PSY 255 1st Edition Lecture 3 Outline of Last Lecture I. History and Future of I-O PsychologyOutline of Current Lecture- Chapter 2 II. ScienceIII. Assumptions and TheoriesIV. MethodsV. Research MethodsCurrent LectureScience: A process for generating knowledge• Science has 4 goals– Description: answers the question of WHAT? Describes the phenomenon.– Explanation: WHY does this phenomenon occur?– Prediction: WHEN? Giving the circumstances, makes predictions on when the phenomenon happens. – Control: Makes changes to observe the phenomenon.• Assumptions– Empiricism: learning via observation and data collection. (ex: surveys, interviews, observing through eyes, etc.)– Determinism: Phenomenon is orderly and systematic. Nothing is random. If we study really hard we can understand and explain why things happen.– Discoverability: Phenomenon is knowable. It is possible for us to understand the mechanisms leading to this phenomenon.• TheoryThese notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.– A set of interrelated concepts and propositions that present systematic view of phenomenon.– Theories are the building block of science.– Theories describe, explain and predictCriteria of what a good theory is• A good theory is:– Parsimonious: Simplest is best. (ex: Gravity is a simple theory)– Precise: Specific and accurate. We don’t want it to be wish-ey washy (a little of this, a tad bit of that is NOT GOOD. Make it precise like a recipe and have exact “measurements”)– Testable: Falsifiable (not yet disconfirmed). We need to be able to test it and see if it works. We assume that all theories are in between totally wrong and absolutely correct.– Useful: Do people care about this? If you think it’s useful, but other people do not, it will be harder to make the theory important.– Generative: It should not be limiting. It should open up ideas for new studies and further research.• Theory or data?– Induction Method• Moving from data to theory• Don’t start with guesses or ideas about how it works…you will just go collect data first. You have no reason to believe one way or another.• Look for patterns from the data– Deduction Method• Moving from theory to data• Design studies to test existing theory• You have some idea of why it happens, and then I come up with some sort of test to see if you are right.– Combination of the both is best!!!!Research• Basic concepts– Causal inference: conclusion about causal relationship between 2 variables:– Independent variable (IV or predictor)…IV causes changes in DV– Dependent variable (DV or outcome)– Extraneous or confounding variable (Z variable)- For IV to CAUSE DV this is the criterion:- IV must occur before DV (x causes y)- There is a correlation between them. They change together in a predictable manner. (x and y are related)- No plausible explanation (this is the hardest to see). - Basic concepts: Internal validity- Accuracy of causal inferences (how accurate is the statement that you made?)- Does IV really cause DV?- USE CONTROLS!- Controlling extraneous variables- Hold it constant: If salary has something to do with job satisfaction, we can study only people making a certain amount of money to keep it constant and control that variable. Select people based on the variable.- Manipulate: Study people consciously that you know only makes minimum wage, then another group that makes a little higherthan minimum, then another group that makes more than $10…see if there are group differences.- Statistically control: Statistically remove variance before I look at data (this is the weakest control).- Random assignment (best control): Everyone has an equal chance of being in control or experimental position.- Double-blind method: Neither researcher, nor participant know whether participants are in the control or experimental group.- Basic concepts: External validity- Extent that findings generalize to/across different people, settings, and time (will I find the same results across populations? MSU, U of M, Akron, OSU, and Carnegie Mellon?)- Trade-off between internal versus external validity- The more control you have, the less realistic it is Research Designs• Experimental methods involve:– Random assignment: Participants are equally likely of being assigned to any condition.– Variable manipulation: Manipulate IV only and control all other variables.• Other experimental methods– Field (vs. laboratory) experiments: Random assignment and manipulation in a “real world” setting. REALLY RARE IN I-O PSYCH.– Quasi-experiments: NO random assignments. You will use in tact groups. This is very common in I-O. We don’t have a lot of control. We can test based on existingdifferences, and by location/department.• Observational methods (correlational design, descriptive methods)– Neither random assignment nor manipulation• In some cases, it is not possible to do random assignment or manipulation (ex: Food intake for a pregnant woman. You can’t make them not eat, for health reasons)– Good for description and prediction, but difficult to make definitive statement about