exam 2

Intro Soundbite: “Welcome to Chapter 1 — where we explore the heart of experimental research: control. Why is it essential? What does it mean? And how do you do it right?” Narration Script: Let’s start with the basics. Control in research is about increasing internal validity — in other words, making sure that what you’re testing is what’s actually causing the effect. No distractions. No imposters. There are two meanings of control: Control experiment: where one group gets the treatment and the other doesn’t — classic control vs. experimental group. Experimental control: all the techniques researchers use to eliminate confounds and keep things clean. Here’s how you do it: Random Assignment: Split people into groups by chance. Why? To make sure personal traits don’t influence results. This boosts internal validity — different from random sampling, which is about generalizing to a bigger population. Matching: Pair people up based on traits like IQ or age before assigning groups — especially helpful in small samples. Subject as Own Control: Also called within-subjects designs — the same person gets every condition. Efficient, but watch out for practice and fatigue effects. To make sure your treatment really worked, we use manipulation checks — like asking, “How stressed did you feel after the task?” They’re not pretests. They’re not attention checks. They’re their own thing. Then there’s instrumentation of response — fancy term for using standardized, reliable tools to measure stuff — like validated anxiety scales. Now for the unsung hero: nuisance variables — annoying things like time of day or noise. You can build them into the design, or use statistical control to clean them out later with ANCOVA or regression. Pro Tip: Random assignment is more effective with larger sample sizes. Smaller samples = more risk of imbalance. We’ll wrap this chapter with an essential distinction: Extraneous Variable: any third wheel in your study. Confounding Variable: the third wheel that’s dating your IV and influencing the outcome — not good. Want to go deeper? Learn to identify: Mediator: explains how the IV causes the DV. Moderator: changes when or for whom the effect happens. Bonus Tip: If a nuisance variable changes the strength of the IV–DV relationship, it becomes a moderator. Oh, and beware of experimenter effects — like bias in how the study is run or how participants are treated. Control them using scripts, automation, or even double-blind procedures. 🎉 Chapter 1 Takeaway: Control is the invisible scaffolding that holds a study together. Master it, and you master experimental logic. 🎙️ Chapter 2: Designing the Perfect Experiment – Tools of the Trade Intro Soundbite: “In Chapter 2, we’re rolling up our sleeves and building experiments from scratch. How do you design one? What makes it ‘true’? Let’s go!” Narration Script: You’ve heard the phrase “true experiment”, but what does it really mean? ✅ It has an independent variable (IV) that you manipulate. ✅ It uses random assignment. ✅ It controls confounding variables. That’s it. That’s the big three. Now, when designing, you can choose lab vs. field: Lab = control city. High internal validity. Field = real-world vibes. High external validity. Let’s talk groups: Control Group: No treatment. Experimental Group: Gets the treatment. Each group is assigned to a condition — like sugar water vs. no sugar water. Pretests? They’re useful — even essential — for three things: Establishing a baseline, Checking if random assignment worked, Serving as a covariate during analysis. Need more control? Enter the Solomon Four-Group Design. Four groups: One with pretest + treatment, One with pretest + no treatment, One with treatment + no pretest, And one with neither. This design lets you test for testing effects, which sneak in when people get better just because they’ve seen the test before. Now let’s look at flawed designs to avoid: One-group posttest-only: You have nothing to compare to. One-group pretest-posttest: You risk effects from time, like people just getting older or smarter. Posttest-only with nonequivalent groups: You didn’t randomly assign, so maybe one group started better than the other. 🎯 Between-subjects vs. Within-subjects: Between = different people in each condition. Within = same people in all conditions. Each has pros and cons — within is efficient but brings order effects. Between avoids order effects but needs more participants. Finally: placebo effects are real. If people think they’re getting treatment, they might get better even if it’s just a sugar pill. That’s why placebo groups and double-blind procedures are gold in clinical research. 🎉 Chapter 2 Takeaway: A solid experiment is about balance — between control and reality, between groups and variables, and between creativity and structure. 🎙️ Chapter 3: Factorial Designs, Interactions, and Interpreting It All Intro Soundbite: “Ready to talk stats, design logic, and how to read those tricky graphs? Chapter 3 is your deep dive into factorials, interactions, and making sense of the numbers.” Narration Script: Imagine an experiment with two IVs. One has 2 levels. The other has 3. That’s a 2 × 3 factorial design. Multiply them — 6 total conditions. Factorial designs let you look at: Main Effects: the effect of one IV across all levels of the other. Interactions: when one IV’s effect depends on the other IV. 📊 Graph Tip: If the lines are parallel, you probably have a main effect. If the lines cross or diverge, that’s an interaction. You’ll also see mixed factorial designs — part between-subjects, part within-subjects. Maybe everyone tries two types of training, but only one group gets caffeine. If you’re shown a graph or description of a study, ask yourself: How many IVs? How many levels per IV? How many groups or conditions total? Are there main effects? Interactions? What’s the best analysis — t-test, ANOVA, regression, or chi-square? 🎯 Pro Tip: T-test for two groups, ANOVA for 3+ groups, Regression for continuous variables, Chi-square for categories. Don’t forget — your lab study (Study 2) is a case study in itself. Describe it clearly: What was the IV and DV? What was the design — within or between? What control strategies did you use? What kind of validity did you maximize? 🎉 Chapter 3 Takeaway: Factorial designs help you move from simple comparisons to complex, real-world interactions. You’re not just testing if something works — you’re testing how, when, and for whom.