PSYC 102 Lecture Notes - Lecture 1: Richard Feynman, Naturalistic Observation, Confirmation Bias

18 views3 pages
18 Feb 2016
School
Department
Course
Page:
of 3
Introduction Jan 5
definition:
- scientific study of BMP
- how we think (cognitive), feel (emotional), and act (behavioural)
goals:
- describe explain predict control
methods:
- understanding = may methods converge to same conclusion
2 basic questions:
1. how we know what we know
2. how we learn what we do not know
reasoning:
- regression to the mean = any spike will level out in time
- leads to the assumption of causation where there is none
- we cannot know what causes what, without carefully controlled studies
- personal experience does NOT = valid evidence
Scientific Method Jan 7
Our best protection against sloppy thinking & human reasoning unfailing method
Common Traps
Hallmarks of Scientific Method
Cal “aga: “iee is a uiue
i of opeess ad skeptiis.
Rihad Fea: “iee is a
wa of tig ot to fool ouself.
You ust dout epets.
1. objective observation and logically necessary conclusions
2. parsimonious explanations stick w/ efficient explanation
3. independent replication
4. skepticism
5. careful designs
6. falsifiability predictions that can potentially be falsified
7. open mind to ideas of others
Research
Process
(a cycle)
1. theory e.g. low self-esteem feeds depression
2. hypothesis e.g. people w/ low self-esteem score higher on depression scale
3. research/observations e.g. administer tests, look for correlation
4. generate or refine the theory
Results
- inaccurate beliefs discredited, replaced w/ accurate beliefs
- no tech to study/progress, so we remain temporarily ignorant
Vocab
1. theory: explanation integrates principles, organizes & predicts behaviour or events
2. hypothesis: testable prediction to endable us to keep, reject, or revise the theory
3. construct: what is being measured
4. operational definition: method used to define/measure variable in study
5. replication: same finding w/ different participants in different situations
6. random sampling: each member of population has = chance of inclusion in sample
(also called unbiased sample)
3 Basic Types of
Scientific Studies
1. descriptive
purpose: careful and accurate description
methods:
- questionnaires & interviews
- naturalistic observation
- case studies
e.g.: amount of sleep and depression
2. correlational
purpose: evaluating relationships
methods:
- correlation coefficient: number between -1 to +1,
shows direction & strength of relationship
- scatterplot: graph comprised of points generated
by values of 2 variables (slope=direction,
scatter=strength)
common error:
- illusory correlation: perception of relationship
where none exists (believe one large outcome)
e.g.: is amount of sleep related to depression
3. experimental
purpose: explore cause and effect
methods:
- focus on effect of independent variable
(manipulated) on dependent variable
(potentially influenced)
- observe experimental condition (exposed to
treatment) vs control condition (unexposed)
- placebo: inert substance/condition replacing
active treatment agent
- random assignment: assigned using chance
alone minimize differences across groups
- double-blind: participants & researchers
uawae of eah patiipat’s status
Statistics:
analyze &
interpret data
using objective
measures (NOT
human judgment),
conveniently
1. descriptive
statistics:
describe data
accurately
and
meaningfully
measures of central tendency:
- mode: most frequently occurring
- median: middle score (rank-ordered distribution)
- mean: arithmetic average (add, divide # scores)
measures of variation:
- range: difference between highest & lowest
- standard variation: avg variation around mean
distribution:
- normal distribution: score frequency decreases as
scores get farther from most frequent
- skewed distribution: extreme values pull the
median and mode
2. inferential
statistics:
statistical statement of
how frequently a result
could occur by chance
reliable when
- large samples
- large mean difference (correlational
relationship)
- less variability
significance
- correlation is unlikely to happen by chance
- p<0.5