Experimental Psych Quiz #2 Review PDF

Chapter 10: Factorial designs, main effects, interactions, design matrix, graphing results, short-hand design notation CHAP. 6 ❖ SYNTHETIC STATEMENTS: ❖ Statements that can be either true or false, and formulate into if-then statements. ❖ Deals with the relationship between variab...

Chapter 10: Factorial designs, main effects, interactions, design matrix, graphing results, short-hand design notation CHAP. 6 ❖ SYNTHETIC STATEMENTS: ❖ Statements that can be either true or false, and formulate into if-then statements. ❖ Deals with the relationship between variables “"If students are hungry, then they will read slowly," is ____.” Analytic statements are true by definition. Their truth can be determined just by understanding the meanings of the words. Example: "All bachelors are unmarried." This is true because the definition of "bachelor" includes "unmarried." Contradictory statements: there is always one false statement: Statement A: "Sleep deprivation leads to decreased cognitive performance." Statement B: "Sleep deprivation enhances cognitive performance." These statements conflict with each other based on well-established research on the effects of sleep on cognitive functioning. Both claims cannot be true at the same time, making them contradictory. INDUCTION VS DEDUCTION: inductive: Inductive reasoning involves starting with specific observations or data and making generalizations or theories based on these observations.REASONS FROM SPECIFIC CASES TO GENERAL PRINCIPLES Researchers collect data from experiments or observations, and from that data, they induce general principles, patterns, or theories. ➔ INDUCTION MAKES A THEORY For example, after observing that people tend to perform better on memory tasks when they are in a good mood, a researcher might generalize that positive emotions improve memory performance. deductive:Deductive reasoning begins with a general theory or hypothesis and uses specific observations or experiments to test whether this theory holds true in particular cases. FROM GENERAL TO SPECIFIC ➔ DEDUCTION TESTS A VALIDITY OF A THEORY OR HYPOTHESIS For example, if a theory suggests that sleep deprivation impairs cognitive performance, a researcher might set up an experiment to test this hypothesis by comparing the performance of sleep-deprived and well-rested participants. ★ EXPERIMENTAL VS NON-EXPERIMENTAL HYPOTHESIS An experimental hypothesis predicts that a change in one variable (the independent variable) will lead to a change in another variable (the dependent variable). -To test cause-and-effect relationships between variables. Example: "Increasing the amount of sleep will improve memory recall in students." ○ In this case, the independent variable is the amount of sleep, and the dependent variable is memory recall. The hypothesis suggests that sleep affects memory. ★ NON EXPERIMENTAL HYPOTHESIS A non-experimental hypothesis predicts a relationship or association between variables but does not involve manipulation. Instead, these hypotheses are based on natural observations or correlations. -​​To explore associations or patterns that naturally exist in the world. Example: "There is a positive relationship between the amount of time spent studying and academic performance." ○ This hypothesis suggests that more study time is related to better academic outcomes, but it does not imply that studying "causes" improved performance, only that a relationship exists. If people live together before marriage, then they will be more likely to divorce. IS A NATURAL OBSERVATION, ISN'T MANIPULATING ANYTHING NOR SAYING THAT IT IS CAUSED BY THAT. ❖ PARSIMONY THE SIMPLE THE BETTER AVOID QUALIFIERS AND CAVEATS WHEN POSSIBLE Walking across campus at night is dangerous if you are alone and unarmed. THIS IS NOT PARSIMONIOUS CHAP. 7 IV and DV Independent variable is what will be manipulated Dependent variable is what will be measured OD Explaining exactly how you will be measuring your experiment If you are going to do an experiment on memory you would explain exactly how you will measure that. "Okay, when I say 'memory,' I mean how many words you can remember after I say them to you." So instead of just using the word "memory," you explain exactly how you'll measure it, like counting the number of words someone remembers. operational definition helps everyone understand exactly what you mean when you talk about something in an experiment ★ Experimental OD This is how a researcher changes or controls something in the experiment. It's the action or condition that gets manipulated to see if it causes any effect. ★ Measured OD This is how the researcher measures something in the experiment. It's the way they define the specific thing they're observing or counting to see what happens. WE NEED OD TO MEASURE HYPOTHETICAL CONSTRUCTS ★ THE 4 LEVELS OF MEASUREMENT NOMINAL:labels or categories (e.g., colors) ORDINAL:has a order (e.g., rankings) INTERVAL:has an order with consistent differences but does not have a true zero(0 does not mean that there isn't nothing of what's being measured for example 0 celsius doesn't mean it doesn’t exist, 0 is a number in the scale) (e.g.,temperature) RATIO:order with consistent measurement but HAS an real 0 (0 actually means “none,nothing”) (e.g.,height,weight,time) 3 types of reliability -Test-retest-If you take the same test at different times, you should get about the same result each time. -Imagine you take a quiz today and then again next week. If the test is reliable, your score should be almost the same both times. -Inter-rater-reliability-If different people give the same test or rating, they should get about the same answer. -If two teachers grade your essay, and they both give it an "A," that means the grading is reliable. They agree! -Intern-item -is a measure of how well the individual items or questions on a test or survey correlate with each other. In other words, it checks if the different questions are all in agreement about what they’re measuring. ➔ Imagine you have a questionnaire about how friendly someone is. You might ask questions like: "Do you smile at people?" "Do you help others when they need it?" "Do you greet people when you see them?" If all these questions are really measuring how friendly someone is, they should all give similar answers. For example, if someone answers "yes" to one of the questions, they should likely answer "yes" to the others, too. 5 types of validity Face validity Face validity is the most basic type of validity. It just means that on the surface, the test looks like it’s measuring what it’s supposed to measure. -Example: If you have a math test with questions like “What is 2 + 2?” and “What is 5 x 3?” It's obvious that the test is about math. This would have high face validity because the test looks like it’s testing math. Content validity:Content validity checks if a test covers all the important parts of the topic it’s supposed to measure. It’s not just about looking right, it’s about being complete. -Example:If you have a history test on the topic of the American Revolution, content validity means making sure the test covers all the important areas—like the causes, important battles, and key people—not just one or two parts. Predictive validity:Predictive validity is about whether a test can predict future behavior or outcomes. It checks if the test results match something that happens in the future. -Example: If a college entrance exam predicts how well students will do in college (like their GPA), the test has predictive validity. The higher the test scores, the better the students tend to do in college. Concurrent validity: Concurrent validity checks if a test gives results that match other tests measuring the same thing, but at the same time. -Example: If you develop a new math test, you could compare its results with another well-known math test that is already trusted. If both tests give similar results for the same group of students, the new test has high concurrent validity. Construct validity: Construct validity is about making sure that a test actually measures the concept or trait it is supposed to measure (like intelligence, happiness, or math ability). It checks whether the test truly reflects the underlying idea or construct. Example: If you have a test for anxiety, construct validity checks if the test is really measuring anxiety and not something else like stress or general nervousness. The questions should clearly relate to anxiety and not to other feelings. ➔ INTERNAL VS EXTERNAL VALIDITY Internal: refers to the degree to which an experiment or study accurately measures the relationship between the variables being studied, without interference from other factors. In other words, it's about whether the changes you see in the dependent variable (what you're measuring) are really caused by the independent variable (what you're manipulating), and not by something else. External: External Validity refers to how well the results of a study can be generalized to other settings, people, times, or situations outside the study itself. It’s about whether the findings apply to real-world scenarios. Chap. 8 ➔ Physical variables: noise, distractions, lighting, etc. ➔ ELIMINATION:removing any factors that could affect the experiment, eliminate those physical variables:sound proof room, do not disturb sign ➔ CONSTANCY:keep the conditions (the levels of the IV) as constant as possible,ensuring they don’t change and affect the results. ➔ BALANCING:This means making sure that variables are evenly distributed or balanced across different experimental groups to avoid bias. ★ Demand characteristics & experimenter bias -Experimenter bias happens when the researcher’s expectations or preferences influence the way they conduct the experiment or interpret the results. This can unintentionally affect the outcomes, making them reflect the experimenter's beliefs rather than the actual results. Example: If a researcher believes a new drug will be effective, they might give more positive feedback to participants in the experimental group or give them more attention, making it more likely that those participants will perform better in the study. -Rosenthal effect EXPERIMENTER BIAS CAN BE AVOIDED BY USING THE DOUBLE BLIND: where both the researcher and the participants don’t know which group is receiving the treatment, can help minimize experimenter bias. SINGLE BLIND:only the participants don't know which condition they are in.  -Demand characteristics are cues or clues in the experiment that influence participants to figure out what the experiment is about or what the researcher expects them to do. In other words, participants might change their behavior because they want to give the researcher the answer they think is expected. ➔ SOCIAL ENVIRONMENT CONFOUNDS: the social environment (other people, social pressures, group dynamics) can influence participants’ behavior and mess with your study's results. ➔ PERSONALITY CONFOUNDS: Individual personality traits (like introversion, anxiety, etc.) can affect how participants respond, introducing extra variability into the data that’s unrelated to the experiment CHAP. 9 IV=>DV cause=>effect independent variable=what will be manipulated dependent variable=what is being measured ➔ Between sub-design: Different participants assigned to different groups Participants are in only one level of the Independent variable Make comparisons between two or more different groups -You need at least 2 participants per condition -Large samples are more likely to get full range of people ➔ Matched groups: A technique used to pair participants in different groups based on similar characteristics to reduce potential confounding variables. This helps ensure that any observed effects are due to the treatment, not other factors. If you are worried about a confound, you should probably use matched groups ➔ Control groups: A group of participants that does not receive the experimental treatment. The control group is used for comparison to determine whether the treatment has an effect on the experimental group. -usually received placebo ➔ Sample size: The number of participants included in a study. ➔ Power:the probability of detecting an effect if one truly exists, the ability of statistical test to detect real differences ➔ Effect size: In short, effect size tells us how strong or weak the relationship is between what we’re testing -Testing how much happiness two different flavors of ice cream brings -You give some people chocolate ice cream and others vanilla ice cream, and then you ask them how happy they feel on a scale from 1 to 10. -effect size is like a measure of how big the difference is in happiness between the two groups of people. If the chocolate ice cream group is super happy, scoring around 9, and the vanilla group is not so happy, scoring around 2, that's a big difference. That would be a large effect size—the ice cream choice really seems to matter a lot. ❖ Random assignment:Imagine you have a big group of people, and you want to test how chocolate vs. vanilla ice cream affects happiness. But instead of letting people choose which ice cream they want, you randomly decide who gets which flavor. So, random assignment means you're not picking people based on who likes chocolate or who likes vanilla As a rule of thumb you should assign at least 20 subjects to each treatment group. In psychotherapy, waiting-list conditions is usually used as the CONTROL GROUP Chapter 10: Factorial designs, main effects, interactions, design matrix, graphing results, short-hand design notation CHAP 10 FACTORIAL DESIGN: A factorial design is when you test more than one thing at the same time to see how they work together. For example, you might want to know how both ice cream flavor (chocolate or vanilla) and temperature (hot or cold) affect happiness. You test all the different combinations of those factors (chocolate and hot, chocolate and cold, vanilla and hot, vanilla and cold) to see what happens. MAIN EFFECTS: The effect of just a single IV in the experiment A main effect is like looking at just one factor on its own. For example, you might ask, “Does ice cream flavor by itself make people happy?” Or, “Does the temperature make a difference by itself?” A main effect tells you what happens with just one thing, like how chocolate makes people happier than vanilla, or how hot temperatures make people happier than cold. INTERACTIONS: An interaction happens when two factors (like ice cream flavor and temperature) work together in a way that’s different from what you’d expect. For example, maybe people like chocolate more when it’s cold, but vanilla is better when it’s hot. That’s an interaction—the combination of ice cream and temperature makes a bigger difference than just one of them alone. Let’s say you’re testing how ice cream flavor (chocolate vs. vanilla) and temperature (hot vs. cold) affect people’s happiness. You might expect the following:  Ice cream flavor: Chocolate might make people happier than vanilla. Temperature: People might like ice cream more when it’s cold, rather than hot. But when you test both factors together, you might find something surprising: Chocolate ice cream is much better when it's cold. Vanilla ice cream is much better when it's hot. This is an interaction because the combination of ice cream flavor and temperature changes the result in a way that wouldn't happen if you looked at them separately. DESIGN MATRIX: A design matrix is like a table that shows all the combinations of the factors you’re testing. It lists out all the different groups you’ll have based on the different flavors and temperatures. It helps you keep track of everything NAME LENGTH NAME TYPE A CHART WITH ONE IV AS A COLUMN HEADER AND THE OTHER AS A ROW A B C M=7 M=9 M=11 M=13 M=11 M=9 Average in first row=9 Average in second row=11 They do better in group than alone First column 7+13/2=10 Second column 10 Third column 10 GRAPHING RESULTS: NO INTERACTION INTERACTION  We use shorthand notation: 2 x 2 (2 IV(factors) each with two levels) 3 x 2 (2 IV, one with 3 levels & one with 2 levels) 2 x 2 x 2 (3 IV, each with two levels) Short-Hand Design Notation Short-hand design notation is a quick way to write down how many factors you have and how many levels (options) there are for each factor. For example: A 2x2 design means two factors, each with two levels (like chocolate vs. vanilla, and hot vs. cold). A 3x2 design means three factors with two levels each.