Critical Thinking & Scientific Method (Unit 2)

Scientific Method

= uses empirical evidence to establish facts 

steps:

1. develop theory

2. create falsifiable hypothesis from it

3. test hypothesis by observing the world

Theory vs. Hypothesis

T: explanations of natural phenomena (specific predictions) 

H: falsifiable prediction made by a theory 

Empirical Method

set of rules/techniques for observation 

2 Aspects of Measurement

1. define property: create operational definition w construct validity 

2. detect property: create instrument with reliability & power 

Defining Properties

• operational definition: description of property in measurable terms

• construct validity: operations are considered good indicators of specified properties 

Detecting Properties

• reliability: detect absence of changes in property 

• power: detect presence of changes in property 

Demand Characteristics

aspects of observational setting causing person to act how they think someone else wants/expects 

Avoiding Demand Characteristics

1. naturalistic observation (observe people in natural environments)

2. privacy/control (anonymous responses/measure invol behaviours)

3. unawareness (dont know true purpose of observation → dont know how to behave)

Problems with Naturalistic Observation

• some events dont naturally occur

• some need direct interaction

Observer Bias

= expectations influence what one believes they observed & whats actually observed

avoid with

• double blind study: researcher & participant dont know how participants are expected to behave

Graphic Representations

= often frequency descriptions (# of times measurement of property takes on each value)

• negatively skewed: lean left

• positively skewed: lean right

• normal description: frequency of measurements highest in middle, ↓ symmetrically both ways

Descriptive Statistics

=summary captures essential info from frequency distribution

• mean: average

• median: value in middle

• mode: occurs most often

Measurements of Variability

• range: largest measurement - smallest in frequency distribution (small range = less variability)

• standard deviation: how each measurement differs from mean

Variables

properties that can have more than 1 value

Correlation

relationship b/w variables where variations in one synchronized w variations in other

• can make predictions, not always accurate 

• measures strength & direction

Correlation: Direction

• positive: variables have more is more relationship

• negative: variables have more is less relationship

Strength: Correlation Coefficient

measures both direction & strength of correlation

• symbol= r (limited range) 

Correlation Types

• perfect positive: r = 1, variables move in same direction @ constant rate

• perfect negative: r = -1, move in opp directions @ constant rate

• no correlation: r = 0, variables dont move symmetrically

(1 & -1 uncommon in real world)

Natural Correlation & Third Variable Problem

NC: correlation observed in natural world 

TVP: correlation of 2 variables not considered causal because a third variable may be causing them both

Experimentation

= establishes causal relationships b/w variables 

1. manipulate: indepen variable (value determined)

2. measure: depend variable (→ effect of indepen)

3. compare: compare depen variable from control/experimental groups

Manipulation

technique for determining causal power of variable by changing value

Self-Selection

=occurs when anything about participant determines value of indepen variable exposed to 

fix: random assignment

• randomly assigns to experimental/control

Internal Validity

confidence that indepen caused change in depen → allows experiment to establish causal relationships

• (high= experiment working to determine causation) 

Case Study

gathers scientific info by studying single person 

External Validity

variables operationally defined in representative way

• (how variable is operationally defined impacts findings)

Population vs. Sample Size

p: N

s: n

• try to get n as close to N as possible → random sampling (every person from pop has equal chance of selection)

• sample representative of pop = can generalize

Type I & II Errors

• I: conclude causal relationship when none (false pos, fluke) 

• II: conclude no causal relationship when there is (false neg, flunk) 

Francis Bacon

= once opinons adopted → find everything that supports

• see what we expect to see

• same evidence → more sure of initial beliefs

• forget evidence we dont see

Ethics: Respecting People

• Tri-Council Policy Statement (TCPS): ethics for human research, main principles:

1. respect for persons (make decisions for self)

2. concern for welfare (maximize benefits/reduce risks) 

3. research is just (distribute risks/benefits equally to participants) 

Most Important Ethical Rules

1. informed consent: verbal agreement to participate by adult informed of risks

2. freedom from coercion

3. protection from harm: prevent physical/psychological harm

4. risk benefit analysis: no risks greater than taken in daily lives, risks outweighed by social benefits/knowledge

5. deception: only when justified

6. debriefing: deception used → must debrief (verbal description of true nature/purpose of study) after

7. confidentiality: private/personal info from study confidential

Ethics Enforcement

• research ethics boards (REBs) → study only after reviewed by

• includes: researchers, uni staff, people from community

• federally funded = nonscientist + person not affiliated w institution 

Ethics: Animals

Canadian Council on Animal Care (CCAC)

1. Replacement: no alternative, justified by scientific/clinical value

2. Reduction: smallest number possible

3. Refinement: minimize discomfort/infection/illness/pain

Ethics: Truth

1. report findings truthfully

2. share credit fairly (contributors/scientists w related work)

3. share data

Reportive Definitions

definitions of words in dictionary

Lay Definitions

simplified explanation of term used in everyday convos/media

Stipulative Definitions

specialized definitions/concepts/vocab specific to academic fields

Uncritical Thinking

aka fast thinking (opinions)

→ lacks careful understanding of stimulus & reflection/reasoning

• jumps from stimulus to action

• skips reasoning

Why Humans Do Uncritical Thinking

• avoid spending effort/thinking resources to solve problems

• prefer quick/easy solutions (even if wrong)

Confirmation Bias

• tendency to seek/interpret/remember info matching pre-existing beliefs

• ignores conflicting evidence

Prestige Bias

• uncritical acceptance of claims from those w high status/prestige

• no training/education

Dogma

= low accuracy/reliability/correctness→ prohibits verification

• beliefs/claims accepted as unquestionably true

• from authority/tradition/etc

• cant question & no evidence

• rigid

Opinion

= low accuracy/reliability/correctness→ ignores verification

• beliefs/claims based on uncritical thinking

• fake news/bias/falsehoods

• questioning ignored

• confirmation bias rejects evidence

• inflexible

Knowledge

= high accuracy/reliability/correctness→ requires verification

• claims must be supported by evidence/reasoning & critical thinking

• from evidence, observation

• needs questioning

• needs reliable evidence & checks biases

• open to revision

Critical Thinking

= slow, reflexive thinking → reflect before taking action

1. seeks careful understanding of stimulus

2. need reflection

Scientific Attitude

• understanding truth more important

• humble/good ignorance

• “what does evidence show?”

• expects questions, testing, challenges

• change conclusions w new evidence

The Critical Thinking Diagram

evaluates claims using reasoning, broader frame work

(stimuli → reasoning → action)

1. U: understanding (claims/ideas before judging them)

2. A: acceptability of language (precise/consistent terms, avoid contradictions)

3. R: relevance (evidence/reasons support claim)

4. E: enough evidence (claims supported by adequate evidence)

The Scientific Method

systematically tests claims w evidence

• focuses on E of critical thinking

1. theory

2. hypothesis

3. research

4. correlational or experimental research

5. outcome (return to theory)

Aristotles Law of Non-Contradiction

A cannot be A and Not-A simultaneously

(acceptability of language)

Triangulation in Research

use multiple methods, data sources, theories or investigators to cross check/validate findings

Methodological Triangulation

use more than 1 method to gather data

Data Triangulation

using different sources of data

Investigator Triangulation

multiple researchers analyze data independently then compare results

Theoretical Triangulation

use more than 1 theoretical perspective to interpret data

Analytical Triangulation

use diff analysis techniques on same data set