Basic Experimental Design Film Experimental Research Methods in
Basic Experimental Design
Film: Experimental Research Methods in Psychology Parts 1 -2 Search by title in Films on Demand
Two or more variables observed in isolation from others n Independent variable: physically manipulated variable; “cause” n Dependent variable: measured variable; “effect”
Levels of the IV n Each possible variation in the IV is known as a level. n For every level you need a “group. ” n Example: Room Temperature – 3 Levels: Hot, Cold, Neutral
Assignment to Groups True experiments: n Random assignment (“allocation” if you’re a Brit) n Not to be confused with random selection
Groups n Control: No treatment; inactive/placebo n Experimental: treatment; active ingredients n Some experiments only have different experimental groups and no control group.
Controlling Variables Extraneous Variables: Undesired variables that might influence the DV. n Examples: n – – Participant characteristics: IQ, mood, marital status, age Environmental factors: room temperature, race of experimenter
Quick Review In a recent study 61 young girls were stressed by having them give speeches to a crowd and work out math problems in front of judges. Afterward, the girls hugged their moms, talked to their mom on the phone or received no contact. The researchers then measured their levels of the “love hormone” oxytocin and the “stress hormone” of cortisol. Both the huggers and phone-talkers experienced a rise in oxytocin and a drop in cortisol, whereas those with no contact showed no such pattern. 1. Identify the IV and its levels 2. Identify the DV(s) 3. Identify some controlled variables other than the IV 4. Identify some uncontrollable variables.
Causal-Comparative Study AKA: Ex Post Facto Study n The causal variable is not actually manipulated by the experimenter. n Naturally occurring IV. n – – IV: Smokers vs. non-smokers DV: rates of lung cancer
Group Activity #1 n Handout: Experimental vs. Non. Experimental n Instructions: – – – Decide if this question can be studied experimentally or not. Identify the likely IV and DV
Group Activity #2 n Handout: Studies from “The Week” Instructions: n Discuss the type of hypothesis most likely in each. n Identify the IV(s) and the DV(s) of each study. n Discuss extraneous variables they controlled for (or probably controlled for).
Experimental Dilemma Tradeoff between control and generalizability Control: limited sample, limited variables. Strong evidence. Generalizability: applies to more people and more variables. More like the real world
Video Continued: Parts 3 -4
Dealing with Extraneous Variables: 1. Hold some variables constant. - Control the most problematic stuff only. 2. Allow some to vary - Random assignment - Keeps the generalizability.
Random Assignment vs. Random Selection n Assignment: equal chance of being assigned to any IV level or group. – n Essential to a true experiment Selection: equal chance of being included in a sample that is selected from a known population.
Confounding Variables (AKA: rd The 3 variable problem) Those variables that systematically vary with the IV such that you can’t tell what caused the change in the DV. All confounding variables are extraneous variables but not all extraneous variables are confounding variables.
Example n Study of ad campaigns and voting attitudes. – Two different ads shown to participants. n Participants’ political party varies (EV). n Participants shown Ad #1 were placed in a room where an SGA poster was hanging. Those shown Ad #2 were not exposed to that same SGA poster. (potential confounder)
When should you worry? If an EV affects all (or most) of one group and none (or few) of the other group. n If your samples are small (< 20 per group). n If you didn’t use random assignment. n If an uncontrolled EV is strongly related to the DV. Be very worried if: n You have more than 1 of these. Your study is worthless if you have all 4. n
Video: Parts 5 -6
Research Design Notation n Short-hand notations used to indicate key aspects of a research design.
Research Design Key R = Random assignment O = Observation (pre or post) X = Event/Intervention/IV N = Non-equivalent groups
True Experiment Types Pre-test-post-test randomized control group design R O X O R O O Posttest-only randomized control group design R X O R O
Group Work n n Use this time to plan your experiment based on what literature you have gathered and your understanding of experimental designs and ethics thus far.
- Slides: 23