Research Methods and Statistics Flashcards

Vocabulary flashcards covering fundamental concepts, definitions, study designs, validity types, measurement scales, and statistical terms from the research methods lecture notes.

Science

A systematic, empirical way of gaining knowledge based on observation and data (not authority/intuition) that tests falsifiable ideas and is public, repeatable, and self-correcting.

Population

The entire group you want to draw conclusions about.

Sample

The subset of the population you actually measure; results are generalized back to the population, so representativeness matters.

Basic research

Research that aims to expand knowledge and test theory with no immediate practical use required (e.g., how muscle fiber type relates to force production).

Applied research

Research that aims to solve a specific practical problem or evaluate an intervention (e.g., does a training program improve balance in older adults?).

Cyclical steps of the research process

Identify question/idea -> review literature -> form hypothesis -> design study & get IRB approval -> collect data -> analyze data -> interpret/conclude -> report results -> new questions (cycle repeats).

Hypothesis

A specific, testable prediction about the relationship between variables.

Theory

A well-supported, organized explanation of a phenomenon that generates testable hypotheses; broader and with more evidence than a hypothesis.

Quantitative research

Research that uses numerical data and statistical analysis.

Qualitative research

Research that uses non-numerical data (interviews, observations, open-ended responses) to understand meaning and experience.

Empirical

Based on direct observation or measurement, not just opinion or logic.

Reliability

Consistency of a measure: gives the same result on repeated measurement (test-retest, inter-rater, internal consistency).

Validity

Whether a measure or study actually measures/shows what it claims to.

Meta-analysis

A statistical method that combines results from many studies on the same question to estimate an overall effect.

Deception (in research)

Withholding or misleading participants about the study's true purpose; allowed only when necessary, risk is minimal, and participants are debriefed afterward.

Debriefing

A process after participation where researchers explain the true purpose, reveal any deception, answer questions, address distress/harm, and let participants withdraw their data; required when deception is used.

Nominal scale

A measurement scale consisting of categories only with no order (e.g., sex, sport played, blood type).

Ordinal scale

A measurement scale with ranked order, but spacing between ranks is unequal or unknown (e.g., finishing place in a race, pain rankings, Likert items).

Interval scale

A measurement scale with equal spacing between values but NO true zero (e.g., temperature in C or F, calendar year).

Ratio scale

A measurement scale with equal spacing AND a true zero, so ratios make sense (e.g., height, weight, VO2max, strength, reaction time).

NOIR

Memory trick for scales of measurement: Nominal, Ordinal, Interval, Ratio; each level includes the properties of the ones before it.

Correlational research

Research that measures the relationship between two variables WITHOUT manipulating either; used when manipulation is impossible/unethical, or to explore relationships and make predictions.

Correlation coefficient ($r$)

A statistic ranging from $-1$ to $+1$. The sign indicates direction ($+$ both rise together, $-$ one rises as other falls), and the absolute value indicates strength (near $0$ is weak, near $1$ is strong).

Correlation vs. causation

Correlation does NOT equal causation because a third variable or reverse direction could explain the relationship.

Coefficient of determination ($r^2$)

Proportion of variance in one variable explained by the other (e.g., $r = .70$ gives $r^2 = .49$, or $49\%$ shared variance).

Experimental research

Research where the investigator manipulates the independent variable, controls other factors, and typically uses random assignment; allows cause-and-effect conclusions.

Non-experimental research

Research with no manipulation (correlational, observational, descriptive, surveys); can show relationships/descriptions but not causation.

Internal validity

Confidence that the independent variable caused the change in the dependent variable and not something else (confounds, history, maturation, selection bias).

External validity

How well results generalize to other people, settings, and times.

Internal vs. external validity trade-off

The principle that tight laboratory control raises internal validity but can lower external validity.

Independent variable (IV)

The variable manipulated or grouped by the researcher (presumed cause).

Dependent variable (DV)

The variable measured as the outcome (presumed effect).

Pilot study

A small-scale trial run before the main study to test feasibility, procedures, timing, and equipment, find problems, and refine the design.

Pseudoexplanation

A statement that sounds like an explanation but only restates/labels the observation (circular reasoning) and cannot be tested (e.g., 'He is aggressive because he has an aggressive personality').

Pseudoscience

Claims presented as scientific but lacking scientific method: unfalsifiable, relies on anecdotes/testimonials, no peer review, resists disconfirming evidence, and uses vague language (e.g., astrology).

Volunteer bias

The tendency for volunteers to differ from non-volunteers (more motivated, educated, interested, healthier), meaning the sample may not represent the population.

Standard error of the mean (SE)

A measure of how much sample means would vary from the true population mean (precision of the mean); calculated as $SE = \frac{s}{\sqrt{n}}$, where $s$ is the sample standard deviation and $n$ is the sample size. Gets smaller as sample size increases.

Standard error vs. standard deviation

Standard deviation (SD) describes the spread of the data; standard error (SE) describes the precision of the sample mean.

Outlier: z-score rule

A rule identifying an outlier as a value beyond $\pm 3\,SD$ from the mean (some use $\pm 2.5$).

Outlier: IQR rule

A rule identifying an outlier as a value falling below $Q1 - 1.5 \times IQR$ or above $Q3 + 1.5 \times IQR$ (what boxplot whiskers show).

Bland-Altman plot

A graphical method to assess agreement by plotting the difference between two methods against their mean, with mean bias and $95\%$ limits of agreement ($\text{bias} \pm 1.96 \times SD\text{ of differences}$).

Intraclass correlation coefficient (ICC)

A statistic used for continuous data and assessing reliability or agreement between raters or trials.

Cohen's kappa

A statistic that measures agreement for categorical data, correcting for chance agreement.

p-value

The probability of getting results at least as extreme as yours IF the null hypothesis were true.

p-value decision rule

If $p < \alpha$ (usually $.05$), reject the null hypothesis (statistically significant). If $p \ge \alpha$, fail to reject the null hypothesis.