Chapter 2 Psych

Three scientific attitudes

Curiosity, skepticism, and open-mindedness

Curiosity (scientific attitude)

An insatiable drive to ask "Why?"

Skepticism (scientific attitude)

Meeting every claim with "show me your evidence" and asking whether there might be another explanation

Open-mindedness (scientific attitude)

Accepting conclusions supported by facts, even when they refute your own beliefs

Scientific method

The set of principles that guide science and distinguish it from other ways of learning about the world

Why psychology adopted the scientific method later than the physical sciences

Its subject matter (e.g., personality) is less tangible and harder to measure than things like the speed of a thrown ball

Confederate

A person who appears to be a participant but actually works for the experimenter

Diffusion of responsibility

A psychological state in which each person feels less personal responsibility for intervening because others are present

Real-world update on the bystander effect (Philpot et al., 2019)

CCTV analysis showed that the larger the group, the more likely at least one person will help

Five steps in the scientific process

1) Identify a question of interest 2) Gather information and form a hypothesis 3) Test the hypothesis by conducting research 4) Analyze data, draw tentative conclusions, and report findings 5) Build a body of knowledge

Step 1 of scientific process

Identify a question of interest

Step 2 of scientific process

Gather information and form a hypothesis

Step 3 of scientific process

Test the hypothesis by conducting research

Step 4 of scientific process

Analyze data, draw tentative conclusions, and report findings

Step 5 of scientific process

Build a body of knowledge: ask further questions, conduct more research, develop and test theories

Hypothesis

A specific prediction about some phenomenon, often written as an "If-Then" statement

Theory

A set of formal statements explaining how and why certain events are related; broader than a hypothesis

Theory of social impact

Latané's broad theory combining diffusion of responsibility with other principles of group behaviour

How science is self-correcting (theories)

When research supports a theory's predictions, confidence increases; when it doesn't, the theory is modified or ultimately discarded

Hindsight reasoning

Explaining behaviour after the fact; many different, even contradictory, explanations can fit the same outcome

Main problem with hindsight reasoning

Related past events can be explained in many creative, reasonable, and contradictory ways, with no sure way to tell which is correct

Understanding through prediction and control

If we understand a behaviour's causes, we should be able to predict when it will occur and produce it by controlling those conditions

Four characteristics of a good theory

1) Organizes existing facts in one framework 2) Is testable 3) Its predictions are supported by new findings 4) Conforms to the law of parsimony

Law of parsimony

If two theories explain and predict equally well, the simpler one is preferred

Prediction vs. understanding

Prediction does not require understanding (e.g., predicting thunder after lightning), but prediction based on understanding increases knowledge and generalizes to new situations

Self-report measures

Ask people to report on their own knowledge, beliefs, feelings, experiences, or behaviour, usually via interviews or questionnaires

Social desirability bias

The tendency to respond in a socially acceptable way rather than according to how one truly feels or behaves

Ways to minimize social desirability bias

Word questions so it isn't relevant, guarantee anonymity and confidentiality, and debrief data collectors regularly

Coding systems

Systems for classifying observed behaviour into defined categories

Reliable (reliability)

Producing consistent observations; if observers using a coding system disagree repeatedly, the data are unreliable

Unobtrusive measures

Ways of recording behaviour so participants are unaware that certain responses are being measured

Archival measures

Records or documents that already exist (e.g., school records)

Personality tests

Tests assessing personality traits; either specialized self-reports or ambiguous stimuli interpreted by the respondent

Intelligence tests

Performance tasks (e.g., assembling objects, solving arithmetic problems) measuring intelligence

Neuropsychological tests

Tests that diagnose normal and abnormal brain functioning by measuring how well people perform mental and physical tasks

Physiological measures

Recordings of heart rate, blood pressure, respiration, hormones, and brain functioning to assess what people are experiencing

Main interpretive problem with physiological measures

We don't always know what a given response (e.g., increased heart rate) means or which emotion or thought it reflects

Using several types of measures in one study

Increases confidence in findings because each measure has advantages and disadvantages

Descriptive research

Research that describes how humans and animals behave, particularly in natural settings; may yield clues about cause-effect relations to test later

Three common descriptive methods

Case studies, naturalistic observation, and surveys

Case study

An in-depth analysis of an individual, a group, or an event

Advantages of case studies

Allow close study of rare phenomena, can challenge a theory or widely held belief, and are a source of new ideas and hypotheses

Limitations of case studies

Poor for determining cause and effect, findings may not generalize, and observers may be subjective or biased

Naturalistic observation

Observing behaviour as it occurs in a natural setting while trying to avoid influencing it

Validity

How well an experimental procedure actually tests what it is designed to test

Internal validity

The degree to which an experiment supports clear causal conclusions (that the independent variable caused changes in the dependent variable)

Confounding of variables

Two variables are intertwined so we can't determine which one influenced the dependent variable

Confound (confounding variable)

A variable that varies along with the independent variable and could account for the results

Simplest way to eliminate a confound

Keep the confounding variable constant across conditions (e.g., hold mood constant across music conditions)

Third-variable problem

A third variable (Z) may be mixed up with correlated variables X and Y, so we can't tell what caused what; why correlation doesn't show causation

Why placebo effects threaten internal validity

They give an alternative explanation for why responses changed after treatment

Experimenter expectancy effects

Subtle, unintentional ways researchers influence participants to respond consistently with the researcher's hypothesis

Blind (keeping researchers blind)

Not telling people who interact with participants or record responses the hypothesis or the participant's condition

Double-blind procedure

Both participant and experimenter are unaware of which experimental condition the participant is in; minimizes placebo and expectancy effects

External validity

The degree to which a study's results can be generalized to other populations, settings, and conditions

Generalizability of underlying principles

External validity usually concerns whether the general principle holds, not whether the exact numbers match

Replication

Repeating a study to see whether the original findings can be duplicated

Why replication matters

Successful replication with other participants and conditions increases confidence in a finding

Meta-analysis

A statistical procedure combining the results of different studies on the same topic, treating each study like a single participant

Why meta-analysis is valued

Many consider it the most objective way to integrate findings and assess direction and strength of relationships

Cross-cultural replication

Examining whether findings generalize across different cultures

Why failed replications can be useful

They can prompt better research and new discoveries as scientists look for factors that explain the differences

Core ethical duties of psychologists

Protect participants' welfare, avoid harm, ensure benefit outweighs risk, obtain informed consent, avoid coercion, and ensure privacy and confidentiality

Informed consent

Explaining all aspects of the procedure and ensuring it is understood; participants can withdraw without penalty

Consent for people unable to give it

Consent must be obtained from parents or guardians (e.g., children, patients with serious mental disturbances)

Incomplete disclosure (deception)

Misleading participants about the true nature of a study

When incomplete disclosure is permitted

Only when no feasible alternative exists and the scientific, educational, or applied benefits clearly outweigh the ethical costs

Debriefing

Telling participants the true purpose of the study at its end

Ethics concerns for Internet research

Protect privacy and confidentiality, obtain informed consent, and debrief participants; apply standard review protocols

Ethical rule for animal research in Canada

Animals shouldn't be subjected to pain, stress, or privation unless no alternative exists and prospective scientific or educational benefits justify it

Survey research

Obtaining information about a topic by giving questionnaires or interviews to many people

Population

All the individuals about whom a researcher is interested in drawing a conclusion

Sample

A subset of individuals drawn from the larger population of interest

Representative sample

A sample that reflects the important characteristics of the population in the same proportions

Random sampling

Procedure in which every member of the population has an equal probability of being chosen for the sample

Stratified random sampling

Dividing the population into subgroups (e.g., gender, ethnic identity) and allocating sample spots in proportion to each subgroup's share of the population, then randomly selecting within each

Main strength of survey research

With a representative sample, we can be confident the findings closely portray the whole population (e.g., about 1000 people can predict a national election)

Sample size vs. representativeness

Large samples are better than small ones, but a smaller representative sample is better than a larger unrepresentative one

Drawbacks of Internet surveys

Little control over data quality, respondents may lie or answer frivolously, and there is no method for randomly sampling Internet users

Strengths of Internet surveys

Collect data from thousands of participants quickly at virtually no cost; results are similar to paper-based surveys and less affected by missing data and socially desirable responding

Problems with crowdsourced data (e.g., Mechanical Turk)

May have flaws that challenge reliability and validity (e.g., VPNs disguise location so it's unknown who is answering)

Four major drawbacks of surveys

1) Can't show cause and effect

2) Rely on self-reports (social desirability bias, interviewer bias, inaccurate self-perception, misinterpreted questions)

3) Unrepresentative samples can lead to faulty generalizations

4) Even random samples occasionally turn out unrepresentative by chance

Habituation

Over time, people and animals adapt to and ignore an observer's presence; researchers may delay data collection until it occurs

Limitations of naturalistic observation

Doesn't permit clear causal conclusions, observer bias in interpretation is possible, and the observer's presence may disrupt behaviour

Correlational research

Research examining associations between naturally occurring events or variables by measuring them (not manipulating them)

Three components of correlational research

1) Measure one variable (X) 2) Measure a second variable (Y) 3) Statistically determine whether X and Y are related

Bidirectionality problem

Two-way causality: X may cause Y, Y may cause X, or both may influence each other

Spurious relationship

An artificial (not genuine) association between two variables; neither actually causes the other

Third-variable problem

A third variable (Z) is responsible for what looks like a relation between X and Y; Z causes both X and Y to change

Correlation coefficient

A statistic indicating the direction and strength of the relation between two variables

Benefits of correlational research

Helps show whether lab findings generalize to the real world, discovers associations to study later in the lab, and examines questions that can't be studied ethically or practically with experiments

Correlation as a basis for prediction

If two variables are correlated, knowing the score on one helps estimate the score on the other (even if the correlation isn't perfect)

Experiment

A research method for examining cause-effect relations

Three essential characteristics of an experiment

1) Researcher manipulates one or more variables 2) Researcher measures whether the manipulation influences other variables 3) Researcher attempts to control extraneous factors that might influence the outcome

Manipulating a variable

Creating different conditions to which participants are exposed (researcher controls the variable)

Extraneous factors

Other factors that might influence the outcome of an experiment; researchers try to control them

Why use a simulator or lab setting for experiments

Provides rigorous environmental control and is safer for ethical reasons

Logic of an experiment

Start with equivalent groups, treat them equally except for the variable of interest, manipulate that variable, measure responses, and compare groups; differences are plausibly caused by the manipulated variable

General vs. operational level of variables

General level is the broad concept (e.g., safe driving); operational level is how it's specifically manipulated or measured (e.g., braking reaction time)

Experimental group

The group that receives a treatment or an active level of the independent variable