AP Psychology Comprehensive Course Notes: Research Methods, Cognition, and Statistics

Revised AP Psychology Course Framework and Exam Structure

• Definition of Psychology: The scientific study of mind and behavior, investigating both underlying cognitive mental processes and observable actions.

• Curriculum Structure: The revised 2024 course framework is organized into five foundational units of instruction:

  • Unit 1: Biological Bases of Behavior: Covers neural transmission, brain anatomy, nervous system structure, neuroplasticity, genetics, and evolutionary influences on human behavior (weighting: $15\text{--}25\%$).

  • Unit 2: Cognition: Explores memory models, problem-solving, cognitive biases, language acquisition, intelligence testing, and information processing (weighting: $15\text{--}25\%$).

  • Unit 3: Development and Learning: Focuses on lifespan physical, cognitive, and moral development, along with behavioral learning paradigms such as classical and operant conditioning (weighting: $15\text{--}25\%$).

  • Unit 4: Social Psychology and Personality: Analyzes group behavior, social perception, attribution theory, conformity, obedience, and structural personality theories (weighting: $15\text{--}25\%$).

  • Unit 5: Mental and Physical Health: Examines psychological disorders, etiology, therapeutic interventions, health psychology, and the physiological consequences of chronic stress (weighting: $15\text{--}25\%$).

• AP Exam Administration Details:

  • Total Duration: 2 hours and 40 minutes ($160\,\text{minutes}$).

  • Section I (Multiple-Choice Questions): Contains 75 discrete multiple-choice items administered over 90 minutes. Accounts for $66.7\%$ of the composite score.

  • Section II (Free-Response Questions): Contains 2 essay-style questions administered over 70 minutes. Accounts for $33.3\%$ of the composite score, divided equally between the two prompts ($16.65\%$ each).

• Free-Response Prompts Specification:

  • Question 1: Article Analysis Question (AAQ): Requires reading an excerpt of psychological research, critically evaluating methodology, identifying experimental components, evaluating validity, and applying concepts.

  • 
    

  • Question 2: Evidence-Based Question (EBQ): Presents three separate research article summaries on an interconnected topic; test-takers formulate a scientific claim and substantiate it using methodological evidence and psychological principles.

Cognitive Architecture: Dual Processing, System 1, and System 2

• Dual Processing Theory: The human brain processes vast amounts of sensory inputs on two parallel tracks—a conscious, deliberate track and an unconscious, automatic track occurring simultaneously without intentional awareness.

  • Real-World Example: When observing high school students playing a football game, conscious thoughts track the score, player strategies, field position, and elapsed time. Concurrently, the brain unconsciously processes each player's movement speed, distance, relative size, jersey colors, and field textures.

  • Unconsciously gathered information only surfaces into conscious awareness when intentionally retrieved or when explicit questions are asked afterward (such as recalling team jersey colors).

• System 1 Cognitive Processing:

  • Operates automatically, unconsciously, swiftly, and with minimal to no mental effort.

  • Heavily error-prone, relying on intuitive reflexes, affective reactions, and heuristic shortcuts.

  • Governs routine everyday choices, such as instinctively grabbing a familiar snack off a supermarket shelf or ducking to avoid a sudden projectile.

• System 2 Cognitive Processing:

  • Operates slowly, consciously, deliberately, and with effortful cognitive strain.

  • Analytical, methodical, and rule-governed, yielding higher accuracy and reliability.

  • Governs complex, novel tasks, such as calculating mathematical equations, parallel parking, or weighing comparative pros and cons when declaring a university major.

Heuristics and Problem-Solving Strategies

• Heuristics: Practical mental shortcuts and rules of thumb deployed by System 1 to reduce the cognitive effort required to reach decisions, make judgments, and solve problems rapidly.

  • Heuristics do not guarantee an optimal or correct solution, but they provide solutions that are "good enough" for immediate evolutionary survival and routine choices.

• Algorithms versus Heuristics:

  • Algorithm: A methodical, step-by-step procedure or exhaustive logical sequence that guarantees the correct solution if executed accurately.

    • Scenario: Navigating to a friend's house by systematically following GPS turn-by-turn routing with street names guarantees exact arrival.

  • Heuristic: A quick rule of thumb that skips exhaustive testing.

    • Scenario: Driving in the general geographic direction of the friend's neighborhood and stopping to ask bystanders for directions once near.

• Educated Guess: A heuristic judgment informed by specialized background knowledge and past experiential schemas, producing a high probability of accuracy without formal proof (e.g., an experienced poker player inferring an opponent is bluffing based on micro-gestures and betting cadences).

• Availability Heuristic: Estimating the likelihood of an event based on how mentally available or vivid instances of it are in memory.

  • Mechanism: Events that are dramatic, sensationalized, or emotionally charged (such as airplane crashes, terrorist strikes, or shark attacks) are recalled effortlessly.

  • Consequence: Causes individuals to severely overestimate the statistical risk of highly publicized anomalies while underestimating statistically common hazards.

  • Empirical Comparisons of Perceived vs. Actual Mortality Risk (Annual US Deaths per $100{,}000$):

    • All Accidents ($37.7$) vs. Strokes ($51.1$)

    • Suicide ($10.9$) vs. Blood Poisoning / Septicemia ($11.2$)

    • Homicide ($5.9$) vs. Diabetes ($24.5$)

    • Motor Vehicle Accidents ($15.3$) vs. Colorectal Cancer ($17.8$)

    • Drowning ($1.3$) vs. Leukemia ($7.1$)

  • Comparative Population Estimates (Knowledge Check):

    • Saudi Arabia ($28\,\text{million}$) vs. Morocco ($34\,\text{million}$)

    • Australia ($21\,\text{million}$) vs. Myanmar ($47\,\text{million}$)

    • South Africa ($48\,\text{million}$) vs. Vietnam ($86\,\text{million}$)

    • Libya ($6\,\text{million}$) vs. Sri Lanka ($20\,\text{million}$)

    • Iraq ($28\,\text{million}$) vs. Tanzania ($38\,\text{million}$)

• Representativeness Heuristic: Judging the probability of an event or categorizing an individual by comparing how well they match an internal prototype, stereotype, or idealized image, while ignoring relevant base-rate statistical facts.

  • Scenario: Looking at an individual with full-body tattoos and unkempt attire and concluding they are an outlaw gang member, or choosing an attractive young female as a child babysitter over an awkward male based purely on maternal prototypes.

  • Consequences: Fosters prejudice, discriminatory behavior, close-minded stereotyping, and persistent failure to consider population base rates.

Cognitive Biases and Critical Thinking Fallacies

• Cognitive Bias Definition: Systematic deviations from normative rationality, logic, or empirical truth that manifest during human judgment, perception, and decision-making.

• Hindsight Bias: The "I-knew-it-all-along" phenomenon; the tendency for people to perceive past events as having been completely predictable after the outcome is already known.

  • Mechanism: Once an outcome occurs (such as an election outcome or sports championship), cognitive schemas rewrite memory to minimize prior uncertainty and inflate perceived foresight.

• Perceiving Order in Random Events: The psychological drive to identify coherent patterns, causal links, or meaning within entirely random, meaningless statistical sequences.

  • Mathematical Law of Truly Large Numbers: Given a sufficiently large sample size, any extraordinarily improbable or outrageous event is mathematically guaranteed to occur purely by chance.

  • Case Example: "Paul the Oracle Octopus," an animal that appeared to predict the outcome of international soccer matches during the 2010 FIFA World Cup, was merely demonstrating random probability amplified by selective media reporting.

• Common Sense Fallacy in Psychology: The false belief that psychological discoveries are merely obvious truisms. In 2013 studies, several seemingly "intuitive" findings demonstrated that once a finding is asserted (e.g., that racists are close-minded, men who cheat have high sex drives, or consumers buy cheaper produce), individuals instantly declare it self-evident through hindsight bias.

• Confirmation Bias: The cognitive tendency to selectively search for, interpret, favor, and recall information that validates preexisting beliefs while discounting, ignoring, or explaining away disconfirming empirical evidence.

  • Wason 4-Card Selection Task (Conditional Logic):

    • Rule: "If a card has a vowel on one side, then it must have an even number on the other side."

    • Cards presented: [A], [Q], [4], [7].

    • Common Error: Most people flip [A] and [4], looking to confirm the hypothesis.

    • Logical Solution: One must turn over card [A] (a vowel to confirm an even number on reverse) and card [7] (an odd number to check if a vowel is on reverse, which would falsify the rule). Flipping [4] proves nothing because the conditional rule does not state that all even numbers must have vowels.

• Self-Serving Bias: A readiness to perceive oneself favorably by attributing personal successes and positive outcomes to internal dispositional factors (ability, intellect, hard work), while blaming personal failures and negative outcomes on external situational factors (bad luck, unfair referees, biased tests).

  • Example: Cristiano Ronaldo crediting his personal physical conditioning and skill when winning a soccer match, but blaming biased referees or poor pitch conditions when suffering a defeat.

• Overconfidence Bias: The tendency to consistently overestimate the accuracy of one's beliefs, judgments, and abilities, being routinely "more confident than correct."

  • Anagram Experiment: When subjects are presented with anagram puzzles (e.g., solving GRABE to BARGE, ETYRN to ENTRY, WREAT to WATER), they predict they will solve each in an average of 10 seconds, but actually require roughly 180 seconds.

  • Meta-Overconfidence: People routinely demonstrate overconfidence regarding their own objectivity, firmly believing that they are less susceptible to cognitive biases than the general public.

• Halo Effect: A cognitive bias where an observer's overall positive impression of an individual or entity in one domain (e.g., physical attractiveness, charisma, polished web design) disproportionately influences their evaluation of that entity's other distinct traits (e.g., morality, competence, intelligence, product utility).

Scientific Methodology: Basic vs. Applied Research

• Basic Research (Pure / Fundamental Science): Investigation driven purely by intellectual curiosity, theoretical inquiry, and the expansion of foundational scientific knowledge without regard for immediate practical utility.

  • Examples: Investigating how short-term working memory capacity declines across varying age cohorts; mapping neurochemical receptor binding profiles in rats; exploring whether COVID-19 is prevalent in Canada.

• Applied Research: Scientific inquiry explicitly designed to solve practical, real-world problems and implement direct interventions or solutions.

  • Examples: Evaluating whether ambient background music improves employee productivity in corporate offices; testing whether a pharmaceutical compound successfully treats symptoms of COVID-19.

• Core Scientific Frameworks:

  • Theory: A broad, comprehensive, well-substantiated explanation that integrates observations, synthesizes empirical findings, and generates testable hypotheses (e.g., "Participating actively in classroom discussions leads to superior academic performance compared to passive listening.").

  • Hypothesis: A specific, testable prediction derived logically from a theory that outlines an operational relationship between two or more variables (e.g., "Students instructed to ask three questions per class period will score higher on final exams than students instructed to remain silent.").

  • Scientific Law: A concise verbal or mathematical statement of an observed regular pattern in nature that appears universally true under specified conditions (e.g., Newton's Law of Universal Gravitation).

  • Falsifiability: The absolute scientific requirement that a hypothesis or theory must be capable of being disproven through empirical observation or experimentation.

    • Falsifiable Statement: "All swans are white" (disproven by the discovery of a single black swan).

    • Non-Falsifiable Statement: "A transcendent deity exists" (cannot be tested or invalidated by any empirical observation).

The Scientific Method and Experimental Design

• Sequential Stages of the Scientific Method:

  1. Question: Formulating an inquiry based on empirical observation (e.g., observing a household toaster that fails to operate).

  2. Research: Reviewing background literature and previous discoveries.

  3. Hypothesis: Deriving an operational, falsifiable predictive claim (e.g., "The toaster fails to heat because the electrical power cord is severed.").

  4. Experiment: Systematically executing a controlled empirical test manipulating only the variable under inspection.

  5. Data Collection & Analysis: Compiling quantitative and qualitative results.

  6. Conclusion & Iteration: Evaluating whether empirical evidence supports or refutes the hypothesis; refining theories and repeating the loop.

• Variables in Experimental Research:

  • Independent Variable (IV): The variable systematically manipulated, altered, or controlled by the experimenter to assess its causal impact.

  • Dependent Variable (DV): The measured outcome variable expected to change in response to the manipulation of the independent variable.

  • Memory Heuristic: The dependent variable depends on the independent variable.

  • Drug Trial Example: The chemical compound administered (or placebo) is the independent variable; the measured reduction in clinical symptoms is the dependent variable.

• Operational Definitions: Clear, precise, explicit descriptions of the exact procedures, measurements, and variables employed in a research protocol.

  • Purpose: Essential to ensure that external researchers can replicate the experiment precisely to verify reliability.

  • Examples:

    • Operationalizing Happiness: Quantified as the total number of distinct facial smiles exhibited during a standardized five-minute social interaction.

    • Operationalizing Aggression: Quantified as the cumulative count of physical hits, kicks, or shoves observed during a 30-minute supervised recess period.

• Confounding Variables: Extraneous variables that are not controlled for by the researcher, which inadvertently covary with the independent variable and offer an alternative explanation for changes in the dependent variable.

  • Example: If a researcher investigates whether cigarette smoking causes cardiovascular disease, confounding variables include participant diet, baseline genetic predispositions, exercise routines, stress levels, and alcohol consumption.

Experimental Controls, Threats, and Corrections

• Placebo Effect: Measurable psychological or physiological improvements following the administration of an inert, sham substance (such as a sugar pill) caused entirely by the participant's positive expectations.

• Nocebo Effect: Physiological or psychological distress and perceived adverse side effects (such as nausea, headaches, or dizziness) triggered by an inert substance due to the participant's negative expectations.

• Order Effects: Confounding influences that arise when the sequence in which experimental conditions, treatments, or questions are presented alters participant responses (caused by fatigue, boredom, or practice).

• Counterbalancing: An experimental control method designed to eliminate order effects by presenting conditions to participants in varying sequences.

  • Application: In a cognitive experiment testing retention of List A versus List B, half of the subjects study List A followed by List B, while the remaining half study List B followed by List A, balancing out order bias.

• Experimental Blind Procedures:

  • Single-Blind Study: Participants remain completely unaware of whether they are assigned to the experimental treatment group or the inert control group; prevents participant expectancy effects and demand characteristics.

  • Double-Blind Study: Neither the participants nor the research assistants administering treatments and collecting outcome data know group assignments; neutralizes both participant expectations and experimenter bias.

• Experimental Taxonomies:

  • Controlled Laboratory Experiments: High internal validity through strict manipulation of variables; suffers from low ecological validity (artificial setting).

  • Field Experiments: Manipulations conducted inside authentic, real-world environments; high realism and ecological validity, but reduced control over extraneous variables.

  • Natural Experiments: Researchers measure outcomes where the independent variable varies due to naturally occurring events (disasters, policy changes) without direct researcher manipulation; cannot definitively demonstrate causality.

Sampling and Assignment Methodologies

• Target Population: The entire, complete category of individuals from which researchers wish to draw conclusions (e.g., all high school students in Canada).

• Sample: A selected subset drawn from the target population that directly participates in the study.

• Random Sampling / Random Selection: Every individual in the target population possesses an equal, nonzero probability of being chosen for participation.

  • Core Benefit: Ensures the sample is representative of the broader population, enabling accurate external validity and generalizability.

• Random Assignment: The procedural assignment of already-recruited participants to either the experimental group or the control group purely by chance.

  • Core Benefit: Equalizes participant characteristics across groups, minimizes individual differences, and eliminates systematic confounding variables, allowing researchers to infer true causality ($A \rightarrow B$).

• Distinction Between Selection and Assignment:

  • Random Selection relates to how participants are drawn from the population (affects generalizability).

  • Random Assignment relates to how participants are distributed into groups (affects internal validity and causal inference).

Inferential Statistics and Hypothesis Testing

• Hypothesis Formulation:

  • Null Hypothesis ($H_0$): Asserts that there is no relationship, effect, or significant difference between the tested variables; any observed disparity is due to chance.

  • Alternative Hypothesis ($H_1$): Asserts that the independent variable exerts a true, measurable effect on the dependent variable.

• Statistical Significance and the $p$-Value:

  • Definition: The $p$-value represents the probability that the observed experimental differences occurred purely by random chance, assuming the null hypothesis is true.

  • Threshold: By psychological convention, an alpha level of $p < 0.05$ is required for statistical significance. If $p < 0.05$, researchers reject the null hypothesis ($H_0$).

  • Example: If a clinical therapy group experiences lower anxiety than a control group with a calculated $p = 0.03$, the difference is statistically significant ($p < 0.05$), indicating an effect.

• Effect Size: The quantitative magnitude or strength of the relationship between variables, independent of sample size.

  • Significance vs. Size: A finding can be statistically significant ($p < 0.05$) with an enormous sample, yet possess a miniscule effect size with negligible practical significance.

  • Empirical Example: A study identifying that firstborn children possess higher intelligence than second-born siblings was statistically significant, yet the actual effect size was tiny—a difference of only $1.5$ IQ points.

• Statistical Errors:

  • Type I Error (False Positive): Incorrectly rejecting a true null hypothesis (detecting an effect that does not exist in reality).

  • Type II Error (False Negative): Failing to reject a false null hypothesis (missing an actual effect that exists in reality).

• Reliability versus Validity:

  • Reliability: The consistency, stability, and repeatability of test results over time across multiple administrations.

    • Enhanced by representative samples, large sample sizes, and quantitative meta-analyses.

  • Validity: The accuracy of an instrument; the extent to which an assessment tool measures precisely what it is designed to measure.

Descriptive and Non-Experimental Research Designs

• Descriptive Research Purpose: Systematically observes, catalogs, and records behaviors without experimental manipulation; cannot establish cause-and-effect relationships.

• Case Studies: An in-depth, intensive clinical or behavioral investigation of a single subject or small group.

  • Advantage: Provides deep descriptive data regarding rare psychological conditions or trauma that cannot be ethically induced in laboratories.

  • Disadvantage: Individual cases can be idiosyncratic outliers that fail to generalize to the broader human population.

  • Classic Historical Case Studies:

    • Phineas Gage (1848): A railroad foreman whose frontal lobes were pierced by an iron tamping rod; his sudden shift from polite and responsible to impulsive and profane demonstrated brain localization and the frontal lobe's role in personality and executive control.

    • Kitty Genovese (1964): Her murder in New York, allegedly witnessed by dozens of neighbors who failed to intervene, catalyzed social psychological investigations into diffusion of responsibility and the Bystander Effect.

    • Sek Yi (2003): A 122-year-old Cambodian man who attributed his extreme longevity to daily smoking and constant prayer; demonstrates why individual case studies cannot be used to infer causal health truths.

• Longitudinal vs. Cross-Sectional Studies:

  • Longitudinal Study: Tracks and assesses the exact same cohort of participants over an extended timeline.

    • Pros: Controls for cohort effects; provides accurate tracking of developmental trajectories.

    • Cons: Extremely expensive, lengthy, and plagued by participant attrition (dropouts).

  • Cross-Sectional Study: Simultaneously evaluates and compares participants across different age brackets at a single point in time.

    • Pros: Rapid, cost-effective, immediate data collection.

    • Cons: Susceptible to cohort effects (generational differences in life experiences that confound the results).

Survey Methods and Observational Research

• Surveys: Standardized self-report questionnaires distributed to samples to infer population beliefs, attitudes, and behaviors.

• Survey Method Biases:

  • False Consensus Effect: The cognitive tendency to overestimate the extent to which other people share our personal attitudes, behaviors, and beliefs (frequently seen in extreme ideological groups).

  • Social Desirability Bias: The inclination of respondents to provide socially acceptable, sanitized answers rather than truthful reports, fearing judgment.

  • Self-Report Bias: Systematic inaccuracy in data arising because human memories are flawed, biased, or intentionally distorted.

  • Sampling Bias / Non-Response Bias: Systematic errors where individuals who choose to respond differ significantly from non-respondents.

  • Wording Effects: Subtle alterations in phrasing, syntax, or interview contexts that dramatically reshape survey outcomes:

    • $77\%$ of respondents expressed deep interest in "plants and trees," whereas only $39\%$ reported an interest in "botany."

    • When interviewed by a male researcher, $64\%$ of women agreed that abortion was a private matter; when interviewed by a female researcher, $84\%$ agreed.

    • $91\%$ of citizens favored "waiting periods and background checks," while only $37\%$ supported a "national gun-registration program costing 20% of crime control dollars."

• Naturalistic Observation: Unobtrusively observing and recording animal or human behavior in natural environments without intervention.

  • Hawthorne Effect: The tendency for subjects to alter their behavior simply because they know they are being observed.

  • Correction: Observers must remain unobtrusive, disguised, or hidden to preserve natural behavior.

Research Ethics and Regulatory Governance

• Historical Origins: Modern ethical guidelines arose following World War II, largely in response to non-consensual Nazi medical experiments uncovered during the Nuremberg Trials (1947), which led to the Nuremberg Code.

• Institutional Review Board (IRB): A formal ethics committee operating within academic and research facilities charged with reviewing, modifying, approving, or rejecting research proposals to safeguard human participants' rights and physical/psychological welfare.

• Mandatory Ethical Tenets for Human Research:

  1. Informed Consent: Participants must voluntarily agree to participate after being clearly informed of the study's general procedures, anticipated risks, and their right to withdraw at any time without penalty.

  2. Deception Limits: Deception may only be employed if the study holds substantial scientific value and no viable alternative exists; it must never disguise physical risks or severe distress.

  3. Absence of Coercion: Participation must be entirely voluntary; excessive incentives or academic threats are prohibited.

  4. Anonymity and Confidentiality: All participant identity markers and personal data must be kept confidential and securely stored.

  5. Risk Minimization: Researchers must protect participants from unnecessary physical trauma, pain, humiliation, and severe psychological distress.

  6. Debriefing: After the conclusion of data collection, researchers must fully disclose the experiment's true purpose, reveal any necessary deception, and address any negative emotional aftereffects.

• Historical Studies with Ethical Concerns:

  • Stanley Milgram's Obedience Studies (1961): Deceived participants into believing they were delivering real, escalating electric shocks (up to 450 volts) to an agonizing learner; raised issues regarding extreme psychological distress and limits of deception.

  • Philip Zimbardo's Stanford County Prison Experiment (1971): Assigned college men as guards or prisoners in a mock facility; aborted after six days due to severe emotional breakdowns, psychological degradation, and physical abuse.

  • Albert Bandura's Bobo Doll Experiment (1961): Exposed nursery children to adult models hitting, punching, and verbally abusing an inflated clown doll; raised concerns over deliberately inducing aggression in young children without parental consent.

  • John B. Watson & Rosalie Rayner's Little Albert Experiment (1920): Classical conditioning of phobic terror in an 11-month-old infant toward white rats and fuzzy objects using loud steel strikes; lacked informed consent, produced severe distress, and provided no deconditioning.

Descriptive Statistics: Central Tendency and Variability

• Descriptive Statistics: Mathematical tools designed to organize, summarize, synthesize, and describe the properties of a specific collected dataset.

• Measures of Central Tendency: Single representative values indicating the center of a distribution:

  • Mean ($\mu$ or $\bar{X}$): The arithmetic average calculated by summing all individual values and dividing by the total count:         $\bar{X} = \frac{\sum X}{N}$         Vulnerability: Highly sensitive to extreme outliers; distorts the true center in skewed distributions.

  • Median: The physical middle score in an ordered rank distribution ($50^{\text{th}}$ percentile). Resistant to extreme outliers; represents the best measure of central tendency for skewed datasets.

  • Mode: The score that appears with the greatest frequency. Useful for categorical data or bimodal distributions.

• Measures of Variation (Dispersion): Describe the spread or dispersion of data points:

  • Range: The gross mathematical difference between the single highest and single lowest score:         $\text{Range} = X_{\text{max}} - X_{\text{min}}$

  • Standard Deviation ($\sigma$): The average measure of how much individual scores vary around the mean:         $\sigma = \sqrt{\frac{\sum (X - \mu)^2}{N}}$         Interpretation: A low standard deviation produces a tall, steep, narrow curve indicating tightly clustered scores; a high standard deviation produces a wide, flat curve indicating widely scattered scores.

  • $z$-Score (Standard Score): Quantifies exactly how many standard deviations an individual score lies above or below the mean:         $z = \frac{X - \mu}{\sigma}$

The Normal Distribution and Skewed Data

• The Normal Distribution (Bell Curve): A continuous, perfectly symmetrical probability distribution where the mean, median, and mode converge at the exact center.

  • Many natural human biological and psychological traits (intelligence scores, height, weight) conform to this distribution.

• The Empirical Rule ($68\text{--}95\text{--}99.7$ Rule):

  • Approximately $68.26\%$ ($\approx 68\%$) of all data points fall within one standard deviation of the mean ($\mu \pm 1\sigma$).

  • Approximately $95.44\%$ ($\approx 95\%$) of all data points fall within two standard deviations of the mean ($\mu \pm 2\sigma$).

  • Approximately $99.74\%$ ($\approx 99.7\%$) of all data points fall within three standard deviations of the mean ($\mu \pm 3\sigma$).

• IQ Distribution Properties ($\mu = 100$, $\sigma = 15$):

  • Score $85$ ($z = -1.0$): Corresponds to the $16^{\text{th}}$ percentile rank (scoring equal to or better than $16\%$ of the population; lower than $84\%$).

  • Score $100$ ($z = 0.0$): Corresponds to the $50^{\text{th}}$ percentile rank.

  • Score $115$ ($z = +1.0$): Corresponds to the $84^{\text{th}}$ percentile rank.

  • Score $130$ ($z = +2.0$): Corresponds to the $98^{\text{th}}$ percentile rank.

• Skewed Distributions:

  • Positive Skew (Right-Skewed): Data contains extreme outliers pulled toward the high positive end (right tail). The mean is pulled upward in the direction of the tail:         $\text{Mode} < \text{Median} < \text{Mean}$

    • Income Disparity Example: In a town of 29 families where 26 earn below $\$70{,}000$ and 3 earn up to $\$710{,}000$, the mode is $\$20{,}000$, the median is $\$30{,}000$, and the mean is pulled up to $\$70{,}000$. The median reflects the true economic center.

  • Negative Skew (Left-Skewed): Data contains extreme outliers pulled toward the low negative end (left tail). The mean is pulled downward:         $\text{Mean} < \text{Median} < \text{Mode}$

    • Quiz Example: On an easy quiz where most students score 9 or 10, but a few unprepared students earn scores of 0, 1, or 2, the tail pulls leftward.

Correlational Analysis and Causation Pitfalls

• Correlation: A statistical measure of the strength and direction of the relationship between two variables, showing how well one variable predicts the other.

• The Pearson Correlation Coefficient ($r$): A mathematical index ranging strictly between $-1.00$ and $+1.00$:

  • Sign ($+\text{ or } -$): Denotes the mathematical direction of the relationship.

  • Absolute Value ($0.00\text{ to }1.00$): Denotes the strength of the relationship.

    • $r = -0.80$ represents a stronger relationship than $r = +0.40$.

    • $r = 0.00$ denotes no linear correlation.

• Directional Types of Correlations:

  • Positive Correlation: Both variables move in the same direction together (as $X$ increases, $Y$ increases).

    • Examples: Study hours and test scores; SAT verbal and SAT math scores ($r = +0.433$); husband's age and wife's age ($r = +0.97$); grip strength and arm strength ($r = +0.63$).

  • Negative Correlation: Variables move in opposite directions (as $X$ increases, $Y$ decreases).

    • Examples: Total lecture length and student attentiveness; party attendance frequency and grade point average.

• Linear Regression Line (Line of Best Fit): A straight mathematical trajectory plotted through a scatterplot minimizing squared residuals, summarizing trends and enabling predictions between variables.

• Crucial Rule: Correlation Does Not Equal Causation: Finding that two variables are correlated does not prove that one causes the other.

  • Directionality Problem: When two variables correlate, it is impossible to determine whether variable $X$ drives changes in variable $Y$, or whether variable $Y$ drives changes in variable $X$ (e.g., does marital status cause elevated happiness, or are happy people more likely to find a spouse?).

  • Third-Variable Problem (Confounding Variable): An extraneous, unmeasured third variable ($Z$) may be the true cause of the observed changes in both $X$ and $Y$.

    • Scenario: A strong positive correlation exists between community ice cream sales and municipal homicide rates. Neither causes the other; hot summer temperatures (the third variable) drive both higher ice cream consumption and outdoor interpersonal altercations.

• Spurious / Illusory Correlations: Imagining a correlation where none exists, driven by selective attention and coincidence:

  • National chocolate consumption per capita correlates strongly with Nobel laureates per capita ($r = +0.791, p < 0.0001$), confounded by underlying national wealth and educational infrastructure.

  • Global average temperature increases correlate with the historical decline in maritime pirates.

  • Annual US per capita cheese consumption correlates with accidental deaths from bedsheet tangling ($r = +0.947$).

• Regression Toward the Mean: The statistical tendency for extreme, unusual, or exceptional scores to return closer to the baseline average on subsequent testing.

  • Historical Error: In the 1950s, researchers observed a correlation between not smoking and developing Alzheimer's or Parkinson's disease. The apparent "protective" effect of smoking was an artifact of mortality—smokers routinely died of cardiovascular failure before reaching the older age brackets where neurodegenerative diseases manifest.