Textbook Notes (270,000)
CA (160,000)
UBC (3,000)
PSYC (800)

PSYC 305 Chapter Notes -Dependent And Independent Variables, Statistical Significance

Course Code
PSYC 305
Sunaina Assanand

of 2
Chapter 2-Methods in the Study of Personality
Gathering Information
-look at yourself/introspection or others (problems: distort what you see, can’t get
into their head)
Seeking Depth: Case Study
Personology-study whole person, not just one aspect (Henry Murray)
-observe person in natural environment, unstructured interviews
-many case studies are also clinical studies (person has problem)
Seeking Generality: Studying many people
-generality (breadth of applicability) is a continuum
Establishing Relationships among Variables
-to see if relationship exists btwn variables, must look at more than 1 variable (eg.
not enough to say low self-esteem=poor GPA…what does high self-esteem
mean?)-can’t do this in case studies
-in many examples, variables/dimensions go together in a systematic way (look at
strength + direction)
Correlation coefficient-shows how strong correlation is (1.0 means perfect positive
-negative 1.0 is perfect inverse correlation
0.6-0.8 is strong, 0.3-0.5 is moderately strong, below 0.3 or 0.2 is weak (more
Statistical Significance-the likelihood of an obtained effect occurring when there is
no true effect
-when probability is small enough, the correlation is said to be statistically
Clinical/Practical Significance-when correlation is believable (statistical significance)
and large enough to have practical importance (eg. can have high s.s. but only
account for tiny bit of behavior-low p.s.)
Causality-variation in one dimension causes variation in another (explains the why)
3rd-variable problem-possibility that an unmeasured variable caused variations in 2
correlated variables
Search for Causality: Experimental Research
Independent Variable-manipulated to see if it’s the cause
Experimental Control-trying to hold all other variables (except independent)
Random Assignment-how to treat variables that can’t be controlled so in a large
group, any significant differences will balance out (race, physique, depressed,
-if do all this, and groups differ in dependent variable at end, must have been b/c of
IV (only difference)
Problem: don’t know what aspect of IV caused difference
-short events in carefully controlled conditions (whereas correlational studies
let you examine elaborate events over a long period)
-correlation studies let you look at eg. effects of divorced parents on smoking
(unethical in experiment)
Multifactor Study/Experimental Personality Research-2 or more predictor variables,
varied separately
eg. People come in with low/high self-esteem (personality variable), success/failure
on test is manipulated (experiment variable)-> dependent variable is
performance on 2nd task
Main Effect-effect of 1 predictor variable independent of other variables (fail=both
esteems do worse)
Interaction-effect of 1 predictor variable differs depending on level of other
predictor variable
eg. failure has different effects on high/low esteems