there? No, 00:00:48.109 --> 00:00:49.590 thankfully that one isn't on the list. Look, 00:00:49.710 --> 00:00:51.369 it's not about memorizing definitions. That's 00:00:51.369 --> 00:00:53.149 the trap everyone falls into. If you just memorize, 00:00:53.270 --> 00:00:54.969 you'll blank out when you see a scenario question. 00:00:55.310 --> 00:00:57.789 We need to connect the dots. We need to understand 00:00:57.789 --> 00:01:01.229 the story of research methods. The story? OK, 00:01:01.450 --> 00:01:04.370 I like stories. Stories I can remember. So our 00:01:04.370 --> 00:01:07.129 mission for this session is simple. We go through 00:01:07.129 --> 00:01:10.010 the must know list. We define the terms. But 00:01:10.010 --> 00:01:12.510 we also translate them into actual human speak. 00:01:13.349 --> 00:01:16.599 No textbook jargon unless we Like, explain it 00:01:16.599 --> 00:01:18.640 first. Exactly. And we're going to look at this 00:01:18.640 --> 00:01:20.819 through the lens of everyday life. We'll use 00:01:20.819 --> 00:01:23.159 analogies, pop culture, and maybe even touch 00:01:23.159 --> 00:01:25.459 on how these things relate to bigger questions 00:01:25.459 --> 00:01:28.340 of, you know, morality and belief. Because honestly, 00:01:28.579 --> 00:01:30.420 that's where the stuff actually gets interesting. 00:01:30.519 --> 00:01:33.079 Right. Because as dry as research method sounds, 00:01:33.719 --> 00:01:35.939 it's basically the rule book for figuring out 00:01:35.939 --> 00:01:38.379 why humans do the weird things we do. Yeah. It's 00:01:38.379 --> 00:01:40.819 the user manual for the human experience. Precisely. 00:01:41.019 --> 00:01:43.280 It is the rules of the game. If you want to play 00:01:43.280 --> 00:01:45.659 the game of science, you have to know the rules. 00:01:45.799 --> 00:01:47.099 Okay, let's start at the very beginning then. 00:01:47.159 --> 00:01:49.599 Part one on the sheet. How we know what we know. 00:01:49.719 --> 00:01:51.959 It starts with non -scientific ways of knowing. 00:01:53.060 --> 00:01:56.900 So, this is the stuff humans did for thousands 00:01:56.900 --> 00:01:58.900 of years before we had labs and white coats, 00:01:59.040 --> 00:02:01.540 right? Right. These are the shortcuts our brains 00:02:01.540 --> 00:02:04.019 use to make sense of the world. The first one 00:02:04.019 --> 00:02:06.640 on the list is tenacity. Tenacity. That sounds 00:02:06.640 --> 00:02:09.280 like a good quality, like grit. He's a tenacious 00:02:09.280 --> 00:02:12.189 worker. In life, sure. But in this context, it's 00:02:12.189 --> 00:02:15.110 actually a bit of a cognitive trap. Tenacity 00:02:15.110 --> 00:02:17.330 is believing something simply because it has 00:02:17.330 --> 00:02:20.150 always been believed. It's habit. It's superstition. 00:02:20.449 --> 00:02:23.650 It's the idea that resists change just because, 00:02:23.650 --> 00:02:25.189 you know, we've always done it this way. Oh, 00:02:25.189 --> 00:02:27.229 I see. Like the idea that opposites attract. 00:02:27.509 --> 00:02:30.650 Exactly. Or black cats are bad luck. Or, oh, 00:02:30.650 --> 00:02:32.349 here's a good one, the idea that you have to 00:02:32.349 --> 00:02:34.889 wait an hour after eating to swim, or you'll 00:02:34.889 --> 00:02:37.509 get a cramp and drown. Wait, is the swimming 00:02:37.509 --> 00:02:39.710 thing not true? I still wait 30 minutes. Just 00:02:39.710 --> 00:02:42.509 to be safe. And that's tenacity in action. There 00:02:42.509 --> 00:02:45.409 is basically zero data supporting the idea that 00:02:45.409 --> 00:02:47.969 recreational swimming after lunch causes fatal 00:02:47.969 --> 00:02:51.050 cramping. But we cling to it tenaciously because, 00:02:51.349 --> 00:02:53.990 well, that's just what we believe. It's comfortable. 00:02:54.250 --> 00:02:57.210 It requires zero mental effort to maintain. OK, 00:02:57.270 --> 00:02:59.330 that makes sense. So tenacity is just habit. 00:02:59.990 --> 00:03:03.530 Got it. Next on the list is intuition. The gut 00:03:03.530 --> 00:03:05.530 feeling. This is when you believe something because 00:03:05.530 --> 00:03:08.009 it just feels right. It's the hunch. Like when 00:03:08.009 --> 00:03:10.129 you meet someone and immediately think, I don't 00:03:10.129 --> 00:03:13.069 trust this guy. Right. Or the stock trader who 00:03:13.069 --> 00:03:15.810 says they just feel the market is turning or 00:03:15.810 --> 00:03:17.990 a parent who senses their kid is in trouble. 00:03:18.490 --> 00:03:21.229 And here's the thing. Intuition isn't always 00:03:21.229 --> 00:03:24.810 wrong. Sometimes it's actually like highly processed 00:03:24.810 --> 00:03:27.009 pattern recognition happening below your conscious 00:03:27.009 --> 00:03:29.330 radar. So what's the problem with it then from 00:03:29.330 --> 00:03:32.180 a scientific standpoint? The problem is it's 00:03:32.180 --> 00:03:35.300 purely subjective. I can't verify your gut feeling. 00:03:35.400 --> 00:03:38.319 I can't peer review your hunch. If your evidence 00:03:38.319 --> 00:03:40.960 is, I feel it in my bones, that's great for you. 00:03:41.120 --> 00:03:43.780 But it's not data. It's just a feeling. Science 00:03:43.780 --> 00:03:46.039 needs to be public and verifiable. You can't 00:03:46.039 --> 00:03:48.460 replicate a gut feeling. OK, next one is authority. 00:03:48.680 --> 00:03:51.680 This is a big one. Huge. This is accepting something 00:03:51.680 --> 00:03:54.639 as fact because a respected source said so. Because 00:03:54.639 --> 00:03:57.139 I said so. The classic parenting move. It's efficient, 00:03:57.240 --> 00:03:59.479 right? We trust the surgeon general or the FDA 00:03:59.479 --> 00:04:01.669 or professors. I don't own an electron microscope, 00:04:01.710 --> 00:04:03.490 so I have to trust the physicist who tells me 00:04:03.490 --> 00:04:06.389 what an atom looks like. We outsource the verification 00:04:06.389 --> 00:04:08.750 to someone with credentials. Yeah, but the text 00:04:08.750 --> 00:04:10.969 makes a specific point here about the difference 00:04:10.969 --> 00:04:13.909 between authority in science and authority in 00:04:13.909 --> 00:04:15.810 religion. I thought that was interesting. It 00:04:15.810 --> 00:04:18.149 is a fascinating distinction. In a religious 00:04:18.149 --> 00:04:20.750 context, authority is often the ultimate proof. 00:04:21.069 --> 00:04:23.269 If it's written in the Bible or the Quran or 00:04:23.269 --> 00:04:25.910 the Torah, that settles the argument for the 00:04:25.910 --> 00:04:28.589 believer. The text is the authority and the text 00:04:28.589 --> 00:04:32.040 is truth. Right. Faith often requires submission 00:04:32.040 --> 00:04:34.920 to authority. Exactly. But in science, authority 00:04:34.920 --> 00:04:37.540 is just a starting point, and often a shaky one. 00:04:37.980 --> 00:04:40.319 The source warns us that authority can be wrong. 00:04:40.800 --> 00:04:43.939 The FDA can make mistakes. Experts can be biased. 00:04:44.579 --> 00:04:47.199 Science demands we question authority, whereas 00:04:47.199 --> 00:04:50.279 faith often demands we submit to it. If Einstein 00:04:50.279 --> 00:04:53.339 says something, we don't just accept it, we test 00:04:53.339 --> 00:04:55.439 it. That's a really helpful distinction. Authority 00:04:55.439 --> 00:04:58.500 is a shortcut, not a destination. Okay, moving 00:04:58.500 --> 00:05:01.779 up the ladder of knowing. Rationalism. Now we're 00:05:01.779 --> 00:05:04.279 using our brains. Rationalism is using logic 00:05:04.279 --> 00:05:06.500 and reasoning to find truth. This is the Socrates 00:05:06.500 --> 00:05:08.939 thing, right? The syllogism. All men are mortal. 00:05:09.180 --> 00:05:11.759 Socrates is a man, therefore Socrates is mortal. 00:05:12.079 --> 00:05:14.480 It's airtight logic. If the first two parts, 00:05:14.639 --> 00:05:17.600 the premises, are true, the third part must be 00:05:17.600 --> 00:05:19.600 true. It is, provided the premises are true. 00:05:20.000 --> 00:05:22.660 But you can be logical and still be completely 00:05:22.660 --> 00:05:24.879 wrong about the real world if your starting point 00:05:24.879 --> 00:05:28.860 is off. Try this one. All birds can fly. Penguins 00:05:28.860 --> 00:05:31.459 are birds, therefore penguins can fly. The logic 00:05:31.459 --> 00:05:33.899 holds, but the conclusion is false because the 00:05:33.899 --> 00:05:37.699 first premise, all birds can fly, is just wrong. 00:05:38.139 --> 00:05:41.000 Exactly. Rationalism is garbage in, garbage out. 00:05:41.399 --> 00:05:43.420 You can sit in an armchair and logically deduce 00:05:43.420 --> 00:05:45.339 things all day, but unless you actually go outside 00:05:45.339 --> 00:05:47.500 and look at a penguin, you won't know the truth. 00:05:47.959 --> 00:05:50.300 And that brings us to the next one. Empiricism. 00:05:50.579 --> 00:05:53.180 Empiricism. That's the, show me the money of 00:05:53.180 --> 00:05:55.459 knowledge. I'll believe it when I see it. Pretty 00:05:55.459 --> 00:05:57.839 much. It's knowledge through direct sensory experience. 00:05:58.019 --> 00:06:00.199 I know fire is hot because I touched it. I know 00:06:00.199 --> 00:06:02.639 the sky is blue because I see it. So science 00:06:02.639 --> 00:06:06.740 is basically rationalism plus empiricism, logic 00:06:06.740 --> 00:06:09.500 plus observation. In a way. But it's organized. 00:06:09.620 --> 00:06:11.399 It's not just random looking and thinking. It's 00:06:11.399 --> 00:06:14.379 the scientific method. And the text uses a great 00:06:14.379 --> 00:06:18.019 analogy here. Science is a game. A game. Like 00:06:18.019 --> 00:06:21.040 Monopoly. Think more like a sport. football, 00:06:21.319 --> 00:06:24.939 or chess. A game is defined by its rules. If 00:06:24.939 --> 00:06:26.600 you pick up the ball in the middle of a soccer 00:06:26.600 --> 00:06:28.980 match and run into the goal using your hands, 00:06:29.220 --> 00:06:31.220 you haven't scored a point. You've just cheated. 00:06:31.579 --> 00:06:33.819 You're not playing soccer anymore. So the methodology 00:06:33.819 --> 00:06:36.459 is the rule book. Exactly. Science has strict 00:06:36.459 --> 00:06:40.560 rules. Randomization, control, peer review. If 00:06:40.560 --> 00:06:42.939 you don't follow the rules, the points you score, 00:06:43.120 --> 00:06:46.029 your results... They don't count. You can claim 00:06:46.029 --> 00:06:48.350 you found a cure for something, but if you didn't 00:06:48.350 --> 00:06:50.670 follow the method, nobody cares. Your claim is 00:06:50.670 --> 00:06:53.269 invalid. I love that. That makes so much sense. 00:06:53.550 --> 00:06:56.410 And the goal of this game, the sheet lists four 00:06:56.410 --> 00:06:59.129 goals of science. Describe, explain, predict, 00:06:59.389 --> 00:07:01.410 control. You need to know these in order. It's 00:07:01.410 --> 00:07:03.410 like a progression. Let's break those down with 00:07:03.410 --> 00:07:07.110 an example. Let's say... Procrastination. We're 00:07:07.110 --> 00:07:08.790 definitely experts on that right now. Perfect. 00:07:09.029 --> 00:07:11.009 First you have to describe. You have to define 00:07:11.009 --> 00:07:13.290 what you're even looking at. What is procrastination? 00:07:13.490 --> 00:07:15.550 Is it delaying a task? Is it doing something 00:07:15.550 --> 00:07:17.350 else instead? You're just painting the picture 00:07:17.350 --> 00:07:19.629 and defining your variables. Then you explain. 00:07:19.930 --> 00:07:23.089 This is the why. Why do students procrastinate? 00:07:23.449 --> 00:07:27.029 Is it fear of failure? Is it laziness? Is it 00:07:27.029 --> 00:07:29.410 like a dopamine issue where you're seeking a 00:07:29.410 --> 00:07:31.430 quick reward? You're looking for causes. Okay, 00:07:31.569 --> 00:07:34.779 then predict. If we know the cause, we can anticipate 00:07:34.779 --> 00:07:37.120 when it will happen again. If a student has high 00:07:37.120 --> 00:07:39.399 fear of failure, they will likely procrastinate 00:07:39.399 --> 00:07:42.500 on the final paper. We're making a testable prediction 00:07:42.500 --> 00:07:45.839 now. And finally, control. This is the ultimate 00:07:45.839 --> 00:07:48.620 power. If we can predict it, can we change it? 00:07:48.720 --> 00:07:51.000 Can we manipulate conditions to make procrastination 00:07:51.000 --> 00:07:54.060 go away? Can we design a workshop that lowers 00:07:54.060 --> 00:07:56.500 fear of failure and thus lowers procrastination? 00:07:56.540 --> 00:07:59.180 That's the end game. That's why we do this, to 00:07:59.180 --> 00:08:01.860 intervene. Okay, so within this game, we have 00:08:01.949 --> 00:08:04.889 Two main types of research on the list, basic 00:08:04.889 --> 00:08:07.230 and applied. Think of it as the difference between 00:08:07.230 --> 00:08:10.629 curiosity and problem solving. Basic research 00:08:10.629 --> 00:08:13.149 is answering questions just for the sake of understanding 00:08:13.149 --> 00:08:16.430 the mechanism. How does memory work? What is 00:08:16.430 --> 00:08:18.829 the nature of love? How does the retina process 00:08:18.829 --> 00:08:20.810 light? We aren't trying to fix anything yet. 00:08:20.810 --> 00:08:22.370 We just want to know how the machine functions. 00:08:22.529 --> 00:08:24.490 And applied research is the practical stuff. 00:08:25.089 --> 00:08:27.730 Exactly. How do we treat Alzheimer's? How do 00:08:27.730 --> 00:08:30.750 we reduce traffic accidents? It's taking that 00:08:30.750 --> 00:08:33.370 basic knowledge and applying it to real -world 00:08:33.370 --> 00:08:36.629 problems. They feed each other, you know? Basic 00:08:36.629 --> 00:08:38.909 research on memory might lead to an applied treatment 00:08:38.909 --> 00:08:41.669 for amnesia. And then there's the evil twin listed 00:08:41.669 --> 00:08:45.250 here. Pseudoscience. The imposter. Pseudoscience 00:08:45.250 --> 00:08:47.950 dresses up like science. It uses the jargon. 00:08:48.129 --> 00:08:50.909 It wears the lab coat. It has charts and graphs. 00:08:50.990 --> 00:08:53.429 Like those late -night commercials. Clinically 00:08:53.429 --> 00:08:56.690 proven to melt belly fat in 48 hours. Usually. 00:08:56.919 --> 00:08:59.299 The key giveaway, the thing to look for on the 00:08:59.299 --> 00:09:01.820 exam, is that pseudoscience only looks for confirmation. 00:09:02.360 --> 00:09:05.059 It ignores anything that proves it wrong. Real 00:09:05.059 --> 00:09:06.899 science is obsessed with trying to prove itself 00:09:06.899 --> 00:09:09.440 wrong. That's a huge difference. So pseudoscience 00:09:09.440 --> 00:09:11.779 starts with the conclusion, this magnet heals 00:09:11.779 --> 00:09:14.120 pain, and only looks for people who say, yes, 00:09:14.120 --> 00:09:16.039 it works, while ignoring the thousand people 00:09:16.039 --> 00:09:18.080 who say it didn't. Precisely. It's confirmation 00:09:18.080 --> 00:09:20.139 bias dressed up in a lab coat. That's a good 00:09:20.139 --> 00:09:23.120 segue into part two, generating ideas. Because 00:09:23.120 --> 00:09:25.960 to do real science, you need an idea. And that 00:09:25.960 --> 00:09:27.980 starts with the battle between hypothesis and 00:09:27.980 --> 00:09:30.679 theory. I feel like people mix these up constantly. 00:09:30.940 --> 00:09:32.860 All the time. In casual conversation, people 00:09:32.860 --> 00:09:34.580 say, oh, I have a theory about why my roommate 00:09:34.580 --> 00:09:36.899 is so cranky. In science, that's not a theory. 00:09:36.960 --> 00:09:38.679 That's a hypothesis. So what's the difference, 00:09:38.840 --> 00:09:42.519 technically? A hypothesis is a specific, testable 00:09:42.519 --> 00:09:45.159 prediction. It's a bet you place on a single 00:09:45.159 --> 00:09:48.059 outcome. If I give my roommate a cookie, their 00:09:48.059 --> 00:09:50.830 mood will improve. That's a hypothesis. It's 00:09:50.830 --> 00:09:52.929 narrow. It's specific. You can test it right 00:09:52.929 --> 00:09:56.570 now. And a theory is bigger. A theory is the 00:09:56.570 --> 00:09:58.950 big picture. It's a broad framework that explains 00:09:58.950 --> 00:10:01.570 a whole body of evidence. It's like the picture 00:10:01.570 --> 00:10:03.909 on the puzzle box after you've put hundreds of 00:10:03.909 --> 00:10:06.549 little hypothesis pieces together. Give me an 00:10:06.549 --> 00:10:08.950 example. Okay. The theory of evolution. It's 00:10:08.950 --> 00:10:11.710 not a guess. It's a massive framework that explains 00:10:11.710 --> 00:10:14.490 fossils, genetics, biology, and observation. 00:10:15.129 --> 00:10:17.669 It's supported by thousands and thousands of 00:10:17.669 --> 00:10:20.190 tested hypotheses. But the text makes an important 00:10:20.190 --> 00:10:23.470 point here. A theory is never proven 100 % right. 00:10:23.629 --> 00:10:26.129 Right. In science, we don't really use the word 00:10:26.129 --> 00:10:29.289 proven. A theory is just the best explanation 00:10:29.289 --> 00:10:32.610 we currently have. It can always be updated if 00:10:32.610 --> 00:10:35.389 new evidence comes along. Newton had a theory 00:10:35.389 --> 00:10:38.759 of gravity. It was great. Then Einstein came 00:10:38.759 --> 00:10:40.440 along and said, actually, there's more to it, 00:10:40.620 --> 00:10:42.659 and updated the theory with general relativity. 00:10:43.100 --> 00:10:44.940 That's the humility of science. It's always a 00:10:44.940 --> 00:10:47.379 work in progress. OK, that makes sense. Now, 00:10:47.639 --> 00:10:50.500 to build these theories, we use logic. The list 00:10:50.500 --> 00:10:53.960 has deduction and induction. Top down versus 00:10:53.960 --> 00:10:55.799 bottom up. I always get these backwards. OK, 00:10:56.059 --> 00:10:57.820 think of the shape. Deduction is like a triangle 00:10:57.820 --> 00:11:00.480 pointing down, top down. You start with a general 00:11:00.480 --> 00:11:02.909 theory at the top. And you deduce a specific 00:11:02.909 --> 00:11:04.710 prediction down at the bottom. Right, right. 00:11:04.710 --> 00:11:06.889 Like the front row example from the book. Exactly. 00:11:07.070 --> 00:11:09.190 General theory. Students who sit in the front 00:11:09.190 --> 00:11:12.429 row are smarter. So your deduction is, therefore, 00:11:12.669 --> 00:11:15.110 if I look at that specific student, Dave, in 00:11:15.110 --> 00:11:18.289 the front row, he will get an A. You go from 00:11:18.289 --> 00:11:20.769 the general rule to the specific case. OK, so 00:11:20.769 --> 00:11:23.470 general to specific. What about induction? Induction 00:11:23.470 --> 00:11:26.110 is the opposite. It's bottom up, like a triangle 00:11:26.110 --> 00:11:28.490 pointing up. You start with specific observations. 00:11:28.830 --> 00:11:31.519 You start with the data. The dirt. So you'd see? 00:11:31.740 --> 00:11:34.059 You'd see. Hey, Dave, in the front row got an 00:11:34.059 --> 00:11:36.179 A. And look, Sarah in the front row got an A. 00:11:36.200 --> 00:11:38.799 And Mike in the front row got an A. You gather 00:11:38.799 --> 00:11:41.220 all these specific alterations. And from them, 00:11:41.279 --> 00:11:44.700 you induce a general theory. Huh. Maybe all front 00:11:44.700 --> 00:11:47.220 row students are smart. You're building the theory 00:11:47.220 --> 00:11:49.559 from the ground up based on what you see. That 00:11:49.559 --> 00:11:52.120 feels like Sherlock Holmes. It is. Although, 00:11:52.379 --> 00:11:54.759 fun fact for the nerds listening, Sherlock Holmes 00:11:54.759 --> 00:11:57.100 usually claims he's using deduction. But he's 00:11:57.100 --> 00:11:59.519 actually using induction most of the time. He 00:11:59.519 --> 00:12:02.100 sees a muddy boot, that's a specific observation, 00:12:02.559 --> 00:12:04.240 and concludes the person came from the country, 00:12:04.440 --> 00:12:07.379 a general conclusion, that's bottom up, that's 00:12:07.379 --> 00:12:10.720 induction. Holmes was a liar. Okay, noted. Before 00:12:10.720 --> 00:12:12.820 we can even form these hypotheses, we have to 00:12:12.820 --> 00:12:15.320 do a literature review, which sounds so boring, 00:12:15.440 --> 00:12:17.860 but the text calls it Standing on the Shoulders 00:12:17.860 --> 00:12:20.440 of Giants. And it is. You have to know what's 00:12:20.440 --> 00:12:22.039 already been done. You look at peer -reviewed 00:12:22.039 --> 00:12:25.080 journals, psych -NFO, PubMed. You don't want 00:12:25.080 --> 00:12:26.960 to waste six months running an experiment that 00:12:26.960 --> 00:12:28.879 someone else already disproved ten years ago. 00:12:29.200 --> 00:12:31.399 It's about being part of a larger conversation. 00:12:31.539 --> 00:12:34.120 And there's a mnemonic here. The three C's of 00:12:34.120 --> 00:12:37.559 a review. Be comprehensive. Be critical. And 00:12:37.559 --> 00:12:40.460 be clever. Don't just read the abstract and call 00:12:40.460 --> 00:12:43.120 it a day. Tear the study apart. Ask questions. 00:12:43.500 --> 00:12:46.059 Why did they use that sample? Was there logic 00:12:46.059 --> 00:12:49.169 sound? Is there a better way to test this? But 00:12:49.169 --> 00:12:52.330 there's a danger lurking in those journals. Publication 00:12:52.330 --> 00:12:55.940 bias. The file drawer problem. This is a really 00:12:55.940 --> 00:12:58.639 serious issue in science. Journals want to publish 00:12:58.639 --> 00:13:02.419 exciting, positive results. We found a cure or 00:13:02.419 --> 00:13:05.600 we found a connection between X and Y. And nobody 00:13:05.600 --> 00:13:07.919 wants to publish, we found nothing. Exactly. 00:13:07.960 --> 00:13:11.059 If you run a perfectly good study and find nothing, 00:13:11.480 --> 00:13:14.120 hey, eating kale has zero effect on math scores. 00:13:14.480 --> 00:13:16.860 That study probably won't get published. It gets 00:13:16.860 --> 00:13:18.840 stuck in the researcher's file drawer. So we 00:13:18.840 --> 00:13:20.740 end up with a skewed view of reality because 00:13:20.740 --> 00:13:24.250 we only see the hits and never the misses. Precisely. 00:13:24.710 --> 00:13:27.450 It makes effects seem stronger or more common 00:13:27.450 --> 00:13:29.090 than they really are because all the failures 00:13:29.090 --> 00:13:31.210 are invisible. It's a huge problem. Okay, let's 00:13:31.210 --> 00:13:33.750 get serious for a second. Part three, the moral 00:13:33.750 --> 00:13:38.950 compass. Research ethics. Because science hasn't 00:13:38.950 --> 00:13:41.690 always been the good guy. No, not at all. History 00:13:41.690 --> 00:13:44.190 is full of the dark side of science. When we 00:13:44.190 --> 00:13:47.610 prioritize knowledge over humanity, we get atrocities. 00:13:47.809 --> 00:13:50.529 The text starts with the Nazi medical experiments. 00:13:50.870 --> 00:13:53.190 It's horrific stuff, freezing people to see how 00:13:53.190 --> 00:13:55.309 long they'd survive, infecting them with diseases 00:13:55.309 --> 00:13:57.610 just to watch them die. This led directly to 00:13:57.610 --> 00:14:00.750 the Nuremberg Code in 1947. The absolute number 00:14:00.750 --> 00:14:03.210 one rule that came out of that, voluntary consent, 00:14:03.570 --> 00:14:05.850 you cannot force people to be guinea pigs, period. 00:14:06.090 --> 00:14:08.629 And it wasn't just the Nazis. We have the Tuskegee 00:14:08.629 --> 00:14:11.139 syphilis study right here in the U .S. This one 00:14:11.139 --> 00:14:13.240 is just heartbreaking. Researchers from the U 00:14:13.240 --> 00:14:15.120 .S. Public Health Service studied the progression 00:14:15.120 --> 00:14:17.700 of syphilis in hundreds of poor African -American 00:14:17.700 --> 00:14:21.419 men for 40 years. 40! And they told the men they 00:14:21.419 --> 00:14:23.159 were being treated. They told them they were 00:14:23.159 --> 00:14:25.980 being treated for bad blood, which was a vague 00:14:25.980 --> 00:14:28.480 local term. But they actually withheld the cure 00:14:28.480 --> 00:14:30.799 penicillin even after it became the standard 00:14:30.799 --> 00:14:33.179 treatment, just so they could watch the disease 00:14:33.179 --> 00:14:36.620 destroy these men's bodies. That is, it's evil. 00:14:36.909 --> 00:14:39.509 There's no other word for it. It is. It violated 00:14:39.509 --> 00:14:41.909 every ethical standard we hold now. That study 00:14:41.909 --> 00:14:44.590 is largely why we have the Belmont report. This 00:14:44.590 --> 00:14:47.210 is guaranteed to be on the exam. The big three 00:14:47.210 --> 00:14:49.350 principles. Yeah. We need to know these cold. 00:14:49.529 --> 00:14:52.970 Okay. Number one, respect for persons. This means 00:14:52.970 --> 00:14:56.200 autonomy. People must be capable of making their 00:14:56.200 --> 00:14:58.940 own informed decisions. This is where informed 00:14:58.940 --> 00:15:01.360 consent comes in. You have to tell them the risks, 00:15:01.659 --> 00:15:03.240 the benefits, what they'll be doing, and that 00:15:03.240 --> 00:15:06.159 they can quit at any time. And if you can't consent, 00:15:06.419 --> 00:15:09.379 like children, or prisoners, or people with cognitive 00:15:09.379 --> 00:15:11.720 impairments? They're considered vulnerable populations. 00:15:11.899 --> 00:15:14.500 They need extra protection. For a child, you 00:15:14.500 --> 00:15:16.460 need permission from a guardian, and you also 00:15:16.460 --> 00:15:18.879 need assent from the child, meaning they agree 00:15:18.879 --> 00:15:21.039 on their own level, even if they can't legally 00:15:21.039 --> 00:15:25.399 consent. OK, number two. Beneficence. Do no harm. 00:15:26.120 --> 00:15:29.799 It's the Hippocratic oath for researchers. Maximize 00:15:29.799 --> 00:15:32.740 benefits, minimize risks. You always have to 00:15:32.740 --> 00:15:35.519 do a risk -benefit analysis. Is the knowledge 00:15:35.519 --> 00:15:38.279 we might gain worth the potential stress or harm 00:15:38.279 --> 00:15:41.019 we might put the participant through? And three, 00:15:42.039 --> 00:15:45.120 justice. Fairness. This goes right back to Tuskegee. 00:15:45.259 --> 00:15:48.080 You can't just test on the poor or the vulnerable 00:15:48.080 --> 00:15:50.460 to find cures that will only benefit the rich. 00:15:50.960 --> 00:15:53.039 The burdens and benefits of research must be 00:15:53.039 --> 00:15:56.259 distributed fairly across society. Okay, let's 00:15:56.259 --> 00:15:58.840 unpack the specific terms on the sheet. Informant 00:15:58.840 --> 00:16:00.980 consent we got. What about deception? Can you 00:16:00.980 --> 00:16:04.039 lie to participants? Surprisingly, yes, but it's 00:16:04.039 --> 00:16:06.559 a very delicate line. Sometimes, if people know 00:16:06.559 --> 00:16:08.440 exactly what you're testing, they act differently. 00:16:08.720 --> 00:16:11.299 If I tell you... I'm testing how honest you are. 00:16:11.720 --> 00:16:13.240 You're suddenly going to be the most honest person 00:16:13.240 --> 00:16:14.899 in the world, so I might have to deceive you 00:16:14.899 --> 00:16:16.639 about the study's true purpose. But there are 00:16:16.639 --> 00:16:18.600 rules for that, right? Strict rules. You can 00:16:18.600 --> 00:16:20.940 only use deception if it's absolutely necessary 00:16:20.940 --> 00:16:22.700 for the study to work, and you must follow it 00:16:22.700 --> 00:16:24.860 with a debriefing. Debriefing. That's the coming 00:16:24.860 --> 00:16:27.139 clean moment. Right. After the study is over, 00:16:27.259 --> 00:16:29.580 you sit them down and say, OK, we're sorry we 00:16:29.580 --> 00:16:31.879 told you the building was on fire. We were actually 00:16:31.879 --> 00:16:34.399 testing your reaction time under stress. Here's 00:16:34.399 --> 00:16:36.480 why we had to do it that way. Here's what we 00:16:36.480 --> 00:16:38.980 hope to learn. And are you OK? You have to restore 00:16:38.980 --> 00:16:40.679 their trust and make sure they leave in the same 00:16:40.679 --> 00:16:44.179 state they came in. Then there's anonymity versus 00:16:44.179 --> 00:16:47.039 confidentiality. These are not the same thing. 00:16:47.379 --> 00:16:49.259 Crucial distinction for the exam. This trips 00:16:49.259 --> 00:16:52.159 people up. Anonymity means even the researcher 00:16:52.159 --> 00:16:54.440 doesn't know who the participant is. There are 00:16:54.440 --> 00:16:57.139 no names, no identifying information linked to 00:16:57.139 --> 00:16:59.759 the data. It's a double -blind data set, effectively. 00:17:00.220 --> 00:17:02.799 Confidentiality. Confidentiality means I know 00:17:02.799 --> 00:17:04.960 who you are. I have your name on a consent form, 00:17:05.039 --> 00:17:07.759 but I promise to keep your data secret and secure. 00:17:08.140 --> 00:17:09.980 Your name won't be in the final publication. 00:17:10.099 --> 00:17:13.660 It's locked away. Got it. Now, we can't talk 00:17:13.660 --> 00:17:16.200 ethics without talking about the bad boy studies, 00:17:16.660 --> 00:17:18.339 the ones that probably couldn't be done today. 00:17:18.900 --> 00:17:22.589 First up... Milgram. The obedience study. Stanley 00:17:22.589 --> 00:17:25.250 Milgram was haunted by the Holocaust. He wanted 00:17:25.250 --> 00:17:27.910 to know, how could ordinary people commit such 00:17:27.910 --> 00:17:30.650 atrocities? So he set up a fake shock machine. 00:17:31.589 --> 00:17:34.490 The participant, the teacher, thought they were 00:17:34.490 --> 00:17:36.910 giving increasingly powerful electric shocks 00:17:36.910 --> 00:17:39.289 to a learner in the other room every time they 00:17:39.289 --> 00:17:41.250 got a question wrong. And the shocks kept getting 00:17:41.250 --> 00:17:43.569 stronger, all the way up to a switch labeled, 00:17:44.190 --> 00:17:47.740 danger, severe shock, and then just... X, X, 00:17:47.740 --> 00:17:50.460 X. And the learner, who is an actor, would scream, 00:17:50.700 --> 00:17:52.460 complain about his heart condition, and eventually 00:17:52.460 --> 00:17:55.019 fall silent. But the experimenter in the white 00:17:55.019 --> 00:17:57.339 lab coat would just calmly say, the experiment 00:17:57.339 --> 00:17:59.539 requires that you continue. And the terrifying 00:17:59.539 --> 00:18:02.180 results. 65 % of people went all the way to the 00:18:02.180 --> 00:18:05.200 maximum voltage, ordinary people. It showed that 00:18:05.200 --> 00:18:07.559 the power of authority can often override our 00:18:07.559 --> 00:18:09.980 own personal morality. It's a profound finding, 00:18:10.299 --> 00:18:12.569 but the stress it caused the participants. yelling 00:18:12.569 --> 00:18:15.069 at the experiment or sweating, shaking. It was 00:18:15.069 --> 00:18:17.269 highly unethical by modern standards. Absolutely. 00:18:17.670 --> 00:18:20.730 Next, Zimbardo, the Stanford prison experiment. 00:18:21.230 --> 00:18:23.990 Oh, man. They assigned psychologically healthy 00:18:23.990 --> 00:18:26.849 college students to be either guards or prisoners 00:18:26.849 --> 00:18:29.970 in a mock prison in the basement of the psychology 00:18:29.970 --> 00:18:32.109 building. It was supposed to last two weeks. 00:18:32.410 --> 00:18:35.210 It lasted six days. Because the guards went full 00:18:35.210 --> 00:18:38.130 dictator. They became sadistic. They were verbally 00:18:38.130 --> 00:18:41.250 abusive. They made prisoners do humiliating things. 00:18:41.690 --> 00:18:44.529 The prisoners, in turn, had emotional breakdowns. 00:18:44.670 --> 00:18:47.349 Zimbardo had to shut it down early. The study 00:18:47.349 --> 00:18:49.769 showed how powerful social roles, the uniform 00:18:49.769 --> 00:18:52.569 you wear, can create evil behavior. It wasn't 00:18:52.569 --> 00:18:54.829 that they were bad apples. It was a bad barrel. 00:18:55.210 --> 00:18:57.490 And finally, the last one on this list, Robber's 00:18:57.490 --> 00:19:00.349 Cave. This one is actually a bit more hopeful 00:19:00.349 --> 00:19:02.910 in the end. Muzaffar Sharif took two groups of 00:19:02.910 --> 00:19:05.410 boys at a summer camp. He named them the Eagles 00:19:05.410 --> 00:19:08.269 and the Rattlers. He first built group cohesion. 00:19:08.569 --> 00:19:10.269 Then he put them in competition against each 00:19:10.269 --> 00:19:12.289 other. And they grew to hate each other, like 00:19:12.289 --> 00:19:14.690 true prejudice. They were raiding cabins, burning 00:19:14.690 --> 00:19:17.390 flags. Lord of the fly style. Basically. But 00:19:17.390 --> 00:19:19.990 then he investigated how to fix it. He found 00:19:19.990 --> 00:19:21.829 that just bringing them together for a movie 00:19:21.829 --> 00:19:23.890 night didn't work. It just gave them another 00:19:23.890 --> 00:19:27.000 chance to fight. What worked was creating super 00:19:27.000 --> 00:19:29.839 -ordinate goals. What's that? Shared goals where 00:19:29.839 --> 00:19:32.700 they had to work together to succeed. The camp's 00:19:32.700 --> 00:19:35.000 water supply broke and all the boys had to work 00:19:35.000 --> 00:19:38.180 together to fix it. A truck got stuck and they 00:19:38.180 --> 00:19:41.119 all had to pull it out. So shared struggle creates 00:19:41.119 --> 00:19:43.539 unity. I like that. That's a powerful lesson. 00:19:44.140 --> 00:19:46.279 Okay, my brain needs a break from the heavy stuff. 00:19:46.759 --> 00:19:50.369 Part four, measuring the invisible. This is the 00:19:50.369 --> 00:19:53.190 core challenge of psychology, right? In physics, 00:19:53.250 --> 00:19:56.109 you can measure mass or velocity. In psych, we 00:19:56.109 --> 00:19:57.809 are measuring feelings. We're trying to measure 00:19:57.809 --> 00:20:00.230 things you can't see or touch. We are measuring 00:20:00.230 --> 00:20:04.390 constructs. Love, anxiety, hunger, intelligence. 00:20:04.609 --> 00:20:07.180 Yep. So the big question is, how do you measure 00:20:07.180 --> 00:20:09.619 hunger? You can't put a ruler next to it. You 00:20:09.619 --> 00:20:11.480 need a way to make it observable. You need an 00:20:11.480 --> 00:20:13.339 operational definition. Yeah. This is the recipe 00:20:13.339 --> 00:20:15.619 for measuring the construct. It's how you operationalize 00:20:15.619 --> 00:20:18.099 it. Exactly. A bad conceptual definition of hunger 00:20:18.099 --> 00:20:20.779 is feeling empty. That's subjective. I don't 00:20:20.779 --> 00:20:23.140 know what your empty feels like. A good operational 00:20:23.140 --> 00:20:26.579 definition is hours since last meal, or grams 00:20:26.579 --> 00:20:29.279 of food consumed in the last 24 hours, or even 00:20:29.279 --> 00:20:31.859 self -reported hunger on a scale of 1 to 10. 00:20:32.099 --> 00:20:35.089 It's concrete and measurable. This matters a 00:20:35.089 --> 00:20:37.609 lot. The prompt mentioned a moral link here. 00:20:37.970 --> 00:20:40.109 Think about defining morality. That's a great 00:20:40.109 --> 00:20:43.750 example. If I operationally define morality as 00:20:43.750 --> 00:20:46.009 frequency of church attendance, I'm going to 00:20:46.009 --> 00:20:48.869 get a very specific and probably very flawed 00:20:48.869 --> 00:20:51.549 set of data. I'm going to completely miss the 00:20:51.549 --> 00:20:53.390 atheists who volunteers at the soup kitchen every 00:20:53.390 --> 00:20:56.089 single day. So how we choose to define our variables 00:20:56.089 --> 00:20:59.210 fundamentally shapes the truth we find. Okay, 00:20:59.890 --> 00:21:02.029 we have our definition. Now we need a scale to 00:21:02.029 --> 00:21:07.029 measure it on. The acronym is N -O -I -R. N -O 00:21:07.029 --> 00:21:10.009 -I -R. Nominal ordinal interval ratio. You have 00:21:10.009 --> 00:21:11.529 to know these. Let's break them down. Nominal 00:21:11.529 --> 00:21:14.910 is N for name. It's just categories. Junger, 00:21:15.109 --> 00:21:18.210 eye color, political party, Ferrari versus Ford. 00:21:18.529 --> 00:21:21.750 There is no order. blue isn't better or more 00:21:21.750 --> 00:21:25.150 than brown. Got it. Ordinal is O for order. Right. 00:21:25.349 --> 00:21:27.269 Rankings, first place, second place, third place, 00:21:27.450 --> 00:21:29.329 small, medium, large. The crucial point here, 00:21:29.349 --> 00:21:31.009 and this is key, is that the gap between the 00:21:31.009 --> 00:21:33.349 ranks isn't necessarily equal. The difference 00:21:33.349 --> 00:21:35.289 in speed between the gold medalist and the silver 00:21:35.289 --> 00:21:38.309 medalist might be 0 .01 seconds. The difference 00:21:38.309 --> 00:21:40.210 between silver and bronze might be 10 seconds. 00:21:40.690 --> 00:21:42.809 We know the order, but not the distance between 00:21:42.809 --> 00:21:45.480 them. OK, that makes sense. Then interval. Interval 00:21:45.480 --> 00:21:47.920 scales have equal gaps. Think temperature in 00:21:47.920 --> 00:21:50.400 Celsius or Fahrenheit. The difference between 00:21:50.400 --> 00:21:52.880 10 degrees and 20 degrees is the same as the 00:21:52.880 --> 00:21:55.059 difference between 20 and 30 degrees. It's 10 00:21:55.059 --> 00:21:58.039 degrees. So the intervals are equal. But there's 00:21:58.039 --> 00:22:00.319 no true zero. Right. Zero degrees doesn't mean 00:22:00.319 --> 00:22:02.579 an absence of temperature. It's just another 00:22:02.579 --> 00:22:04.539 point on the scale. You can even have negative 00:22:04.539 --> 00:22:06.970 temperature. Which brings us to the last one 00:22:06.970 --> 00:22:10.329 ratio ratio scales are like interval scales But 00:22:10.329 --> 00:22:13.230 they have a true zero a meaningful zero point 00:22:13.230 --> 00:22:16.089 height weight time number of questions answered 00:22:16.089 --> 00:22:19.309 correctly So zero pounds means no weight zero 00:22:19.309 --> 00:22:21.910 seconds means no time And because there is a 00:22:21.910 --> 00:22:24.950 true zero, you can make ratios. You can say 100 00:22:24.950 --> 00:22:27.630 pounds is twice as heavy as 50 pounds. You can't 00:22:27.630 --> 00:22:29.829 say 100 degrees Fahrenheit is twice as hot as 00:22:29.829 --> 00:22:32.089 50 degrees Fahrenheit. That math doesn't work. 00:22:32.250 --> 00:22:34.609 That is the trickiest distinction, I think. Interval 00:22:34.609 --> 00:22:37.390 versus ratio. The key question is, does zero 00:22:37.390 --> 00:22:40.849 mean none of the thing? If yes, it's ratio. You 00:22:40.849 --> 00:22:43.990 nailed it. OK. Next big concept pair, reliability 00:22:43.990 --> 00:22:46.390 versus validity. The bathroom scale analogy. 00:22:46.470 --> 00:22:49.079 This one is clutch. It's the classic for a reason. 00:22:49.420 --> 00:22:52.519 Reliability is consistency. If I step on my bathroom 00:22:52.519 --> 00:22:55.420 scale five times in a row and it says 150, 150, 00:22:55.559 --> 00:22:58.859 150, 150, 150, that scale is highly reliable. 00:22:59.220 --> 00:23:01.740 It gives me the same result every time. OK, so 00:23:01.740 --> 00:23:05.359 what's validity? Validity is accuracy. Does the 00:23:05.359 --> 00:23:08.740 scale reflect reality? If I actually weigh 150 00:23:08.740 --> 00:23:12.720 pounds and the scale says 150, it is both reliable 00:23:12.720 --> 00:23:15.980 A and D valid. But what if I actually weigh 150 00:23:15.980 --> 00:23:19.519 pounds and the scale consistently says 170, 170, 00:23:19.740 --> 00:23:21.660 170? Well, that's reliable, but it's not valid. 00:23:21.859 --> 00:23:24.539 It's consistently wrong. Exactly. A broken clock 00:23:24.539 --> 00:23:26.920 is perfectly reliable. It shows the exact same 00:23:26.920 --> 00:23:29.099 time twice a day, but it's only valid for two 00:23:29.099 --> 00:23:31.380 minutes out of the entire day. For science, you 00:23:31.380 --> 00:23:33.359 need both. You need your measure to be consistent 00:23:33.359 --> 00:23:35.440 and accurate. The text mentions different types. 00:23:35.690 --> 00:23:38.710 test -retest reliability. Are scores stable over 00:23:38.710 --> 00:23:40.930 time? If I give you an IQ test today and again 00:23:40.930 --> 00:23:42.470 in a month, the scores should be pretty similar. 00:23:42.769 --> 00:23:44.690 And inter -rater reliability. You two judges 00:23:44.690 --> 00:23:46.549 see the same thing. Think of Olympic diving. 00:23:46.710 --> 00:23:49.049 You want the judges' scores to be highly correlated. 00:23:49.150 --> 00:23:51.769 If one gives a 9 .5 and another gives a 2 .0, 00:23:51.789 --> 00:23:54.769 you have low inter -rater reliability. And for 00:23:54.769 --> 00:23:58.490 validity, there's construct validity, which is 00:23:58.490 --> 00:24:01.029 the big one. Are we actually measuring what we 00:24:01.029 --> 00:24:04.130 think we are? Is our love scale really measuring 00:24:04.130 --> 00:24:06.589 love, or is it just measuring, like, attachment? 00:24:06.789 --> 00:24:09.349 And then there's face validity, which is just, 00:24:09.630 --> 00:24:11.769 does it look like a good test on the surface? 00:24:11.849 --> 00:24:13.910 Which is the least scientific type, right? Totally. 00:24:14.490 --> 00:24:17.970 Buzzfeed quizzes have high face validity, which 00:24:17.970 --> 00:24:20.390 Disney princess are you? Looks like it's measuring 00:24:20.390 --> 00:24:22.549 that, but they have absolutely zero construct 00:24:22.549 --> 00:24:25.410 validity. Okay, part five. We're getting there. 00:24:25.990 --> 00:24:27.869 Sampling and research design. The who and the 00:24:27.869 --> 00:24:30.420 how. Let's start with the who. Population versus 00:24:30.420 --> 00:24:32.920 sample. Population is everyone you want your 00:24:32.920 --> 00:24:35.059 results to apply to. The entire group you're 00:24:35.059 --> 00:24:37.119 interested in. All college students in the United 00:24:37.119 --> 00:24:39.779 States. All people with depression. Which is 00:24:39.779 --> 00:24:41.960 usually impossible to study. So you take a sample. 00:24:42.279 --> 00:24:44.339 The sample is the small subset of the population 00:24:44.339 --> 00:24:47.119 that you actually study. The 20 kids in my intracite 00:24:47.119 --> 00:24:50.119 class. The 100 people who responded to my flyer. 00:24:50.240 --> 00:24:52.180 And ideally you want a representative sample. 00:24:52.380 --> 00:24:54.539 One that actually looks like the population in 00:24:54.539 --> 00:24:57.160 terms of, you know, age, gender, ethnicity, etc. 00:24:57.359 --> 00:24:59.619 But usually what we get in psychology is what's 00:24:59.619 --> 00:25:01.839 called a weird sample. This is a great acronym. 00:25:02.019 --> 00:25:06.180 I remember this Western. Western, educated, industrialized, 00:25:06.440 --> 00:25:09.619 rich and democratic. Most psychology research 00:25:09.619 --> 00:25:11.980 is done on American undergrads who need course 00:25:11.980 --> 00:25:14.319 credit. So the question is, does a finding from 00:25:14.319 --> 00:25:17.180 a study of 19 year old Stanford students really 00:25:17.180 --> 00:25:20.240 apply to a 60 year old farmer in Kenya? Maybe. 00:25:21.000 --> 00:25:23.400 Maybe not. That's a sampling bias, and it's a 00:25:23.400 --> 00:25:25.700 big limitation on the generalizability of our 00:25:25.700 --> 00:25:28.859 findings. Okay, now for the how. The designs. 00:25:29.319 --> 00:25:32.579 Non -experimental versus experimental. Non -experimental 00:25:32.579 --> 00:25:36.380 includes naturalistic observation. Just watching 00:25:36.380 --> 00:25:38.819 people in the wild. Yep. Like Jane Goodall with 00:25:38.819 --> 00:25:41.579 the chimps. It has high external validity because 00:25:41.579 --> 00:25:43.960 it's real life, but you have very low control. 00:25:44.140 --> 00:25:45.940 You can't stop it from raining, and you can't 00:25:45.940 --> 00:25:48.000 ask the chimps to repeat that cool thing they 00:25:48.000 --> 00:25:50.259 just did. And it has some ethical issues, right? 00:25:50.349 --> 00:25:52.750 Privacy. For sure. Then you have surveys and 00:25:52.750 --> 00:25:54.609 correlational research. This is where you measure 00:25:54.609 --> 00:25:56.329 two or more variables and see how they relate. 00:25:56.490 --> 00:25:58.390 And this is where we get the golden rule of stats, 00:25:58.829 --> 00:26:01.650 tattooed on every researcher's brain. Correlation 00:26:01.650 --> 00:26:03.990 does not equal causation. Say it again for the 00:26:03.990 --> 00:26:06.349 people in the back. Just because two things move 00:26:06.349 --> 00:26:09.250 together doesn't mean one causes the other. The 00:26:09.250 --> 00:26:12.809 classic example. Ice cream sales are positively 00:26:12.809 --> 00:26:15.349 correlated with shark attacks. So does eating 00:26:15.349 --> 00:26:17.650 ice cream cause you to be attacked by a shark? 00:26:17.829 --> 00:26:20.789 No. There's a third variable causing both. Heat. 00:26:21.269 --> 00:26:23.609 In the summer, it's hot, so more people go swimming 00:26:23.609 --> 00:26:26.269 and more people eat ice cream. OK, so to prove 00:26:26.269 --> 00:26:28.730 cause and effect, you need an experiment. The 00:26:28.730 --> 00:26:30.930 gold standard. The gold standard. And you need 00:26:30.930 --> 00:26:33.029 three pillars. The professor will definitely 00:26:33.029 --> 00:26:35.450 ask about these. Number one is manipulation. 00:26:36.089 --> 00:26:38.569 You have to actively change the independent variable. 00:26:38.809 --> 00:26:41.210 Number two is control. You need a control group 00:26:41.210 --> 00:26:42.930 that doesn't get the manipulation, and you have 00:26:42.930 --> 00:26:44.839 to keep everything else constant. And number 00:26:44.839 --> 00:26:47.200 three, the most important one is randomization. 00:26:47.480 --> 00:26:49.799 You must use random assignment to put people 00:26:49.799 --> 00:26:52.420 into your experimental and control groups. This 00:26:52.420 --> 00:26:54.559 helps ensure the groups are equal to begin with. 00:26:54.720 --> 00:26:56.559 And if you're missing one of those, especially 00:26:56.559 --> 00:26:58.380 random assignment, then you have what's called 00:26:58.380 --> 00:27:01.680 a quasi experiment. It looks like an experiment, 00:27:01.680 --> 00:27:04.660 but it's not quite there. For example, comparing 00:27:04.660 --> 00:27:07.000 men and women, you can't randomly assign someone 00:27:07.000 --> 00:27:09.920 to be man or woman. So any differences you find 00:27:09.920 --> 00:27:12.220 might be due to gender or they might be due to 00:27:12.220 --> 00:27:14.099 something else you couldn't control for. Got 00:27:14.099 --> 00:27:16.880 it. Even with a true experiment, there are threats. 00:27:17.619 --> 00:27:20.220 Internal validity versus external validity. This 00:27:20.220 --> 00:27:22.559 is a classic trade -off. It's like a seesaw. 00:27:22.799 --> 00:27:25.740 Internal validity is your confidence that the 00:27:25.740 --> 00:27:28.019 independent variable really caused the change 00:27:28.019 --> 00:27:30.779 in the dependent variable. It's high in a controlled 00:27:30.779 --> 00:27:32.819 lab setting where you can eliminate all other 00:27:32.819 --> 00:27:34.980 explanations. But the more controlled your lab 00:27:34.980 --> 00:27:37.859 is, the less it looks like the real world. Exactly. 00:27:38.119 --> 00:27:40.539 Which lowers your external validity, the degree 00:27:40.539 --> 00:27:42.940 to which your findings apply or generalize to 00:27:42.940 --> 00:27:45.660 the real world. A field study has high external 00:27:45.660 --> 00:27:48.299 validity but low internal validity because of 00:27:48.299 --> 00:27:51.180 all the chaos of real life. And there are specific 00:27:51.180 --> 00:27:53.890 threats to internal validity. The list has history. 00:27:54.210 --> 00:27:56.430 An outside event that happens during your study 00:27:56.430 --> 00:27:59.069 and affects the results. You're doing a study 00:27:59.069 --> 00:28:02.390 on anxiety and a sudden stock market crash happens 00:28:02.390 --> 00:28:05.420 in the middle of it. That crash is a history 00:28:05.420 --> 00:28:08.460 threat. Maturation. People just grow or change 00:28:08.460 --> 00:28:10.880 naturally over time. If you're studying a reading 00:28:10.880 --> 00:28:13.119 program for first graders over a year, they're 00:28:13.119 --> 00:28:14.579 going to get better at reading anyway just because 00:28:14.579 --> 00:28:16.700 they're getting older. That's maturation. People 00:28:16.700 --> 00:28:18.980 dropping out of your study. And it's usually 00:28:18.980 --> 00:28:21.519 not random. The people who drop out might be 00:28:21.519 --> 00:28:23.839 different from those who stay, which biases your 00:28:23.839 --> 00:28:26.819 results. And regression toward the mean. This 00:28:26.819 --> 00:28:29.569 one's tricky. It just means that extreme scores 00:28:29.569 --> 00:28:32.829 tend to get less extreme over time. If you score 00:28:32.829 --> 00:28:34.950 super high on a test, you'll probably score a 00:28:34.950 --> 00:28:37.390 little lower the next time. If you score super 00:28:37.390 --> 00:28:39.809 low, you'll probably score a little higher. Extreme 00:28:39.809 --> 00:28:41.809 scores have nowhere to go but back toward the 00:28:41.809 --> 00:28:44.910 average. Okay, last section. My brain is almost 00:28:44.910 --> 00:28:48.750 at capacity. Part 6. Making sense of the data. 00:28:49.369 --> 00:28:52.170 Statistics. Don't panic. We just need the basics. 00:28:52.630 --> 00:28:55.789 First, descriptive statistics. These just describe 00:28:55.789 --> 00:28:58.160 or summarize your data. This is your measures 00:28:58.160 --> 00:29:02.099 of central tendency. Mean, median, and mode. 00:29:02.660 --> 00:29:06.339 Right. Mean is the mathematical average. Mode 00:29:06.339 --> 00:29:09.099 is the most frequent score. And median is the 00:29:09.099 --> 00:29:11.359 one right in the middle. Why do we need the median 00:29:11.359 --> 00:29:13.859 if we have the mean? It's good for skewed data, 00:29:14.160 --> 00:29:17.000 like income. If Bill Gates walks into a bar with 00:29:17.000 --> 00:29:20.519 10 regular people, the mean income skyrockets 00:29:20.519 --> 00:29:23.259 to millions of dollars. But that doesn't really 00:29:23.259 --> 00:29:25.380 describe the bar. The median income, the guy 00:29:25.380 --> 00:29:27.279 in the middle, stays pretty much the same. It 00:29:27.279 --> 00:29:30.019 gives a better picture. OK. And then there's 00:29:30.019 --> 00:29:32.559 variability, how spread out the scores are. Standard 00:29:32.559 --> 00:29:35.200 deviation. Right. Now for the hard part, inferential 00:29:35.200 --> 00:29:37.720 statistics. This is where we make inferences 00:29:37.720 --> 00:29:40.200 from our sample back to the population. This 00:29:40.200 --> 00:29:42.140 is for decision making. And it all starts with 00:29:42.140 --> 00:29:45.059 the null hypothesis. The assumption of no effect. 00:29:45.279 --> 00:29:47.599 It's the boring default state of the world. This 00:29:47.599 --> 00:29:50.059 new drug does nothing. There is no difference 00:29:50.059 --> 00:29:52.539 between these two groups. In the legal system, 00:29:52.619 --> 00:29:55.539 it's innocent until proven guilty. In science, 00:29:55.579 --> 00:29:58.339 it's no effect until proven otherwise. And the 00:29:58.339 --> 00:30:00.440 whole goal of the experiment is to find enough 00:30:00.440 --> 00:30:03.680 evidence to reject the null hypothesis. Exactly. 00:30:03.940 --> 00:30:06.900 And if we can, we say the result is statistically 00:30:06.900 --> 00:30:09.559 significant. But we can make mistakes. The sheet 00:30:09.559 --> 00:30:13.140 lists type I and type II errors. Right. A type 00:30:13.140 --> 00:30:16.380 I error is a false positive. It's when you reject 00:30:16.380 --> 00:30:18.599 the null hypothesis when you shouldn't have. 00:30:18.940 --> 00:30:20.819 You conclude the drug works, but it actually 00:30:20.819 --> 00:30:23.079 doesn't. You found an effect that isn't really 00:30:23.079 --> 00:30:26.099 there. So like convicting an innocent person. 00:30:26.269 --> 00:30:29.769 Perfect analogy and a type 2 error is a false 00:30:29.769 --> 00:30:32.269 negative It's when you fail to reject the null 00:30:32.269 --> 00:30:34.589 when you should have the drug actually works, 00:30:34.589 --> 00:30:37.390 but your study missed it You fail to find a real 00:30:37.390 --> 00:30:40.430 effect letting a guilty person go free exactly 00:30:40.430 --> 00:30:45.380 Wow We did it. That is the entire list. My brain 00:30:45.380 --> 00:30:47.460 is full, but I feel like I actually understand 00:30:47.460 --> 00:30:49.460 the logic now. It's not just a list of words. 00:30:49.599 --> 00:30:51.420 It's a toolkit. That's what's important. These 00:30:51.420 --> 00:30:53.359 aren't just terms for the exam tomorrow. They 00:30:53.359 --> 00:30:55.079 are tools for critical thinking for the rest 00:30:55.079 --> 00:30:57.779 of your life. Right. Because once you leave this 00:30:57.779 --> 00:31:00.220 class, you're going to be bombarded with claims 00:31:00.220 --> 00:31:03.799 online, on the news. This superfood cures cancer. 00:31:04.319 --> 00:31:06.980 Video games cause violence. Understanding methods 00:31:06.980 --> 00:31:09.099 allows you to ask the right questions. What was 00:31:09.099 --> 00:31:11.900 the sample? Was it just a correlation? How do 00:31:11.900 --> 00:31:14.759 they operationally define violence? Exactly. 00:31:14.960 --> 00:31:18.539 It's about being a skeptical, informed consumer 00:31:18.539 --> 00:31:20.799 of information instead of just accepting everything 00:31:20.799 --> 00:31:23.859 you hear. Which brings up a final thought. The 00:31:23.859 --> 00:31:26.279 text mentioned the Facebook emotional contagion 00:31:26.279 --> 00:31:29.039 study where they manipulated people's news feeds 00:31:29.039 --> 00:31:31.640 to see if they could make them happier or sadder 00:31:31.640 --> 00:31:34.240 without their consent. It was a huge ethical 00:31:34.240 --> 00:31:37.740 controversy. It just makes you think. With social 00:31:37.740 --> 00:31:40.940 media and big data, we're all participants in 00:31:40.940 --> 00:31:43.559 massive, unconsented experiments every single 00:31:43.559 --> 00:31:47.079 day. Every A -B test on a website, every algorithm 00:31:47.079 --> 00:31:49.680 change, are we okay with that? That is a deep 00:31:49.680 --> 00:31:51.660 and troubling question. Something to mull over 00:31:51.660 --> 00:31:54.180 after the exam. For sure. Okay, well, good luck 00:31:54.180 --> 00:31:56.099 on the exam. You're gonna crush it. We're both 00:31:56.099 --> 00:31:57.319 gonna crush it. See you there.