Unit O Notes: The Scientific Attitude, Research Methods, and Statistical Reasoning (Modules 0.1–0.6)
Module 0.1 The Scientific Attitude, Critical Thinking, and Developing Arguments
AP® Exam four science practices to develop throughout AP Psychology: Concept Application, Research Methods & Design, Data Interpretation, and Argu-mentation. Look for AP Science Practice features throughout units.
Learning targets:
0.1-1 Explain how psychology is a science.
0.1-2 Describe the three key elements of the scientific attitude and how they support scientific inquiry.
0.1-3 Explain how critical thinking feeds a scientific attitude, and smarter thinking for everyday life.
AP® Science Practice: Research emphasis; terminology learned in context; Research methods and design important for AP exam.
The Amazing Randi (James Randi, 1928–2020) tested and debunked supposed psychic phenomena using evidence-based approaches.
Key elements of the scientific attitude::
Skeptical but not cynical; open-minded but not gullible.
Core idea: The scientific attitude helps sift reality from fantasy; Right ideas stick around, wrong ideas get discarded.
Practical mottoes and examples:
The rat is always right: humility in psychology—humility means we accept when data contradict our ideas.
Examples of claims to test: extrasensory mind-reading; whether parental behaviors determine children’s sexual orientation (Module 3.3).
Developing Arguments section outline:
By addressing critical thinking questions, you develop arguments based on scientifically derived evidence; essential for AP exam.
Critical thinking and the scientific attitude:
Curiosity: Does it work? Can predictions be confirmed?
Skepticism: What do you mean? How do you know?
Humility: Be open to surprises; acknowledge that observed behaviors may not match own beliefs.
Visual cue: The “Internet quotes” diagram contrasts truth as facts vs misattributed or misleading quotes.
Examine the Concept and Check Your Understanding prompts encourage you to articulate what critical thinking involves and to defend psychology as a science.
Module 0.2 The Need for Psychological Science
Learning targets:
0.2-1 Explain how cognitive biases (hindsight bias, overconfidence, tendency to perceive order in random events) illustrate why science-based answers are more valid than common-sense answers.
Common-sense caution: intuition can be right (e.g., belonging needs relate to happiness) but many “commonsense truths” are overturned by research (e.g., love and happiness; brain concepts; dreams predicting the future; 10% brain myth).
Myth vs fact example: Cold-weather myth vs actual facts about colds/viruses, humidity, and hypothermia risks.
Hindsight bias (I-knew-it-all-along): the tendency to believe, after learning an outcome, that one would have foreseen it.
Demonstrations: two groups told opposing findings about romance and separation; after explanation, both groups find the finding unsurprising, illustrating hindsight bias.
Deepwater Horizon example: with 20/20 hindsight, warnings were recognized as obvious mistakes.
Overconfidence: tendency to overestimate one's knowledge; evidenced by quick answers to trivial questions or predictions about social events being correct less than half the time.
Tetlock (1998, 2005) large dataset of expert predictions showed experts were right <40% of the time; superforecasters excel by gathering facts and balancing arguments.
Perceiving order in random events: humans see patterns where none exist; randomness often looks nonrandom (e.g., sequences of computer-generated numbers, coin flips).
Why cognitive biases persist: patterns-seeking helps us make sense of a random world and reduce stress, but it risks mistaking luck for causation.
AP® Science Practice: Examine the Concept, Check Your Understanding, Apply the Concept prompts emphasize distinguishing fact from intuition; defending psychology as a science.
The Need for Psychological Science takeaway: critical thinking helps discard myths and seek truth; psychology rests on scientific inquiry.
Module 0.3 The Scientific Method
Learning targets:
0.3-1 Describe how theories advance psychological science.
0.3-2 Explain how psychologists use case studies, naturalistic observations, and surveys to observe and describe behavior, and why random sampling is important.
Important preface: The scientific method is a set of principles/procedures, not a finite list of facts. You should understand how psychology is done, not only what it has discovered.
Core vocabulary:
peer reviewers: experts who evaluate a study's theory, originality, and accuracy before publication.
theory: an explanation using integrated principles that organize observations, predict behaviors/events; underpins hypotheses.
hypothesis: a testable prediction implied by a theory; falsifiable.
falsifiable: a hypothesis can be disproven by observation or experiment.
The Scientific Method overview:
Self-correcting process of evaluating ideas via observation/analysis; theories are supported when data align with predictions; if not, revise or reject.
Peer review introduces safeguards for theory quality and reproducibility.
Constructing Theories:
Theories organize observations and generate hypotheses; e.g., sleep improves memory; more sleep often corresponds with better memory performance.
Hypotheses specify what results would support or disconfirm the theory; falsifiability is key.
Theories should organize observations and imply testable predictions for practical applications.
Observations may be biased by existing theories; therapists/psychologists must use operational definitions to define variables (e.g., what counts as sleep deprivation; aggression measures via pins in a doll; etc.).
Replication strengthens confidence when different participants/situations yield similar results.
Operational definitions: precise statements of procedures used to measure/define variables; enable replication.
Replication: repeating a study with different participants to see if the same pattern holds.
Non-experimental methods: Case studies, naturalistic observations, and surveys:
Case study: in-depth analysis of one person or group; useful for revealing universal principles but may mislead due to atypical cases.
Freud and Little Hans example: early case study supporting sexual theory; contemporary psychology questions Freud’s sexual theory but acknowledges unconscious processes.
Naturalistic observation: behavior observed in natural environments without manipulation; increasingly enabled by digital tech/big data (e.g., GPS data during Covid, 504 million tweets mood tracking).
Positive correlation: both variables rise or fall together. Negative correlation: one rises as the other falls.
Perfect correlations are rare in real life; interpret with caution.
Practical example (Table 0.4-1): Fear and disgust across 24 animals; example data show a positive correlation r = +0.72 with substantial scatter, illustrating that correlation does not imply perfect prediction.
AP® Exam Tip: Positive/negative indicates direction, not value judgment about desirability; correlation does not imply causation.
Directionality problem: correlation cannot determine which variable causes the other; a third variable may be involved.
Illusory correlations and regression toward the mean:
Illusory correlation: perceiving a relationship where none exists due to selective attention to confirm beliefs.
Regression toward the mean: extreme scores tend to move toward the average on subsequent measurements; explains why apparently causative explanations (hot streaks, curses) may be mistaken.
Experimental method to establish causation:
Random assignment and manipulation of an independent variable to observe effects on a dependent variable, while controlling confounds.
Example: Allcott et al. (2020) deactivated Facebook accounts for 4 weeks; randomized assignment; observed changes in depression/happiness; demonstrates causal inference beyond correlation.
Placebo and bias controls:
Single-blind: participants unaware of treatment status; controls for social desirability bias.
Double-blind: neither participants nor researchers know who receives treatment; reduces experimenter bias and placebo effects.
Independent variable (IV): the manipulated factor; Dependent variable (DV): the measured outcome.
Confounding variables: other variables that could influence results; random assignment helps balance these across conditions.
Operational definitions: precise definitions of how IV and DV are measured/manipulated to enable replication.
Experimental design example: Facebook deactivation study with IV = deactivated Facebook account (yes/no); DV = depression score after 4 weeks; randomized groups; potential placebo-like effects discussed.
Experimental validity and measurement:
Validity: whether the experiment tests what it is supposed to test.
The rental housing study example: Independent variable = perceived ethnicity of the sender’s name; Dependent variable = rate of positive invitations; demonstrates manipulation and measurement clarity.
AP® Science Practice: Examine the Concepts on random assignment, double-blind, placebo, and the difference between random assignment and random sampling.
Control for confounds through random assignment and careful operationalization.
Use of single/double-blind procedures to curb biases and placebo effects.
Module 0.4 Review
Summary of correlations:
Correlation describes relationships; it does not prove causation; reliable and significant differences are needed for generalization.
Illusory correlations and regression toward the mean are common biases that can mislead interpretation.
Experimental design essentials:
Random assignment to experimental vs control groups.
Manipulation of IV, measurement of DV, control of confounds.
Validity and reliability considerations.
The 0- execution example: rental housing independent/dependent variables clarified.
Module 0.5 Research Design and Ethics in Psychology
Learning targets:
0.5-1 Explain how to determine which research design to use.
0.5-2 Explain the value of simplified laboratory conditions for illuminating everyday life.
0.5-3 Explain why psychologists study animals and outline ethical guidelines safeguarding human and animal welfare.
0.5-4 Describe how psychologists' values influence what they study and how they apply results.
Research design overview (Table 0.5-1):
Non-experimental methods: Case studies, naturalistic observations, surveys; while correlational studies detect relationships, they cannot specify cause-and-effect.
Experimental method: Tests cause-and-effect; involves manipulation of IV, random assignment, and measurement of DV; potential limitations include generalizability and ethical constraints.
Key points about measurement and design:
Quantitative vs qualitative methods; operational definitions; replication and generalizability.
Diversity, equity, and inclusion considerations in study design.
Predicting everyday behavior: simplified laboratory conditions test theoretical principles rather than exact real-world replication (Mook, 1983).
The ethics of research:
Animal research: animals used to understand human biology and behavior; potential benefits include medical advances; debates around animal welfare.
Human participants: focus on minimizing harm, informed consent, debriefing, confidentiality; IRBs (Institutional Review Boards) oversee ethical compliance; APA and BPS guidelines.
Informed consent before participation; debriefing after; risks and benefits disclosed; deception allowed only if essential and justifiable.
Animal and human welfare discussions:
Ethical standards emphasize humane care, minimizing discomfort, and ensuring welfare; guidelines vary by region (APA, BPS, European Parliament).
Real-world benefits of animal research include improved animal care and enrichment and greater empathy for humans/animals.
Values in psychology:
Researchers’ values influence topic choices, theories, interpretations, labeling, and how results are used in practice.
Language and labels reflect attitudes (e.g., rigidity vs consistency, fanaticism vs faith).
Psychology’s power to influence society can be used for good or ill; science aims to enlighten and address societal problems (inequality, climate change, prejudice).
AP® Practice prompts and cross-cutting questions:
Distinguish between quantitative and qualitative methods.
Explain informed consent and debriefing and their importance.
Reflect on how values influence research questions and interpretations.
Module 0.6 Statistical Reasoning in Everyday Life
Learning targets:
0.6-1 Describe descriptive statistics.
0.6-2 Describe measures of central tendency (mean, median, mode) and percentile rank.
0.6-3 Explain the relative usefulness of the two measures of variation.
0.6-4 Describe inferential statistics.
0.6-5 Explain how we determine whether an observed difference can be generalized to other populations.
Why statistics matter:
Stats help see what the unaided eye might miss; avoid overreliance on big, round numbers; prevent misinformation from undocumented numbers.
Covid vaccine example: “95% effective” meaning in primary trial terms; absolute vs relative risk clarification (example breakdown with 162 vs 8 cases among 44k participants).
Descriptive statistics:
Histogram: simple bar graph showing a distribution; careful with axis scales; two histograms can mislead by scale choices.
Measures of central tendency:
Mode: most frequent score(s).
Mean: arithmetic average: ar{x} = rac{1}{n} ext{(sum of all } x_i)
Median: middle score when data are ordered.
Percentile rank: percentage of scores below a given score.
Skewed distributions: mean can be distorted by extreme scores; example with income distribution (Elon Musk’s wealth skews mean upward while median stays representative of typical income).
The bottom line: note which measure is reported; consider the impact of outliers.
Measures of variation:
Range: crude measure of spread (highest minus lowest).
Standard deviation: more informative; captures average deviation from the mean.
Bell-shaped (normal) distribution: most scores cluster near the mean; 68% within one SD; 95% within two SDs.
Standard deviation formula (note the textbook’s form): s=n−1∑<em>i=1n(x</em>i−xˉ)2
Normal curve example: Wechsler IQ scale with mean 100 and SD 15; approx. 68% within ±1 SD (i.e., 85–115) and 95% within ±2 SD (i.e., 70–130).
Inferential statistics:
Inferential statistics generalize from a sample to a population; distinguish descriptive vs inferential.
Key questions involve reliability (consistency) and statistical significance (probability results reflect a real effect rather than chance).
Confidence intervals: range likely to contain the population mean; precision improves with larger samples and lower variability.
Statistical significance vs practical significance:
A result can be statistically significant but have little practical importance, especially with very large samples.
The p-value threshold commonly used is p < 0.05 for significance (though context matters).
The concept of effect size (strength of relationship) is separate from p-values; larger effects are more practically meaningful.
Principles for inferring population differences from samples (0.6-5):
1) Representative samples are better than biased samples.
2) Bigger samples are better than smaller ones.
3) More estimates (meta-analysis) are better than fewer.
Key takeaways:
Descriptive statistics summarize data; inferential statistics determine generalizability to populations.
Always consider sampling methods, sample size, and the presence of outliers when interpreting results.
Use critical thinking to interpret statistics encountered in media and research reports.
Practical exercise prompts include calculating mean/median/mode for a data set, interpreting percentile ranks, and evaluating graphs with scale considerations.
Quick reference: Key formulas and concepts
Correlation coefficient: r∈[−1,+1]; direction: positive or negative; strength inferred from scatter.
Mean: xˉ=n1∑<em>i=1nx</em>i
Standard deviation (sample): s=n−1∑<em>i=1n(x</em>i−xˉ)2
Normal curve properties: ~68% within one SD; ~95% within two SDs.
Statistical significance threshold: p < 0.05 (commonly used).
Confidence interval (general form): xˉ±z∗ns (for known z or t-based equivalents).
Independent variable (IV) = manipulated factor; Dependent variable (DV) = measured outcome; Confounding variable = other factors that may affect results.
Random sampling vs random assignment:
Random sampling: representative population sample for generalization (surveys).
Random assignment: equalizes groups in experiments to infer causation.
Descriptive vs inferential statistics:
Descriptive: summarizes data from a sample.
Inferential: generalizes findings to a population; uses p-values, confidence intervals, etc.