Why is psychology a science? Scientific Method: System of gathering data so that bias and error in measurement are reduced
Why is psychology a science? Scientific Method: 1. Perceive question 2. Form hypothesis 3. Test hypothesis 4. Draw conclusions 5. Report results -Replication
Descriptive methods Naturalistic observation Watching behavior in normal environment Watching behavior in normal environment Major Advantage: Realistic data Realistic data
Descriptive methods Naturalistic observation Disadvantages: Observer effect Observer effect Observer bias Observer bias Lack generalizeability Lack generalizeability
Descriptive methods Laboratory observation Watching behavior in laboratory environment Watching behavior in laboratory environment Advantages: Control Control Specialized equipment Specialized equipment
Descriptive methods Laboratory observation Disadvantage: Artificial situation = artificial behavior Artificial situation = artificial behavior
Descriptive methods Case study One individual, great detail One individual, great detailAdvantage: Tremendous detailTremendous detailDisadvantages: Lack of generealizeabilityLack of generealizeability Biases in self-reportBiases in self-report Phineas GagePhineas Gage
Descriptive methods Surveys series of questions (e.g., Mate Preferences) series of questions (e.g., Mate Preferences) Representative sample - randomly selected sample of from a larger population Population - entire group of people/animals of interest
Sampling
Descriptive methods Surveys Advantages: Data from large numbers (large N) Data from large numbers (large N) Private information Private informationDisadvantages: Ensure representative sample Ensure representative sample Inaccurate data (courtesy bias, memory) Inaccurate data (courtesy bias, memory) Order, wording effects Order, wording effects
Measurement Scientific measurement requires: Reliability = consistency Reliability = consistency Validity = accuracy Validity = accuracy
Measure must be reliable: Coke vs. Pepsi Reliability: Interrater reliability-Do Interrater reliability-Do observers agree? Test-retest reliability- Test-retest reliability- Are measures consistent over time? BUT people might self report liking Coke very consistently simply because they reliably like the advertising.
Advertising preferences
Measure must be valid: Coke vs. Pepsi Internal validity-Are effects due to conditions effects due to conditions manipulated? manipulated? e.g., preference may come from product advertising or from product taste (e.g., initial sweetness)
Measure must be valid: Coke vs. Pepsi Internal validity-Are effects due to conditions effects due to conditions manipulated? manipulated?... A blind taste-test would be a better measure…it would eliminate brand loyalty as a variable ( = higher internal validity)
Types of research designs: relationships between variables Correlational Experimental Quasi-experimental
Types of research designs: relationships between variables Correlational: Association between variables Association between variables No manipulation No manipulation Correlation coefficient: r (range: ) Correlation coefficient: r (range: ) Correlations does not equal causation! Correlations does not equal causation!
Correlational research Correlation relationship between two variables relationship between two variablesVariable Anything that can change or varyAnything that can change or vary
Types of research designs: relationships between variables Experimental: Causal relationship between variables Causal relationship between variables Experimental manipulation Experimental manipulation
Types of research designs: relationships between variables Quasi-experimental: No manipulation possible No manipulation possible
Correlational research
Correlation Measures of two variables Measures of two variables Mathematical formula Mathematical formula Correlation coefficient (r), represents Correlation coefficient (r), represents 1. direction of relationship (+/-) 2. strength of relationship
Correlational research Correlation coefficient (r): –1.00 to Closer to 1.00 or -1.00, the stronger the relationship No correlation = 0.0 No correlation = 0.0 Perfect correlation = OR Perfect correlation = OR +1.00
Correlational research Positive correlation – same direction One increases, the other increases; one decreases, the other decreases One increases, the other increases; one decreases, the other decreases Negative correlation – opposite direction One increases, other decreases. One increases, other decreases. CORRELATION DOES NOT EQUAL CAUSATION!!!
Experimental research manipulation of a variable to see if changes in behavior result determination of cause-and-effect relationships
Experimental research Operational definition Definition variable of interest Definition variable of interest Allows precise, direct measurement Allows precise, direct measurement Example: Aggression
Variables in experimental design Independent variable-what is manipulated e.g., Prozac dosage e.g., Prozac dosage e.g., Violent TV e.g., Violent TV Dependent variable-what is measured, the outcome e.g., Depression e.g., Depression e.g., aggressive play e.g., aggressive play
Participants in experimental design Experimental group: group whose environment or experience has changed e.g., Prozac e.g., Prozac e.g., TV e.g., TV Control group: group treated as much like experimental group as possible Does NOT participate in experimental manipulation Does NOT participate in experimental manipulation e.g., No Prozac, sugar pille.g., No Prozac, sugar pill No TVNo TV
Random assignment Assignment to experimental or control groups randomly Assignment to experimental or control groups randomly Participants have an equal chance of being in either group Participants have an equal chance of being in either group Controls for confounding variables
Experimental design Placebo effect expectations of the participants influence behavior (e.g., Viagra vs. sugar pill) expectations of the participants influence behavior (e.g., Viagra vs. sugar pill) Single-blind study Participants do NOT know if they are in the experimental or the control group (reduces placebo effect) Participants do NOT know if they are in the experimental or the control group (reduces placebo effect)
Experimental design Experimenter effect Experimenter’s expectations influence results Experimenter’s expectations influence results Double-blind study Neither the experimenter nor the participants know group assignment (reduces placebo and experimenter effect) Neither the experimenter nor the participants know group assignment (reduces placebo and experimenter effect)
Correlational design
Results:
Correlational design Conclusions: NO causal relationship can be inferred NO causal relationship can be inferred
Experimental design Question: Does close physical contact promote secure infant-mother attachment? Method: Women randomly assigned to use “Snuglis” (experimental) or standard infant seats (control) to carry their infants
Experimental design Results: Carrier effect on attachment Snuglis: more responsive moms, infants more likely to develop secure attachment Snuglis: more responsive moms, infants more likely to develop secure attachment Standard carrier: less responsive moms, infants less likely to develop secure attachment Standard carrier: less responsive moms, infants less likely to develop secure attachment Conclusion: causal relationship between type of carrier and attachment