Research methods

MODULE 0.1–0.2: SCIENTIFIC THINKING

Scientific attitude / skepticism: Questioning claims and looking for evidence. Example: Checking statistics before believing a Patriots player is the best.

Critical thinking: Carefully analyzing evidence before making a conclusion. Example: Comparing a quarterback's stats and opponents.

Overconfidence: Being more confident than you should be. Example: A player guarantees a win despite poor performances.

Barnum effect: Believing a vague description specifically describes you. Example: A horoscope says you're competitive and sometimes nervous.

Hindsight bias: Believing you knew an outcome after it happened. Example: Saying you knew the Celtics would win after the game.

MODULE 0.3: SCIENTIFIC METHOD

Theory: An explanation of why something happens. Example: Sleep helps athletes perform better.

Hypothesis: A testable prediction. Example: Players who sleep eight hours will score more points.

Operational definition: Exactly how a variable is measured or manipulated. Example: Measuring performance by points scored.

Replication: Repeating a study to check whether results are consistent. Example: Repeating a study on athletes' sleep.

Case study: Studying one person or small group in depth. Example: Studying one MLB player's recovery.

Naturalistic observation: Watching behavior in a natural setting without interfering. Example: Observing athletes during practice.

Survey: Asking people questions to collect information. Example: Asking 100 athletes about confidence.

Generalizability: How well findings apply to a larger population. Example: Whether results from 50 MLB players apply to other professional baseball players.

Wording effect: How a question is phrased influences responses. Example: Asking whether unfair referees should be punished.

False consensus effect: Overestimating how many people agree with you. Example: A Celtics fan assumes everyone thinks Boston is the best.

Social desirability bias: Giving answers that make you look good. Example: An athlete exaggerates how often they train.

Sampling bias: Choosing a sample that doesn't represent the population. Example: Surveying only Patriots fans about America's favorite NFL team.

Population: The entire group researchers want to study. Example: Every NFL player.

Simple random sampling: Randomly selecting individuals from the population. Example: Randomly choosing 100 NFL players.

Cluster random sampling: Randomly selecting entire groups and studying everyone in them. Example: Choosing four NFL teams and studying every player.

Stratified random sampling: Dividing people into categories and randomly selecting individuals from each. Example: Selecting players from every NFL position.

MODULE 0.4: CORRELATION AND EXPERIMENTATION

Correlation: A relationship between two variables. Example: Practice hours and shooting accuracy.

Positive correlation: Both variables move in the same direction. Example: More practice is associated with more successful shots.

Negative correlation: Variables move in opposite directions. Example: More missed practices are associated with fewer points.

Correlation coefficient: A number from -1 to +1 showing the strength and direction of a linear relationship. Example: -0.90 means a strong negative correlation.

Scatterplot: A graph of dots showing the relationship between two variables. Example: Practice hours plotted against batting averages.

Correlation vs. causation: A relationship doesn't prove one variable causes another. Example: Larger crowds and more wins don't prove crowds cause wins.

Illusory correlation: Believing two things are related without sufficient evidence. Example: Thinking a lucky jersey makes the Patriots win.

Ex post facto study: Comparing existing groups after differences or events have already occurred. Example: Comparing athletes with and without previous concussions.

Experiment: Manipulating a variable to test cause and effect. Example: Assigning extra practice to see whether performance improves.

Experimental group: The group receiving the treatment being tested. Example: Players receiving extra training.

Control group: The comparison group that doesn't receive the treatment being tested. Example: Players continuing normal training.

Random assignment: Placing participants into experimental groups by chance. Example: Randomly assigning athletes to training groups.

Single-blind procedure: Participants don't know which treatment condition they're in. Example: Players don't know whether their drink contains the tested ingredient.

Double-blind procedure: Neither participants nor researchers interacting with them know who receives which treatment. Example: Neither athletes nor testing staff know who received the active drink.

Placebo effect: Changes caused by expectations rather than an active treatment. Example: Players improve because they believe a drink boosts performance.

Independent variable: What the researcher manipulates. Example: Number of extra practice hours assigned.

Dependent variable: What the researcher measures. Example: Number of successful free throws.

Confounding variable: An outside factor that could affect the results. Example: Some athletes have more experience.

Validity: How accurately a study measures or tests what it claims to. Example: Using shooting accuracy to measure shooting skill.

Demand characteristics: Clues that cause participants to change behavior because they guess the study's purpose. Example: Players try harder because they know effort is being studied.

Experimenter bias: Researcher expectations unintentionally influence results. Example: A coach rates favored players more positively.

MODULE 0.5: RESEARCH ETHICS

Quantitative research: Research using numerical data. Example: Comparing batting averages.

Qualitative research: Research using descriptions or experiences. Example: Interviewing athletes about motivation.

Research ethics: Rules protecting participants and ensuring responsible research. Example: Protecting athletes participating in a study.

Informed consent: Participants understand important study details and voluntarily agree. Example: Players agree after being informed of the risks.

Debriefing: Explaining the study's purpose and procedures afterward. Example: Researchers explain why athletes received different instructions.

Deception: Intentionally withholding or misleading participants about some study details. Example: Players aren't initially told the study is about confidence.

Psychological stress: Emotional distress researchers must minimize. Example: Avoiding unnecessary pressure during testing.

Culture in research: Considering cultural differences and avoiding cultural bias. Example: Studying athletes from different countries fairly.

Scientific integrity: Conducting and reporting research honestly. Example: Reporting accurate results even when they contradict a prediction.

MODULE 0.6: STATISTICS VOCABULARY

Descriptive statistics: Numbers used to summarize data. Example: The Celtics' average points per game.

Histogram: A graph showing frequencies of numerical values in intervals. Example: Number of MLB players in different home-run ranges.

Positive skew: A distribution with a long tail to the right. Example: Most players hit few home runs, but a few hit many.

Negative skew: A distribution with a long tail to the left. Example: Most athletes score highly on an easy skills test, but a few score very low.

Normal curve: A symmetrical bell-shaped distribution with most scores near the mean. Example: Most scores cluster around the average.

Effect size: How large a difference or relationship is. Example: How much extra training improves shooting accuracy.

Inferential statistics: Using sample data to draw conclusions about a population. Example: Studying 100 athletes to estimate a trend among all athletes in a league.

Statistical significance: Evidence that results are unlikely under the null hypothesis, often using p < 0.05. Example: A training study finds a statistically significant improvement.

P-value: Probability of results at least as extreme as those observed, assuming the null hypothesis is true. Example: p = 0.03 means a 3% probability under that assumption.

Null hypothesis: The assumption that there is no effect or relationship. Example: Extra training doesn't change shooting accuracy.

Type I error: Rejecting a true null hypothesis, also called a false positive. Example: Concluding training works when it doesn't.

Type II error: Failing to reject a false null hypothesis, also called a false negative. Example: Missing a real improvement caused by training.